Adaptive thinking

.github/workflows/go-ci.yaml

@@ -6,44 +6,32 @@ on:
   pull_request:
     branches: [ main ]
 
-env:
-  GO_VERSION: '1.27.3'
-
 jobs:
   # Lint job
   lint:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v7
-      - uses: actions/setup-go@v7
-        with:
-          go-version: ${{ env.GO_VERSION }}
+      - uses: actions/checkout@v6
       - uses: golangci/golangci-lint-action@v9
         with:
-          version: v2.15.0
+          version: v2.12.2
           args: --timeout 5m
 
   # Test job
   test:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v7
-      - uses: actions/setup-go@v7
+      - uses: actions/checkout@v6
+      - uses: actions/setup-go@v6
         with:
-          go-version: ${{ env.GO_VERSION }}
+          go-version: '1.26.0'
       - run: go test ./... -v
 
   # Security scan job
   security:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v7
-      - uses: actions/setup-go@v7
+      - uses: actions/checkout@v6
+      - uses: securego/[email protected]
         with:
-          go-version: ${{ env.GO_VERSION }}
-      # gosec is installed rather than run as its container action: the
-      # published image ships its own Go, which lags behind and fails to load
-      # the module ("go.mod requires go >= X"). Installing it here builds gosec
-      # with the toolchain above, so CI's Go version is ours to choose.
-      - run: go install github.com/securego/gosec/v2/cmd/[email protected]
-      - run: gosec ./...
+          args: ./...

Dockerfile

@@ -1,6 +1,6 @@
 # Multi-stage build for Go Telegram Bot
 # Build stage
-FROM golang:1.27-alpine AS builder
+FROM golang:1.26-alpine AS builder
 
 # Install build dependencies including C compiler for CGO
 RUN apk add --no-cache git ca-certificates tzdata gcc musl-dev
@@ -18,7 +18,7 @@ RUN go mod download
 COPY . .
 
 # Build the application
-RUN CGO_ENABLED=1 GOOS=linux go build -o telegram-bot .
+RUN CGO_ENABLED=1 GOOS=linux go build -a -installsuffix cgo -o telegram-bot .
 
 # Runtime stage
 FROM alpine:latest

README.md

@@ -40,7 +40,12 @@ A scalable, multi-bot solution for Telegram using Go, GORM, and the Anthropic AP
 
 ### Native Deployment
 
-1. Clone the repository as above. Building needs Go 1.27 or later and a C compiler, because the SQLite driver uses cgo.
+1. Install using `go get`:
+
+   ```bash
+   go get -u github.com/HugeFrog24/go-telegram-bot
+   cd go-telegram-bot
+   ```
 
 2. Configure as above, then build:
    ```bash
@@ -68,121 +73,11 @@ Each bot is one JSON file in `config/` (see `config/default.json` for the templa
 
 | Key                | Type   | Default     | Description |
 | ------------------ | ------ | ----------- | ----------- |
-| `max_tokens`       | number | `1000`      | Maximum output tokens per reply. **Thinking tokens count toward this limit** — raise it (e.g. `4000`+) whenever `thinking` is `"adaptive"`, or a turn can spend the whole budget on reasoning and produce no text. Claude 5-generation models (Haiku 5.5, Sonnet 5 and later, Opus 5 and later, Fable) think even when `thinking` is omitted; on those, keep `effort` low or raise this limit. |
-| `temperature`      | number | *(ignored)* | No longer sent: since API version `2026-10-21` the Claude API rejects `temperature`, `top_p` and `top_k` on every model. Old configs still load, and the bot logs a startup notice; remove the key. |
+| `max_tokens`       | number | `1000`      | Maximum output tokens per reply. **Thinking tokens count toward this limit** — raise it (e.g. `4000`+) whenever `thinking` is `"adaptive"`, or a turn can spend the whole budget on reasoning and produce no text. |
 | `thinking`         | string | *(omitted)* | Reasoning mode: `"adaptive"` (the model decides when and how much to think) or `"disabled"`. Omit the key entirely to use the model's own API default. If the configured model doesn't support the chosen mode, the API rejects the request with a 400 — owners/admins see the raw error, regular users get the generic fallback. Check [Anthropic's model docs](https://platform.claude.com/docs/en/build-with-claude/adaptive-thinking) for per-model support. |
 | `thinking_display` | string | *(omitted)* | `"summarized"` or `"omitted"`. Only valid together with `"thinking": "adaptive"`. Controls whether the API returns a readable summary of the reasoning (logged, never sent to chat). Thinking is billed the same either way; when omitted, the API's per-model default applies. |
-| `effort`           | string | *(omitted)* | `"low"`, `"medium"`, `"high"`, `"xhigh"` or `"max"`. Caps how much reasoning and token spend a turn may use; the API default is `high` on most models and `medium` on Haiku 5.5 and Opus 5.5. For short-reply chat bots `"low"` or `"medium"` cuts cost noticeably. Rejected with a 400 on models older than Claude 4.6 (Haiku 4.5, Sonnet 4.5) — omit it there. |
-| `debounce_ms`      | number | *(omitted)* | Quiet window, in milliseconds, for coalescing rapid follow-up messages into a single turn. Omit or set `0` to disable. See [Coalescing rapid messages](#coalescing-rapid-messages) below. |
-| `stream_drafts`    | bool   | `false`     | Streams each reply into one Telegram draft that grows in place, with a native Stop button, and sends it as a single message when done. Private chats only. See [Draft streaming and the Stop button](#draft-streaming-and-the-stop-button) below. |
-| `cache_history`    | bool   | `true`      | Places a prompt-cache breakpoint on the trailing conversation block, so each turn reads the prior history from cache instead of reprocessing it at full price. Set `false` to cache only the system prompt. |
-| `web_search`       | object | *(omitted)* | Enables Anthropic's server-side web search (and, optionally, web fetch), sandboxed to a domain allowlist. Omit the key entirely to leave both tools off — an absent block sends byte-identical requests to before. See [Web search / fetch](#web-search--fetch) below. |
 
-Every reply logs one accounting line — `[usage] model=... in=... out=... thinking=... cache_read=... cache_write=... stop=...` — so thinking spend (billed even when its text is omitted) stays visible in `journalctl`/`docker compose logs`. A `stop=max_tokens` line is accompanied by an error-level warning that the reply was truncated.
-
-> [!TIP]
-> Watch `cache_read`/`cache_write` after changing prompts or models. A cache breakpoint below the model's minimum cacheable prefix fails **silently** — no error, just `cache_write=0` forever. The minimum is model-specific and not monotonic across generations (Haiku 4.5 needs 4096 tokens; Sonnet 4.6 needs 1024), so a short system prompt that caches fine on one model may never cache on another. Note also that chat memory is a sliding window: once it is full, each turn evicts the oldest message and changes the prefix, so `cache_read` on long-running chats will be lower than on fresh ones.
-
-### Coalescing rapid messages
-
-A user who sends "how do I do X", then "sorry typo", then "lmao" in five seconds would otherwise get three separate replies — the bot starts a full turn per message, because each Telegram update independently drives one. `debounce_ms` holds text messages in a per-chat buffer and resets the window on every new message, dispatching a single turn once the user stops typing.
-
-```json
-"debounce_ms": 2500
-```
-
-Reasonable windows are 1500–3000ms for ordinary chat and up to 8000ms for Telegram Business, where a person writing to a business account tends to send longer bursts. The maximum is 30000ms.
-
-Nothing is discarded while buffering. Each message is still persisted and added to chat memory the moment it arrives, so the single coalesced turn sees all of them — the buffer only decides *when* to answer, never *what* the model reads. Reply metadata (language, premium status, business connection) follows the most recent message in the batch.
-
-What deliberately does **not** wait:
-
-- **Commands** (`/stats`, `/clear`, …) dispatch immediately.
-- **Photos, albums, voice, and stickers** dispatch immediately and cancel any pending text window. The buffered text is not lost — it is already in memory, so the media turn answers it too.
-- **`/clear` and `/clear_hard`** cancel the buffer outright. Without this, the window would fire seconds after the wipe and replay the just-deleted messages back into memory.
-
-**Edits count as typing.** Mobile clients deliver autocorrect fixes as `edited_message` updates a few seconds after the send. The stored text (or photo caption) is replaced with the edit. If the message is still inside its quiet window, the window restarts. If a turn is already reading the old wording and has shown nothing yet, that turn restarts with the correction. An edit to a message that was already answered only corrects the record.
-
-#### Messages that arrive mid-turn
-
-No window catches every slow typist, and no LLM API lets you append to a request already in flight, so there is also a rule for a message that lands *after* the window closed, while the model is still working. At most one turn runs per chat, and a text flush that finds one running is handled in one of three ways:
-
-- **Dropped** if the running turn's context already includes the message. Its reply answers the new message too.
-- **Restarted** if nothing has been shown to the user yet. The running request is cancelled and rerun with the new message included. This is the only case that wastes tokens, and it is bounded by the time to first output; with `cache_history` on, the rerun reads the conversation prefix from cache.
-- **Collected** if output is already on screen. The message gets its own turn as soon as the running one finishes, instead of a second concurrent one. If the finished turn's context turns out to have included the message after all, the extra turn is skipped.
-
-Photos, voice, and stickers wait for a running turn to finish and are never restarted. Photo uploads still start immediately and overlap the running turn; only the model call queues behind it.
-
-While a turn is running the bot shows Telegram's "typing…" indicator (or "recording audio" while synthesising a voice reply), refreshed every 4 seconds because Telegram expires the status after 5. Without it, a debounce window reads as the bot ignoring you — which is what prompts users to send more messages in the first place.
-
-> [!NOTE]
-> A cancelled request is not free. Anthropic bills the input tokens and any output generated before the connection closed, and a web search the turn already ran is billed again on the rerun. That is why a restart only happens before anything was shown, and why a message that never dispatched is still the cheapest message of all.
-
-### Draft streaming and the Stop button
-
-```json
-"stream_drafts": true
-```
-
-With `stream_drafts` on, a reply in a private chat is streamed into one Telegram draft that grows in place (`sendMessageDraft`, available to every bot since Bot API 9.5), then sent as a single real message when the model finishes. Long replies split at 4096 characters. The draft is updated at most once per second, matching Telegram's per-chat send guidance, and re-sent every 10 seconds through silent stretches such as a web search so the preview does not expire.
-
-The draft carries Telegram's native **Stop** button (Bot API 10.3). Tapping it cancels the Anthropic stream; whatever the user had watched being written is sent as a normal message and stored in memory, so the next turn continues from the reply the user actually saw. A message the user sent before tapping Stop still gets its own turn afterwards. The draft text is capped at 4096 characters; the final message is split instead.
-
-What falls back to the previous one-message-per-text-block delivery:
-
-- **Business connections and group chats.** Telegram documents drafts for private chats only, and `sendMessageDraft` has no `business_connection_id` parameter.
-- **A rejected draft call.** After the first error the bot stops touching the draft and just sends the final message.
-
-> [!IMPORTANT]
-> While a draft is on screen, Telegram Android disables the composer: the user cannot send text or commands until the final message lands, and Telegram has said this is intended. The Stop button is their way out. For that reason the draft only appears once the model starts producing text; tool calls before that are covered by the typing indicator, which locks nothing. A March 2026 report described drafts rendering under "Pinned Messages" on iOS, so check the behaviour on your own client before enabling this for users.
-
-### Web search / fetch
-
-The `web_search` block wires in Anthropic's server-side `web_search` (and optionally `web_fetch`) tools, mirroring how `mcp_servers` wires in MCP toolsets: the engine provides the generic mechanism, and the per-bot config carries the policy. **No domains are hardcoded in the engine** — each profile declares its own allowlist. When the model decides a question needs current information, it runs a search server-side; results and any fetch targets are confined to the domains you list.
-
-```json
-"web_search": {
-  "allowed_domains": [
-    "example.com/help",
-    "docs.example.com"
-  ],
-  "max_uses": 3,
-  "fetch": true,
-  "max_content_tokens": 50000
-}
-```
-
-| Field                  | Type     | Description |
-| ---------------------- | -------- | ----------- |
-| `allowed_domains`      | string[] | Restricts **`web_search`** results (and, by default, `web_fetch` targets) to these. Each entry is a **domain with an optional path prefix and no scheme** — `example.com` or `example.com/help`. Mutually exclusive with `blocked_domains`. **The path only scopes `web_search`** — see the asymmetry note below. |
-| `blocked_domains`      | string[] | Excludes these domains instead of allowlisting. Mutually exclusive with `allowed_domains` (setting both fails validation at boot, since the API 400s). |
-| `fetch_allowed_domains`| string[] | **Host-only** allowlist for `web_fetch`, independent of the search list. Omit it and `web_fetch` reuses the *hosts* of `allowed_domains`. Set it to fetch a **narrower** set than you search — e.g. to keep a shared host (a social platform) search-only. Only meaningful with `fetch: true`. |
-| `max_uses`             | number   | Caps how many searches the model may run per turn. Omit for no cap. |
-| `fetch`                | bool     | Also enable `web_fetch`, which pulls full page content for URLs already surfaced by a search or pasted by the user (it cannot fetch model-invented URLs). Citations are always on when fetch is enabled, so fetched passages are sourceable. |
-| `max_content_tokens`   | number   | Caps the tokens a single `web_fetch` may pull into context. Only meaningful with `fetch: true`. |
-| `dynamic_filtering`    | bool     | Defaults to `true`. Lets the tools filter results inside Anthropic's sandbox before they enter context — see below. Set `false` for models older than Claude 4.6, which reject the tool unless it is called directly. |
-
-**Search/fetch path asymmetry (important).** Anthropic filters the two tools differently: `web_search` honors a path prefix (`example.com/help` matches only `example.com/help/...`), but **`web_fetch` matches on the host only — an entry that includes a path never matches any fetch URL**. The engine bridges this: `web_search` gets your entries verbatim, while `web_fetch` gets the host portion of each entry (path stripped, deduped). So a path-scoped `allowed_domains` still permits fetching across the whole host. If that host isn't wholly trusted, list the fetchable hosts explicitly in `fetch_allowed_domains` instead.
-
-**Search-only pattern for shared hosts (e.g. social).** To let the model *search* a single account on a shared platform without ever *fetching* the wider platform, put the account path in `allowed_domains` and leave its host out of `fetch_allowed_domains`:
-
-```json
-"web_search": {
-  "allowed_domains": ["helpcenter.example/hc", "x.com/youraccount"],
-  "fetch_allowed_domains": ["helpcenter.example"],
-  "fetch": true
-}
-```
-
-Here search is confined to your help center *and* your one social account, but fetch can only ever reach `helpcenter.example` — a pasted `x.com/someone-else` URL is not fetchable. Note the search side only sandboxes cleanly on platforms that keep an account's content under its handle path (X/Twitter, TikTok `@handle`, Facebook); use the `handle/` or `handle/*` form to avoid a prefix bleed. Instagram is an exception — posts live at `instagram.com/p/<code>`, not under the handle — so it can't be account-sandboxed by path.
-
-The allowlist is a **hard, server-side sandbox**, not a prompt request — off-list results are dropped and an off-list fetch target returns `url_not_allowed`, which is the exfiltration mitigation Anthropic recommends for bots processing untrusted input. Keep the list tight. A fetch can still fail for site-side reasons (bot protection such as Cloudflare returns `url_not_accessible`); the search snippet plus citations remain the reliable signal, so a full fetch is a bonus, not a dependency.
-
-`web_fetch` can only open a URL that a `web_search` result surfaced (or the user pasted) — never one the model invents or rebuilds from a page title; those return `url_not_in_prior_context`. So a fetch-enabled bot's prompt should steer it to **search first, then fetch a returned URL**, and to **re-search with a more specific query** (rather than guess a URL) when the exact page it wants isn't in the results.
-
-**Dynamic filtering.** The engine uses the `web_search_20260318` / `web_fetch_20260318` tools, which run the search inside Anthropic's code-execution sandbox and let the model discard irrelevant parts of a result before it reaches the context window. On token-heavy sources (a long docs page, a GitHub README) this is the difference between paying for the whole page and paying for the part that answered the question; there is no extra charge for the code execution it uses. It needs a **Claude 4.6 or later** model — on anything older (Haiku 4.5, Sonnet 4.5) the API rejects the request unless the tool is called directly, so set `"dynamic_filtering": false` on those profiles. Disabling it also restores Zero Data Retention eligibility, which the sandbox otherwise forfeits.
-
-Web search is **not free** — Anthropic bills per search (plus the tokens the results add) — so scope the allowlist and `max_uses` deliberately, and lean on a prompt that tells the bot to search only when a question actually turns on current or authoritative information. Search activity logs as `[web] ...` lines, and long multi-search turns are handled transparently (the engine follows Anthropic's `pause_turn` continuations up to a small cap, so a slow search doesn't truncate the reply).
+Every reply logs one accounting line — `[usage] model=... in=... out=... thinking=... stop=...` — so thinking spend (billed even when its text is omitted) stays visible in `journalctl`/`docker compose logs`. A `stop=max_tokens` line is accompanied by an error-level warning that the reply was truncated.
 
 > [!TIP]
 > For deep request/response debugging, the Anthropic Go SDK ships `option.WithDebugLog(...)` (dumps full HTTP bodies with auth headers redacted). It is not wired into the bot — dev-only, add it temporarily to the client constructor if you ever need wire-level traces.

anthropic.go

@@ -9,16 +9,14 @@ import (
 	"sync/atomic"
 	"time"
 
-	"github.com/anthropics/anthropic-sdk-go/v2"
-	"github.com/anthropics/anthropic-sdk-go/v2/packages/param"
+	"github.com/anthropics/anthropic-sdk-go"
+	"github.com/anthropics/anthropic-sdk-go/packages/param"
 )
 
 var ErrModelNotFound = errors.New("model not found or deprecated")
 
 const maxFileNotFoundRetries = 3
 
-const maxPauseTurnContinuations = 5
-
 const defaultMaxTokens = 1000
 
 const mcpUnsupportedSentinel = "format not currently supported by the Anthropic API"
@@ -27,7 +25,7 @@ var mcpUnsupportedCount atomic.Uint64
 
 type mcpCall struct{ server, name, input string }
 
-func (b *Bot) getAnthropicResponse(ctx context.Context, chatID int64, messages []anthropic.BetaMessageParam, isEmojiOnly bool, username string, firstName string, lastName string, isPremium bool, languageCode string, messageTime int, onSegment func(string) error, onProgress func(string)) (string, error) {
+func (b *Bot) getAnthropicResponse(ctx context.Context, chatID int64, messages []anthropic.BetaMessageParam, isEmojiOnly bool, username string, firstName string, lastName string, isPremium bool, languageCode string, messageTime int, onSegment func(string) error) (string, error) {
 	staticPrompt := strings.TrimSpace(b.config.SystemPrompts["custom_instructions"])
 
 	InfoLogger.Printf("Sending %d messages to Anthropic", len(messages))
@@ -40,6 +38,7 @@ func (b *Bot) getAnthropicResponse(ctx context.Context, chatID int64, messages [
 		Model:     b.config.Model,
 		MaxTokens: maxTokens,
 		Messages:  messages,
+		Betas:     []anthropic.AnthropicBeta{anthropic.AnthropicBetaFilesAPI2025_04_14},
 	}
 
 	if staticPrompt != "" {
@@ -58,20 +57,17 @@ func (b *Bot) getAnthropicResponse(ctx context.Context, chatID int64, messages [
 		params.System = blocks
 	}
 
+	if b.config.Temperature != nil {
+		params.Temperature = param.NewOpt(float64(*b.config.Temperature))
+	}
+
 	if thinking, ok := thinkingParamFromConfig(b.config.Thinking, b.config.ThinkingDisplay); ok {
 		params.Thinking = thinking
 	}
 
-	if b.config.Effort != "" {
-		params.OutputConfig = anthropic.BetaOutputConfigParam{
-			Effort: anthropic.BetaOutputConfigEffort(b.config.Effort),
-		}
-	}
-
-	var tools []anthropic.BetaToolUnionParam
-
 	if len(b.config.MCPServers) > 0 {
 		mcpServers := make([]anthropic.BetaRequestMCPServerURLDefinitionParam, 0, len(b.config.MCPServers))
+		tools := make([]anthropic.BetaToolUnionParam, 0, len(b.config.MCPServers))
 		for _, s := range b.config.MCPServers {
 			srv := anthropic.BetaRequestMCPServerURLDefinitionParam{
 				Name: s.Name,
@@ -99,23 +95,15 @@ func (b *Bot) getAnthropicResponse(ctx context.Context, chatID int64, messages [
 			tools = append(tools, anthropic.BetaToolUnionParam{OfMCPToolset: toolset})
 		}
 		params.MCPServers = mcpServers
+		params.Tools = tools
 		params.Betas = append(params.Betas, anthropic.AnthropicBetaMCPClient2025_11_20)
 	}
 
-	tools = append(tools, webSearchTools(b.config.WebSearch)...)
-
-	if len(tools) > 0 {
-		params.Tools = tools
+	for attempt := 0; attempt < maxFileNotFoundRetries; attempt++ {
+		joined, streamErr := b.streamMessages(ctx, params, onSegment)
+		if streamErr == nil {
+			return joined, nil
 		}
-
-	var full string
-	var lastMsg anthropic.BetaMessage
-	fileRetries, pauseContinuations := 0, 0
-	for {
-		// full is passed as the progress prefix so that, across pause_turn
-		// continuations, the user watches the whole reply so far.
-		joined, msg, streamErr := b.streamMessages(ctx, params, onSegment, onProgress, full)
-		if streamErr != nil {
 		var apiErr *anthropic.Error
 		if !errors.As(streamErr, &apiErr) || apiErr.StatusCode != http.StatusNotFound {
 			return "", fmt.Errorf("error creating Anthropic message: %w", streamErr)
@@ -124,107 +112,15 @@ func (b *Bot) getAnthropicResponse(ctx context.Context, chatID int64, messages [
 		if missingFileID == "" {
 			return "", fmt.Errorf("%w: %s", ErrModelNotFound, b.config.Model)
 		}
-			fileRetries++
-			if fileRetries > maxFileNotFoundRetries {
-				return "", fmt.Errorf("max self-heal retries (%d) exceeded: too many file_ids gone from anthropic", maxFileNotFoundRetries)
-			}
 		ErrorLogger.Printf("[%s] self-heal: stripping dead file_id %s from chat %d (attempt %d/%d)",
-				b.config.ID, missingFileID, chatID, fileRetries, maxFileNotFoundRetries)
+			b.config.ID, missingFileID, chatID, attempt+1, maxFileNotFoundRetries)
 		b.stripDeadFileIDFromMemory(chatID, missingFileID)
 		if _, cleanupErr := b.markFilesPendingCleanup(ctx, chatID, []string{missingFileID}); cleanupErr != nil {
 			ErrorLogger.Printf("[%s] mark files pending cleanup: %v", b.config.ID, cleanupErr)
 		}
 		params.Messages = b.prepareContextMessages(b.getOrCreateChatMemory(chatID))
-			continue
 	}
-
-		lastMsg = msg
-		full = joinNonEmpty(full, joined)
-
-		if msg.StopReason == anthropic.BetaStopReasonPauseTurn {
-			pauseContinuations++
-			if pauseContinuations > maxPauseTurnContinuations {
-				ErrorLogger.Printf("[%s] pause_turn continuations exceeded (%d); returning partial answer",
-					b.config.ID, maxPauseTurnContinuations)
-				break
-			}
-			params.Messages = append(params.Messages, msg.ToParam())
-			continue
-		}
-		break
-	}
-
-	if full == "" {
-		return "", emptyStreamError(string(lastMsg.StopReason), string(lastMsg.StopDetails.Category),
-			lastMsg.Usage.OutputTokensDetails.ThinkingTokens, params.MaxTokens)
-	}
-	return full, nil
-}
-
-func webSearchTools(cfg *WebSearchConfig) []anthropic.BetaToolUnionParam {
-	if cfg == nil {
-		return nil
-	}
-	// Dynamic filtering runs the tool inside code execution so results are
-	// filtered before they reach the context window. It requires a Claude 4.6+
-	// model; older ones must call the tool directly.
-	var callers []string
-	if !cfg.DynamicFilteringEnabled() {
-		callers = []string{"direct"}
-	}
-	search := &anthropic.BetaWebSearchTool20260318Param{
-		AllowedDomains: cfg.AllowedDomains,
-		BlockedDomains: cfg.BlockedDomains,
-		AllowedCallers: callers,
-	}
-	if cfg.MaxUses > 0 {
-		search.MaxUses = param.NewOpt(int64(cfg.MaxUses))
-	}
-	tools := []anthropic.BetaToolUnionParam{{OfWebSearchTool20260318: search}}
-
-	if cfg.Fetch {
-		fetchAllowed := cfg.FetchAllowedDomains
-		if len(fetchAllowed) == 0 {
-			fetchAllowed = cfg.AllowedDomains
-		}
-		fetch := &anthropic.BetaWebFetchTool20260318Param{
-			AllowedDomains: fetchHosts(fetchAllowed),
-			BlockedDomains: fetchHosts(cfg.BlockedDomains),
-			AllowedCallers: callers,
-			Citations:      anthropic.BetaCitationsConfigParam{Enabled: param.NewOpt(true)},
-		}
-		if cfg.MaxUses > 0 {
-			fetch.MaxUses = param.NewOpt(int64(cfg.MaxUses))
-		}
-		if cfg.MaxContentTokens > 0 {
-			fetch.MaxContentTokens = param.NewOpt(int64(cfg.MaxContentTokens))
-		}
-		tools = append(tools, anthropic.BetaToolUnionParam{OfWebFetchTool20260318: fetch})
-	}
-	return tools
-}
-
-// fetchHosts strips any path from each domain entry. web_fetch matches on host
-// only, so a path-scoped entry (valid for web_search) would otherwise never match
-// any fetch URL. Search keeps the path-scoped entries; fetch gets host-only.
-func fetchHosts(entries []string) []string {
-	if len(entries) == 0 {
-		return nil
-	}
-	seen := make(map[string]bool, len(entries))
-	hosts := make([]string, 0, len(entries))
-	for _, e := range entries {
-		host := e
-		if i := strings.IndexByte(host, '/'); i >= 0 {
-			host = host[:i]
-		}
-		if host == "" || seen[host] {
-			continue
-		}
-		seen[host] = true
-		hosts = append(hosts, host)
-	}
-	return hosts
+	return "", fmt.Errorf("max self-heal retries (%d) exceeded: too many file_ids gone from anthropic", maxFileNotFoundRetries)
 }
 
 func buildUserContext(username, firstName, lastName string, isPremium bool, languageCode string, messageTime int) string {
@@ -283,12 +179,7 @@ func thinkingParamFromConfig(mode, display string) (anthropic.BetaThinkingConfig
 	}
 }
 
-// streamMessages runs one streaming request and returns its text blocks joined.
-// onSegment fires with each completed text block; onProgress fires on every
-// text delta with the whole reply so far, prefix included, for sinks that
-// redraw in place rather than append. prefix is what earlier requests of the
-// same turn already produced and is not part of the returned text.
-func (b *Bot) streamMessages(ctx context.Context, params anthropic.BetaMessageNewParams, onSegment func(string) error, onProgress func(string), prefix string) (string, anthropic.BetaMessage, error) {
+func (b *Bot) streamMessages(ctx context.Context, params anthropic.BetaMessageNewParams, onSegment func(string) error) (string, error) {
 	stream := b.anthropicClient.Beta.Messages.NewStreaming(ctx, params)
 	defer func() {
 		if err := stream.Close(); err != nil {
@@ -297,9 +188,7 @@ func (b *Bot) streamMessages(ctx context.Context, params anthropic.BetaMessageNe
 	}()
 
 	var (
-		message                                           anthropic.BetaMessage
-		done                                              string   // this request's completed text blocks, joined
-		doneAll                                           = prefix // same, with the turn's earlier requests in front
+		allSegments                                       []string
 		currentKind                                       string
 		currentText                                       strings.Builder
 		currentThinking                                   strings.Builder
@@ -308,24 +197,31 @@ func (b *Bot) streamMessages(ctx context.Context, params anthropic.BetaMessageNe
 		currentTResultUseID, currentTResultServer         string
 		currentTResultIsError                             bool
 		currentTResultContent                             string
-		currentServerToolName, currentServerToolID        string
-		currentServerResult                               string
 		mcpCalls                                          = map[string]mcpCall{}
+		startInputTokens                                  int64
+		finalUsage                                        anthropic.BetaMessageDeltaUsage
+		stopReason                                        string
 	)
 
 	for stream.Next() {
 		e := stream.Current()
-		if accErr := message.Accumulate(e); accErr != nil {
-			ErrorLogger.Printf("[stream] accumulate failed: %v", accErr)
-		}
 		switch e.Type {
+		case "message_start":
+			startInputTokens = e.AsMessageStart().Message.Usage.InputTokens
+
+		case "message_delta":
+			md := e.AsMessageDelta()
+			finalUsage = md.Usage
+			if md.Delta.StopReason != "" {
+				stopReason = string(md.Delta.StopReason)
+			}
+
 		case "content_block_start":
 			cbs := e.AsContentBlockStart()
 			currentKind = cbs.ContentBlock.Type
 			currentText.Reset()
 			currentThinking.Reset()
 			currentInputJSON.Reset()
-			currentServerResult = ""
 			switch currentKind {
 			case "mcp_tool_use":
 				currentTUseName = cbs.ContentBlock.Name
@@ -336,11 +232,6 @@ func (b *Bot) streamMessages(ctx context.Context, params anthropic.BetaMessageNe
 				currentTResultServer = cbs.ContentBlock.ServerName
 				currentTResultIsError = cbs.ContentBlock.IsError
 				currentTResultContent = cbs.ContentBlock.JSON.Content.Raw()
-			case "server_tool_use":
-				currentServerToolName = cbs.ContentBlock.Name
-				currentServerToolID = cbs.ContentBlock.ID
-			case "web_search_tool_result", "web_fetch_tool_result":
-				currentServerResult = cbs.ContentBlock.JSON.Content.Raw()
 			}
 
 		case "content_block_delta":
@@ -349,16 +240,13 @@ func (b *Bot) streamMessages(ctx context.Context, params anthropic.BetaMessageNe
 			case "text_delta":
 				if currentKind == "text" {
 					currentText.WriteString(cbd.Delta.Text)
-					if onProgress != nil {
-						onProgress(joinNonEmpty(doneAll, currentText.String()))
-					}
 				}
 			case "thinking_delta":
 				if currentKind == "thinking" {
 					currentThinking.WriteString(cbd.Delta.Thinking)
 				}
 			case "input_json_delta":
-				if currentKind == "mcp_tool_use" || currentKind == "server_tool_use" {
+				if currentKind == "mcp_tool_use" {
 					currentInputJSON.WriteString(cbd.Delta.PartialJSON)
 				}
 			}
@@ -368,8 +256,7 @@ func (b *Bot) streamMessages(ctx context.Context, params anthropic.BetaMessageNe
 			case "text":
 				seg := strings.TrimSpace(currentText.String())
 				if seg != "" {
-					done = joinNonEmpty(done, seg)
-					doneAll = joinNonEmpty(doneAll, seg)
+					allSegments = append(allSegments, seg)
 					if onSegment != nil {
 						if cbErr := onSegment(seg); cbErr != nil {
 							ErrorLogger.Printf("[stream] onSegment failed: %v", cbErr)
@@ -398,15 +285,6 @@ func (b *Bot) streamMessages(ctx context.Context, params anthropic.BetaMessageNe
 						"(total=%d): server=%q tool=%q input=%s tool_use_id=%q",
 						b.config.ID, total, call.server, call.name, call.input, currentTResultUseID)
 				}
-			case "server_tool_use":
-				InfoLogger.Printf("[web] %s id=%q input=%s",
-					currentServerToolName, currentServerToolID, currentInputJSON.String())
-			case "web_search_tool_result", "web_fetch_tool_result":
-				preview := currentServerResult
-				if len(preview) > 500 {
-					preview = preview[:500] + "...(truncated)"
-				}
-				InfoLogger.Printf("[web] %s content=%s", currentKind, preview)
 			case "thinking", "redacted_thinking":
 				if summary := strings.TrimSpace(currentThinking.String()); summary != "" {
 					if len(summary) > 500 {
@@ -418,7 +296,7 @@ func (b *Bot) streamMessages(ctx context.Context, params anthropic.BetaMessageNe
 				}
 			default:
 				if currentKind != "" {
-					InfoLogger.Printf("[stream] block type=%q (unhandled)", currentKind)
+					InfoLogger.Printf("[mcp] block type=%q (unhandled)", currentKind)
 				}
 			}
 			currentKind = ""
@@ -426,52 +304,33 @@ func (b *Bot) streamMessages(ctx context.Context, params anthropic.BetaMessageNe
 	}
 
 	if err := stream.Err(); err != nil {
-		return "", message, err
+		return "", err
 	}
 
-	stopReason := string(message.StopReason)
-	if stopReason != "" || message.Usage.OutputTokens > 0 {
-		// cache_read/cache_write make the caching configuration falsifiable: a
-		// breakpoint below the model's minimum cacheable prefix fails silently,
-		// reporting cache_write=0 rather than raising an error.
-		InfoLogger.Printf("[usage] model=%s in=%d out=%d thinking=%d cache_read=%d cache_write=%d stop=%s",
-			params.Model, message.Usage.InputTokens, message.Usage.OutputTokens,
-			message.Usage.OutputTokensDetails.ThinkingTokens,
-			message.Usage.CacheReadInputTokens, message.Usage.CacheCreationInputTokens,
-			stopReason)
-		if message.StopReason == anthropic.BetaStopReasonMaxTokens {
+	if stopReason != "" || finalUsage.OutputTokens > 0 {
+		inputTokens := finalUsage.InputTokens
+		if inputTokens == 0 {
+			inputTokens = startInputTokens
+		}
+		InfoLogger.Printf("[usage] model=%s in=%d out=%d thinking=%d stop=%s",
+			params.Model, inputTokens, finalUsage.OutputTokens,
+			finalUsage.OutputTokensDetails.ThinkingTokens, stopReason)
+		if stopReason == "max_tokens" {
 			ErrorLogger.Printf("[usage] response truncated at max_tokens=%d - raise max_tokens (thinking counts toward it)",
 				params.MaxTokens)
 		}
 	}
 
-	return done, message, nil
+	if len(allSegments) == 0 {
+		return "", emptyStreamError(stopReason, finalUsage.OutputTokensDetails.ThinkingTokens, params.MaxTokens)
+	}
+	return strings.Join(allSegments, "\n\n"), nil
 }
 
-// joinNonEmpty is the one rule for assembling reply text from parts: a blank
-// line between them, empty parts skipped. The live draft, the final message,
-// and the stored partial all go through it, so they can never disagree.
-func joinNonEmpty(a, b string) string {
-	switch {
-	case a == "":
-		return b
-	case b == "":
-		return a
-	default:
-		return a + "\n\n" + b
-	}
-}
-
-func emptyStreamError(stopReason, refusalCategory string, thinkingTokens, maxTokens int64) error {
-	switch stopReason {
-	case "max_tokens":
+func emptyStreamError(stopReason string, thinkingTokens, maxTokens int64) error {
+	if stopReason == "max_tokens" {
 		return fmt.Errorf("output budget exhausted before any text (thinking used %d of %d max_tokens) - raise max_tokens",
 			thinkingTokens, maxTokens)
-	case "refusal":
-		if refusalCategory == "" {
-			return fmt.Errorf("the model declined the request (refusal)")
-		}
-		return fmt.Errorf("the model declined the request (refusal, category %s)", refusalCategory)
 	}
 	return fmt.Errorf("unexpected response format from Anthropic")
 }

anthropic_files.go

@@ -9,7 +9,7 @@ import (
 	"strings"
 	"time"
 
-	"github.com/anthropics/anthropic-sdk-go/v2"
+	"github.com/anthropics/anthropic-sdk-go"
 )
 
 const fileNotFoundPrefix = "File not found: "
@@ -19,8 +19,9 @@ func formatUploadFilename(botID uint, chatID int64, tgMessageID int, ext string)
 }
 
 func (b *Bot) uploadImageToAnthropic(ctx context.Context, data []byte, filename, contentType string) (string, error) {
-	resp, err := b.anthropicClient.Files.Upload(ctx, anthropic.FileUploadParams{
+	resp, err := b.anthropicClient.Beta.Files.Upload(ctx, anthropic.BetaFileUploadParams{
 		File:  anthropic.File(bytes.NewReader(data), filename, contentType),
+		Betas: []anthropic.AnthropicBeta{anthropic.AnthropicBetaFilesAPI2025_04_14},
 	})
 	if err != nil {
 		return "", fmt.Errorf("anthropic files upload: %w", err)
@@ -29,7 +30,9 @@ func (b *Bot) uploadImageToAnthropic(ctx context.Context, data []byte, filename,
 }
 
 func (b *Bot) deleteFileFromAnthropic(ctx context.Context, fileID string) error {
-	_, err := b.anthropicClient.Files.Delete(ctx, fileID, anthropic.FileDeleteParams{})
+	_, err := b.anthropicClient.Beta.Files.Delete(ctx, fileID, anthropic.BetaFileDeleteParams{
+		Betas: []anthropic.AnthropicBeta{anthropic.AnthropicBetaFilesAPI2025_04_14},
+	})
 	if err == nil {
 		return nil
 	}

anthropic_test.go

@@ -6,7 +6,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/anthropics/anthropic-sdk-go/v2"
+	"github.com/anthropics/anthropic-sdk-go"
 )
 
 func TestTimeContextFor(t *testing.T) {
@@ -127,232 +127,17 @@ func TestBackwardCompatibleParams(t *testing.T) {
 	}
 }
 
-func TestWebSearchTools(t *testing.T) {
-	t.Run("nil config yields no tools", func(t *testing.T) {
-		if tools := webSearchTools(nil); tools != nil {
-			t.Errorf("webSearchTools(nil) = %v, want nil", tools)
-		}
-	})
-
-	t.Run("search only when fetch off", func(t *testing.T) {
-		tools := webSearchTools(&WebSearchConfig{
-			AllowedDomains: []string{"example.com/hc"},
-			MaxUses:        3,
-		})
-		if len(tools) != 1 {
-			t.Fatalf("got %d tools, want 1 (search only)", len(tools))
-		}
-		if tools[0].OfWebSearchTool20260318 == nil {
-			t.Fatalf("tools[0] is not a web_search tool")
-		}
-		if tools[0].OfWebFetchTool20260318 != nil {
-			t.Error("web_fetch tool present but fetch is off")
-		}
-	})
-
-	t.Run("search + fetch with allowlist and citations", func(t *testing.T) {
-		tools := webSearchTools(&WebSearchConfig{
-			AllowedDomains:   []string{"example.com/hc", "docs.example.com"},
-			MaxUses:          3,
-			Fetch:            true,
-			MaxContentTokens: 50000,
-		})
-		if len(tools) != 2 {
-			t.Fatalf("got %d tools, want 2 (search + fetch)", len(tools))
-		}
-
-		search := tools[0].OfWebSearchTool20260318
-		if search == nil {
-			t.Fatalf("tools[0] is not a web_search tool")
-		}
-		if !sameStrings(search.AllowedDomains, []string{"example.com/hc", "docs.example.com"}) {
-			t.Errorf("search AllowedDomains = %v, want the path-scoped list unchanged", search.AllowedDomains)
-		}
-		if search.MaxUses.Value != 3 {
-			t.Errorf("search MaxUses = %d, want 3", search.MaxUses.Value)
-		}
-
-		fetch := tools[1].OfWebFetchTool20260318
-		if fetch == nil {
-			t.Fatalf("tools[1] is not a web_fetch tool")
-		}
-		if !sameStrings(fetch.AllowedDomains, []string{"example.com", "docs.example.com"}) {
-			t.Errorf("fetch AllowedDomains = %v, want host-only [example.com docs.example.com]", fetch.AllowedDomains)
-		}
-		if fetch.MaxContentTokens.Value != 50000 {
-			t.Errorf("fetch MaxContentTokens = %d, want 50000", fetch.MaxContentTokens.Value)
-		}
-		if fetch.Citations.Enabled.Value != true {
-			t.Error("fetch citations not enabled")
-		}
-	})
-
-	t.Run("fetch hosts are deduped", func(t *testing.T) {
-		tools := webSearchTools(&WebSearchConfig{
-			AllowedDomains: []string{"a.com/x", "a.com/y", "b.com"},
-			Fetch:          true,
-		})
-		fetch := tools[1].OfWebFetchTool20260318
-		if fetch == nil {
-			t.Fatalf("tools[1] is not a web_fetch tool")
-		}
-		if !sameStrings(fetch.AllowedDomains, []string{"a.com", "b.com"}) {
-			t.Errorf("fetch AllowedDomains = %v, want deduped [a.com b.com]", fetch.AllowedDomains)
-		}
-	})
-
-	t.Run("social host is search-only via fetch_allowed_domains", func(t *testing.T) {
-		tools := webSearchTools(&WebSearchConfig{
-			AllowedDomains:      []string{"helpshift.example/hc", "x.com/thatskygame"},
-			FetchAllowedDomains: []string{"helpshift.example"},
-			Fetch:               true,
-		})
-		search := tools[0].OfWebSearchTool20260318
-		if search == nil {
-			t.Fatalf("tools[0] is not a web_search tool")
-		}
-		var searchHasSocial bool
-		for _, d := range search.AllowedDomains {
-			if d == "x.com/thatskygame" {
-				searchHasSocial = true
-			}
-		}
-		if !searchHasSocial {
-			t.Errorf("search AllowedDomains = %v, want it to include x.com/thatskygame", search.AllowedDomains)
-		}
-
-		fetch := tools[1].OfWebFetchTool20260318
-		if fetch == nil {
-			t.Fatalf("tools[1] is not a web_fetch tool")
-		}
-		if !sameStrings(fetch.AllowedDomains, []string{"helpshift.example"}) {
-			t.Errorf("fetch AllowedDomains = %v, want only [helpshift.example]", fetch.AllowedDomains)
-		}
-		for _, d := range fetch.AllowedDomains {
-			if strings.HasPrefix(d, "x.com") {
-				t.Errorf("fetch AllowedDomains leaked the social host: %v", fetch.AllowedDomains)
-			}
-		}
-	})
-
-	t.Run("wire shape carries allowed_domains", func(t *testing.T) {
-		tools := webSearchTools(&WebSearchConfig{
-			AllowedDomains: []string{"thatgamecompany.helpshift.com/hc"},
-			MaxUses:        2,
-			Fetch:          true,
-		})
-		raw, err := json.Marshal(tools)
-		if err != nil {
-			t.Fatalf("marshal: %v", err)
-		}
-		body := string(raw)
-		for _, want := range []string{
-			"web_search_20260318",
-			"web_fetch_20260318",
-			"thatgamecompany.helpshift.com/hc",
-			"allowed_domains",
-		} {
-			if !strings.Contains(body, want) {
-				t.Errorf("wire body missing %q:\n%s", want, body)
-			}
-		}
-	})
-}
-
-func sameStrings(got, want []string) bool {
-	if len(got) != len(want) {
-		return false
-	}
-	for i := range got {
-		if got[i] != want[i] {
-			return false
-		}
-	}
-	return true
-}
-
-func TestFetchHosts(t *testing.T) {
-	cases := []struct {
-		name string
-		in   []string
-		want []string
-	}{
-		{"nil in nil out", nil, nil},
-		{"empty in nil out", []string{}, nil},
-		{"host passthrough", []string{"example.com"}, []string{"example.com"}},
-		{"strip path", []string{"example.com/hc/en"}, []string{"example.com"}},
-		{"dedup after strip", []string{"a.com/x", "a.com/y"}, []string{"a.com"}},
-		{"preserve order and subdomains", []string{"docs.example.com/a", "example.com"}, []string{"docs.example.com", "example.com"}},
-		{"drop empty leading slash", []string{"/oops", "ok.com"}, []string{"ok.com"}},
-	}
-	for _, tc := range cases {
-		t.Run(tc.name, func(t *testing.T) {
-			if got := fetchHosts(tc.in); !sameStrings(got, tc.want) {
-				t.Errorf("fetchHosts(%v) = %v, want %v", tc.in, got, tc.want)
-			}
-		})
-	}
-}
-
 func TestEmptyStreamError(t *testing.T) {
-	err := emptyStreamError("max_tokens", "", 3900, 4000)
+	err := emptyStreamError("max_tokens", 3900, 4000)
 	for _, want := range []string{"output budget exhausted", "3900", "4000"} {
 		if !strings.Contains(err.Error(), want) {
 			t.Errorf("max_tokens case: %q missing %q", err.Error(), want)
 		}
 	}
-	if got := emptyStreamError("refusal", "cyber", 0, 1000).Error(); got != "the model declined the request (refusal, category cyber)" {
-		t.Errorf("refusal case = %q", got)
-	}
-	if got := emptyStreamError("refusal", "", 0, 1000).Error(); got != "the model declined the request (refusal)" {
-		t.Errorf("uncategorized refusal case = %q", got)
-	}
-	if got := emptyStreamError("end_turn", "", 0, 1000).Error(); got != "unexpected response format from Anthropic" {
+	if got := emptyStreamError("end_turn", 0, 1000).Error(); got != "unexpected response format from Anthropic" {
 		t.Errorf("generic case = %q", got)
 	}
-	if got := emptyStreamError("", "", 0, 1000).Error(); got != "unexpected response format from Anthropic" {
+	if got := emptyStreamError("", 0, 1000).Error(); got != "unexpected response format from Anthropic" {
 		t.Errorf("no-stop-reason case = %q", got)
 	}
 }
-
-func TestDynamicFilteringEnabled(t *testing.T) {
-	off := false
-	on := true
-	cases := []struct {
-		name string
-		cfg  *WebSearchConfig
-		want bool
-	}{
-		{"nil config defaults on", nil, true},
-		{"unset defaults on", &WebSearchConfig{}, true},
-		{"explicit true", &WebSearchConfig{DynamicFiltering: &on}, true},
-		{"explicit false", &WebSearchConfig{DynamicFiltering: &off}, false},
-	}
-	for _, tc := range cases {
-		if got := tc.cfg.DynamicFilteringEnabled(); got != tc.want {
-			t.Errorf("%s: got %v, want %v", tc.name, got, tc.want)
-		}
-	}
-}
-
-// Models older than Claude 4.6 reject the 2026 web tools unless the tool is
-// restricted to direct calls, so the opt-out must reach both tools.
-func TestWebSearchTools_AllowedCallers(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	off := false
-
-	tools := webSearchTools(&WebSearchConfig{Fetch: true})
-	if got := tools[0].OfWebSearchTool20260318.AllowedCallers; got != nil {
-		t.Errorf("default should omit allowed_callers, got %v", got)
-	}
-	if got := tools[1].OfWebFetchTool20260318.AllowedCallers; got != nil {
-		t.Errorf("default should omit allowed_callers on fetch, got %v", got)
-	}
-
-	tools = webSearchTools(&WebSearchConfig{Fetch: true, DynamicFiltering: &off})
-	if got := tools[0].OfWebSearchTool20260318.AllowedCallers; len(got) != 1 || got[0] != "direct" {
-		t.Errorf("search allowed_callers = %v, want [direct]", got)
-	}
-	if got := tools[1].OfWebFetchTool20260318.AllowedCallers; len(got) != 1 || got[0] != "direct" {
-		t.Errorf("fetch allowed_callers = %v, want [direct]", got)
-	}
-}

bot.go

@@ -8,8 +8,8 @@ import (
 	"sync"
 	"time"
 
-	"github.com/anthropics/anthropic-sdk-go/v2"
-	"github.com/anthropics/anthropic-sdk-go/v2/option"
+	"github.com/anthropics/anthropic-sdk-go"
+	"github.com/anthropics/anthropic-sdk-go/option"
 	"github.com/go-telegram/bot"
 	"github.com/go-telegram/bot/models"
 	"gorm.io/gorm"
@@ -29,12 +29,6 @@ type Bot struct {
 	botID           uint
 	albumBuffers    map[string]*pendingAlbum
 	albumBuffersMu  sync.Mutex
-	intakeBuffers   map[int64]*pendingIntake
-	intakeBuffersMu sync.Mutex
-	intakeSeq       uint64
-	turns           map[int64]*chatTurn
-	turnsMu         sync.Mutex
-	turnSeq         uint64
 }
 
 func messageType(msg *models.Message) string {
@@ -83,13 +77,7 @@ func NewBot(db *gorm.DB, config BotConfig, clock Clock, tgClient TelegramClient)
 		return nil, err
 	}
 
-	// An explicitly passed credential is used as is, so an empty key would go
-	// out as an empty bearer token; leave it out to use the SDK's default chain.
-	var anthropicOpts []option.RequestOption
-	if config.AnthropicAPIKey != "" {
-		anthropicOpts = append(anthropicOpts, option.WithAPIKey(config.AnthropicAPIKey))
-	}
-	anthropicClient := anthropic.NewClient(anthropicOpts...)
+	anthropicClient := anthropic.NewClient(option.WithAPIKey(config.AnthropicAPIKey))
 
 	b := &Bot{
 		db:              db,
@@ -102,8 +90,6 @@ func NewBot(db *gorm.DB, config BotConfig, clock Clock, tgClient TelegramClient)
 		botID:           botEntry.ID,
 		tgBot:           tgClient,
 		albumBuffers:    make(map[string]*pendingAlbum),
-		intakeBuffers:   make(map[int64]*pendingIntake),
-		turns:           make(map[int64]*chatTurn),
 	}
 
 	if tgClient == nil {
@@ -286,54 +272,14 @@ func (b *Bot) addMessageToChatMemory(chatMemory *ChatMemory, message Message) {
 	}
 }
 
-// prepareContextMessages converts chat memory into API messages.
 func (b *Bot) prepareContextMessages(chatMemory *ChatMemory) []anthropic.BetaMessageParam {
 	b.chatMemoriesMu.RLock()
 	defer b.chatMemoriesMu.RUnlock()
-	return b.buildContextMessages(chatMemory)
-}
 
-// snapshotForTurn builds the model context for turn and records on it the
-// newest user message that context includes, both read under one lock so they
-// cannot disagree. The recorded id is what lets a later flush tell whether the
-// running turn already answers it.
-func (b *Bot) snapshotForTurn(turn *chatTurn, chatMemory *ChatMemory) []anthropic.BetaMessageParam {
-	b.chatMemoriesMu.RLock()
-	contextMessages := b.buildContextMessages(chatMemory)
-	covered := newestUserMessageID(chatMemory)
-	b.chatMemoriesMu.RUnlock()
-
-	b.markCovered(turn, covered)
-	return contextMessages
-}
-
-// newestUserMessageID returns the highest database id among the user messages
-// in memory. Ids are monotonic, so "newer" and "higher" agree.
-func newestUserMessageID(chatMemory *ChatMemory) uint {
-	var newest uint
-	for _, msg := range chatMemory.Messages {
-		if msg.IsUser && msg.ID > newest {
-			newest = msg.ID
-		}
-	}
-	return newest
-}
-
-func (b *Bot) newestUserMessageIDInChat(chatID int64) uint {
-	chatMemory := b.getOrCreateChatMemory(chatID)
-	b.chatMemoriesMu.RLock()
-	defer b.chatMemoriesMu.RUnlock()
-	return newestUserMessageID(chatMemory)
-}
-
-// buildContextMessages is prepareContextMessages without the lock.
-func (b *Bot) buildContextMessages(chatMemory *ChatMemory) []anthropic.BetaMessageParam {
 	InfoLogger.Printf("Chat memory contains %d messages", len(chatMemory.Messages))
-	if b.config.DebugScreening {
 	for i, msg := range chatMemory.Messages {
 		InfoLogger.Printf("Message %d: IsUser=%v, Text=%q Images=%d", i, msg.IsUser, msg.Text, len(msg.ImageFileIDs))
 	}
-	}
 
 	var contextMessages []anthropic.BetaMessageParam
 	for _, msg := range chatMemory.Messages {
@@ -345,51 +291,16 @@ func (b *Bot) buildContextMessages(chatMemory *ChatMemory) []anthropic.BetaMessa
 		if msg.IsUser {
 			param = anthropic.NewBetaUserMessage(blocks...)
 		} else {
-			param = anthropic.BetaMessageParam{OfAssistant: &anthropic.BetaAssistantMessageParam{
+			param = anthropic.BetaMessageParam{
+				Role:    anthropic.BetaMessageParamRoleAssistant,
 				Content: blocks,
-			}}
+			}
 		}
 		contextMessages = append(contextMessages, param)
 	}
-
-	if b.config.CacheHistoryEnabled() {
-		markTrailingCacheBreakpoint(contextMessages)
-	}
 	return contextMessages
 }
 
-// markTrailingCacheBreakpoint puts a cache_control breakpoint on the final
-// content block of the conversation, so the next turn reads the whole prefix
-// from cache instead of reprocessing it. The system prompt keeps its own
-// breakpoint; tools and system render ahead of messages, so the two compose.
-//
-// Caveat worth knowing when reading [usage] lines: chat memory is a sliding
-// window. Once it is full, each new turn evicts the oldest message, which
-// changes the prefix and forces a miss. Until then, and for any chat shorter
-// than the window, this converts a full-price reread into a cache read.
-func markTrailingCacheBreakpoint(messages []anthropic.BetaMessageParam) {
-	if len(messages) == 0 {
-		return
-	}
-	var blocks []anthropic.BetaContentBlockParamUnion
-	switch last := messages[len(messages)-1]; {
-	case last.OfUser != nil:
-		blocks = last.OfUser.Content
-	case last.OfAssistant != nil:
-		blocks = last.OfAssistant.Content
-	}
-	if len(blocks) == 0 {
-		return
-	}
-
-	switch last := &blocks[len(blocks)-1]; {
-	case last.OfText != nil:
-		last.OfText.CacheControl = anthropic.NewBetaCacheControlEphemeralParam()
-	case last.OfImage != nil:
-		last.OfImage.CacheControl = anthropic.NewBetaCacheControlEphemeralParam()
-	}
-}
-
 func contentBlocksForMessage(msg Message) []anthropic.BetaContentBlockParamUnion {
 	var blocks []anthropic.BetaContentBlockParamUnion
 	if msg.IsUser && len(msg.ImageFileIDs) > 0 {
@@ -756,7 +667,6 @@ func (b *Bot) screenIncomingMessage(message *models.Message) (Message, error) {
 	}
 
 	userMessage := b.createMessage(message.Chat.ID, message.From.ID, message.From.Username, userRole, messageText, true)
-	userMessage.TelegramMessageID = message.ID
 
 	if message.Sticker != nil {
 		userMessage.StickerFileID = message.Sticker.FileID
@@ -796,18 +706,16 @@ func (b *Bot) screenOutgoingMessage(chatID int64, response string) (Message, err
 		return Message{}, err
 	}
 
-	// Mark every outstanding user message in the chat, not just the newest one.
-	// A coalesced turn answers the whole batch, so a single-row update would
-	// leave the earlier messages permanently unanswered. This also drops an
-	// UPDATE ... ORDER BY ... LIMIT, which stock SQLite builds do not support.
 	now := time.Now()
 	err := b.db.Model(&Message{}).
 		Where("chat_id = ? AND bot_id = ? AND is_user = ? AND answered_on IS NULL",
 			chatID, b.botID, true).
+		Order("timestamp DESC").
+		Limit(1).
 		Update("answered_on", now).Error
 
 	if err != nil {
-		ErrorLogger.Printf("Error marking user messages as answered: %v", err)
+		ErrorLogger.Printf("Error marking user message as answered: %v", err)
 	}
 
 	chatMemory := b.getOrCreateChatMemory(chatID)

config.go

@@ -5,9 +5,7 @@ import (
 	"fmt"
 	"os"
 	"path/filepath"
-	"slices"
 	"strings"
-	"time"
 )
 
 type MCPServer struct {
@@ -17,20 +15,6 @@ type MCPServer struct {
 	AllowedTools       []string `json:"allowed_tools,omitempty"`
 }
 
-type WebSearchConfig struct {
-	AllowedDomains      []string `json:"allowed_domains,omitempty"`
-	BlockedDomains      []string `json:"blocked_domains,omitempty"`
-	FetchAllowedDomains []string `json:"fetch_allowed_domains,omitempty"`
-	DynamicFiltering    *bool    `json:"dynamic_filtering,omitempty"`
-	MaxUses             int      `json:"max_uses,omitempty"`
-	Fetch               bool     `json:"fetch,omitempty"`
-	MaxContentTokens    int      `json:"max_content_tokens,omitempty"`
-}
-
-// validEfforts are the output_config.effort levels the API accepts. Models
-// older than Claude 4.6 (Haiku 4.5, Sonnet 4.5) reject the parameter entirely.
-var validEfforts = []string{"low", "medium", "high", "xhigh", "max"}
-
 const (
 	ThinkingModeAdaptive      = "adaptive"
 	ThinkingModeDisabled      = "disabled"
@@ -38,36 +22,6 @@ const (
 	ThinkingDisplayOmitted    = "omitted"
 )
 
-// maxDebounceMs bounds debounce_ms. Beyond this the bot reads as unresponsive
-// rather than deliberate, and the coalesced turn drifts far enough from the
-// user's last message that the reply feels stale.
-const maxDebounceMs = 30000
-
-// DebounceWindow is the quiet period an intake buffer waits before dispatching a
-// coalesced turn. Zero disables debouncing entirely (the default), matching the
-// opt-in behavior of comparable gateways.
-func (c *BotConfig) DebounceWindow() time.Duration {
-	if c.DebounceMs <= 0 {
-		return 0
-	}
-	return time.Duration(c.DebounceMs) * time.Millisecond
-}
-
-// DynamicFilteringEnabled reports whether the web tools may run inside code
-// execution to filter results before they reach the context window. Defaults to
-// true; set "dynamic_filtering": false for models older than Claude 4.6, which
-// reject the tool unless it is called directly.
-func (c *WebSearchConfig) DynamicFilteringEnabled() bool {
-	return c == nil || c.DynamicFiltering == nil || *c.DynamicFiltering
-}
-
-// CacheHistoryEnabled reports whether a cache_control breakpoint should be placed
-// on the trailing conversation block in addition to the system prompt. Defaults
-// to true; set "cache_history": false to opt out.
-func (c *BotConfig) CacheHistoryEnabled() bool {
-	return c.CacheHistory == nil || *c.CacheHistory
-}
-
 type BotConfig struct {
 	ID                string            `json:"id"`
 	TelegramToken     string            `json:"telegram_token"`
@@ -76,14 +30,10 @@ type BotConfig struct {
 	MessagePerDay     int               `json:"messages_per_day"`
 	TempBanDuration   string            `json:"temp_ban_duration"`
 	Model             string            `json:"model"`
-	Temperature       *float32          `json:"temperature,omitempty"` // never sent (rejected since API 2026-10-21); parsed only to flag stale configs
+	Temperature       *float32          `json:"temperature,omitempty"`
 	MaxTokens         int               `json:"max_tokens,omitempty"`
 	Thinking          string            `json:"thinking,omitempty"`
 	ThinkingDisplay   string            `json:"thinking_display,omitempty"`
-	Effort            string            `json:"effort,omitempty"`
-	DebounceMs        int               `json:"debounce_ms,omitempty"`
-	StreamDrafts      bool              `json:"stream_drafts,omitempty"`
-	CacheHistory      *bool             `json:"cache_history,omitempty"`
 	SystemPrompts     map[string]string `json:"system_prompts"`
 	Active            bool              `json:"active"`
 	OwnerTelegramID   int64             `json:"owner_telegram_id"`
@@ -93,7 +43,6 @@ type BotConfig struct {
 	ElevenLabsModel   string            `json:"elevenlabs_model"`
 	DebugScreening    bool              `json:"debug_screening"`
 	MCPServers        []MCPServer       `json:"mcp_servers,omitempty"`
-	WebSearch         *WebSearchConfig  `json:"web_search,omitempty"`
 	ConfigFilePath    string            `json:"-"`
 }
 
@@ -158,7 +107,10 @@ func loadAllConfigs(dir string) ([]BotConfig, error) {
 				continue
 			}
 
-			logConfigAdvisories(&config)
+			if config.Thinking == ThinkingModeAdaptive && config.MaxTokens > 0 && config.MaxTokens < 4000 {
+				InfoLogger.Printf("[%s] thinking=adaptive with max_tokens=%d: thinking tokens count toward max_tokens; consider >= 4000",
+					config.ID, config.MaxTokens)
+			}
 
 			config.ConfigFilePath = validPath
 			configs = append(configs, config)
@@ -172,61 +124,6 @@ func loadAllConfigs(dir string) ([]BotConfig, error) {
 	return configs, nil
 }
 
-// logConfigAdvisories emits non-fatal boot-time notes about settings that are
-// valid but likely to surprise: silently ineffective, or costlier than intended.
-func logConfigAdvisories(config *BotConfig) {
-	if config.Temperature != nil {
-		InfoLogger.Printf("[%s] temperature=%g is ignored: the Claude API no longer accepts sampling parameters; remove the key",
-			config.ID, *config.Temperature)
-	}
-
-	if config.Thinking == ThinkingModeAdaptive && config.MaxTokens > 0 && config.MaxTokens < 4000 {
-		InfoLogger.Printf("[%s] thinking=adaptive with max_tokens=%d: thinking tokens count toward max_tokens; consider >= 4000",
-			config.ID, config.MaxTokens)
-	}
-
-	if config.DebounceMs > 0 {
-		InfoLogger.Printf("[%s] intake debounce enabled: coalescing rapid text messages over a %dms quiet window",
-			config.ID, config.DebounceMs)
-	} else {
-		InfoLogger.Printf("[%s] intake debounce disabled: every message dispatches its own turn (set debounce_ms to coalesce rapid follow-ups)",
-			config.ID)
-	}
-
-	if config.Effort != "" {
-		InfoLogger.Printf("[%s] effort=%s: caps reasoning depth and token spend; the API rejects it on models older than Claude 4.6",
-			config.ID, config.Effort)
-	}
-
-	if config.StreamDrafts {
-		InfoLogger.Printf("[%s] draft streaming enabled: private chats get one growing draft with a Stop button; business and group chats keep per-block messages",
-			config.ID)
-	}
-
-	if ws := config.WebSearch; ws != nil && len(ws.AllowedDomains) == 0 && len(ws.BlockedDomains) == 0 {
-		InfoLogger.Printf("[%s] web_search enabled with no allowed_domains/blocked_domains: the model may search the open web",
-			config.ID)
-	}
-
-	if ws := config.WebSearch; ws != nil && !ws.DynamicFilteringEnabled() {
-		InfoLogger.Printf("[%s] web_search dynamic_filtering disabled: results enter context unfiltered (set it true on Claude 4.6+ models to cut tokens)",
-			config.ID)
-	}
-
-	if ws := config.WebSearch; ws != nil && len(ws.FetchAllowedDomains) > 0 {
-		if !ws.Fetch {
-			InfoLogger.Printf("[%s] web_search.fetch_allowed_domains is set but fetch is disabled: it has no effect",
-				config.ID)
-		}
-		for _, d := range ws.FetchAllowedDomains {
-			if strings.Contains(d, "/") {
-				InfoLogger.Printf("[%s] web_search.fetch_allowed_domains entry %q includes a path: web_fetch matches host-only, so the whole host is fetchable",
-					config.ID, d)
-			}
-		}
-	}
-}
-
 func validateConfig(config *BotConfig, ids, tokens map[string]bool) error {
 	if config.ID == "" {
 		return fmt.Errorf("missing 'id' field")
@@ -267,35 +164,10 @@ func validateConfig(config *BotConfig, ids, tokens map[string]bool) error {
 			config.ThinkingDisplay, ThinkingDisplaySummarized, ThinkingDisplayOmitted)
 	}
 
-	if config.Effort != "" && !slices.Contains(validEfforts, config.Effort) {
-		return fmt.Errorf("invalid 'effort' value %q: must be one of %s (or omitted)",
-			config.Effort, strings.Join(validEfforts, ", "))
-	}
-
 	if config.MaxTokens < 0 {
 		return fmt.Errorf("'max_tokens' must be greater than 0 when set")
 	}
 
-	if config.DebounceMs < 0 {
-		return fmt.Errorf("'debounce_ms' must be greater than 0 when set")
-	}
-	if config.DebounceMs > maxDebounceMs {
-		return fmt.Errorf("'debounce_ms' of %d exceeds the maximum of %d (Telegram drops long-idle updates and users read silence as failure)",
-			config.DebounceMs, maxDebounceMs)
-	}
-
-	if ws := config.WebSearch; ws != nil {
-		if len(ws.AllowedDomains) > 0 && len(ws.BlockedDomains) > 0 {
-			return fmt.Errorf("'web_search' cannot set both allowed_domains and blocked_domains (the API rejects that)")
-		}
-		if ws.MaxUses < 0 {
-			return fmt.Errorf("'web_search.max_uses' must be greater than 0 when set")
-		}
-		if ws.MaxContentTokens < 0 {
-			return fmt.Errorf("'web_search.max_content_tokens' must be greater than 0 when set")
-		}
-	}
-
 	if config.MessagePerHour <= 0 {
 		return fmt.Errorf("'messages_per_hour' must be greater than 0")
 	}

config/default.json

@@ -11,10 +11,9 @@
     "messages_per_hour": 20,
     "messages_per_day": 100,
     "temp_ban_duration": "24h",
-    "model": "claude-haiku-5-5",
+    "model": "claude-haiku-4-5",
+    "temperature": 0.7,
     "max_tokens": 1000,
-    "effort": "low",
-    "debounce_ms": 2500,
     "debug_screening": false,
     "system_prompts": {
         "custom_instructions": "You are Atom, a helpful assistant texting through a limited Telegram interface with a 15-word maximum. Write like texting a friend - use shorthand, skip grammar, use slang/abbreviations. The system cuts off anything longer than 15 words.\n\n- Address the user by their first name, and reply in their preferred language (both are in the conversation context).\n- Use time-appropriate greetings based on the user's local time of day.\n- If a user asks about buying apples, inform them that we don't sell apples.\n- When asked for a joke, tell a clean, family-friendly joke about programming or technology.\n- If someone inquires about our services, explain that we offer AI-powered chatbot solutions.\n- For any questions about pricing, direct users to contact our sales team at [email protected].\n- If asked about your capabilities, be honest about what you can and cannot do.\nAlways maintain a friendly and professional tone.",

config_test.go

@@ -875,84 +875,3 @@ func TestThinkingConfigLoad(t *testing.T) {
 		t.Errorf("MaxTokens = %d, want 4096", cfg.MaxTokens)
 	}
 }
-
-func TestWebSearchConfig(t *testing.T) {
-	t.Run("absent leaves nil", func(t *testing.T) {
-		var cfg BotConfig
-		if err := json.Unmarshal([]byte(`{}`), &cfg); err != nil {
-			t.Fatalf("unmarshal: %v", err)
-		}
-		if cfg.WebSearch != nil {
-			t.Errorf("WebSearch = %+v, want nil when absent", cfg.WebSearch)
-		}
-	})
-
-	t.Run("loads allowlist", func(t *testing.T) {
-		jsonData := `{
-			"web_search": {
-				"allowed_domains": ["example.com/hc", "docs.example.com"],
-				"max_uses": 3,
-				"fetch": true,
-				"max_content_tokens": 50000
-			}
-		}`
-		var cfg BotConfig
-		if err := json.Unmarshal([]byte(jsonData), &cfg); err != nil {
-			t.Fatalf("unmarshal: %v", err)
-		}
-		if cfg.WebSearch == nil {
-			t.Fatalf("WebSearch = nil, want populated")
-		}
-		if len(cfg.WebSearch.AllowedDomains) != 2 {
-			t.Errorf("AllowedDomains = %v, want 2 entries", cfg.WebSearch.AllowedDomains)
-		}
-		if cfg.WebSearch.MaxUses != 3 {
-			t.Errorf("MaxUses = %d, want 3", cfg.WebSearch.MaxUses)
-		}
-		if !cfg.WebSearch.Fetch {
-			t.Error("Fetch = false, want true")
-		}
-		if cfg.WebSearch.MaxContentTokens != 50000 {
-			t.Errorf("MaxContentTokens = %d, want 50000", cfg.WebSearch.MaxContentTokens)
-		}
-	})
-
-	cases := []struct {
-		name    string
-		ws      *WebSearchConfig
-		wantErr string
-	}{
-		{"nil ok", nil, ""},
-		{"allowlist ok", &WebSearchConfig{AllowedDomains: []string{"example.com"}, MaxUses: 3}, ""},
-		{"blocklist ok", &WebSearchConfig{BlockedDomains: []string{"evil.com"}}, ""},
-		{"empty ok", &WebSearchConfig{}, ""},
-		{"both allow and block rejected",
-			&WebSearchConfig{AllowedDomains: []string{"a.com"}, BlockedDomains: []string{"b.com"}},
-			"cannot set both allowed_domains and blocked_domains"},
-		{"negative max_uses rejected",
-			&WebSearchConfig{AllowedDomains: []string{"a.com"}, MaxUses: -1},
-			"'web_search.max_uses' must be greater than 0"},
-		{"negative max_content_tokens rejected",
-			&WebSearchConfig{AllowedDomains: []string{"a.com"}, MaxContentTokens: -1},
-			"'web_search.max_content_tokens' must be greater than 0"},
-	}
-	for i, tc := range cases {
-		t.Run(tc.name, func(t *testing.T) {
-			cfg := thinkingTestConfig(fmt.Sprintf("bot-web-%d", i))
-			cfg.WebSearch = tc.ws
-			err := validateConfig(&cfg, map[string]bool{}, map[string]bool{})
-			if tc.wantErr == "" {
-				if err != nil {
-					t.Fatalf("validateConfig(web_search=%+v) = %v, want nil", tc.ws, err)
-				}
-				return
-			}
-			if err == nil {
-				t.Fatalf("validateConfig(web_search=%+v) = nil, want error containing %q", tc.ws, tc.wantErr)
-			}
-			if !contains(err.Error(), tc.wantErr) {
-				t.Errorf("error %q does not contain %q", err.Error(), tc.wantErr)
-			}
-		})
-	}
-}

database.go

@@ -4,7 +4,6 @@ import (
 	"fmt"
 	"log"
 	"os"
-	"strings"
 	"time"
 
 	"gorm.io/driver/sqlite"
@@ -12,27 +11,6 @@ import (
 	"gorm.io/gorm/logger"
 )
 
-// dbLogLevel reads DB_LOG_LEVEL. The default is warn (GORM's own default):
-// errors and slow queries only. "info" logs every statement, which is useful
-// alone at a terminal and, in a deployed bot, writes each user message and
-// reply into the container log twice over.
-func dbLogLevel() logger.LogLevel {
-	raw := strings.ToLower(strings.TrimSpace(os.Getenv("DB_LOG_LEVEL")))
-	switch raw {
-	case "", "warn", "warning":
-		return logger.Warn
-	case "silent":
-		return logger.Silent
-	case "error":
-		return logger.Error
-	case "info":
-		return logger.Info
-	default:
-		InfoLogger.Printf("unknown DB_LOG_LEVEL %q; falling back to warn", raw)
-		return logger.Warn
-	}
-}
-
 func initDB() (*gorm.DB, error) {
 	if err := os.MkdirAll("data", 0750); err != nil {
 		return nil, fmt.Errorf("failed to create data directory: %w", err)
@@ -42,15 +20,8 @@ func initDB() (*gorm.DB, error) {
 		log.New(log.Writer(), "\r\n", log.LstdFlags),
 		logger.Config{
 			SlowThreshold: time.Second,
-			LogLevel:      dbLogLevel(),
+			LogLevel:      logger.Info,
 			Colorful:      false,
-			// A missing row is routine here (the owner probe, a first-time
-			// user), so it must not surface as an error once the level drops
-			// below info.
-			IgnoreRecordNotFoundError: true,
-			// Log placeholders rather than inline values: a slow-query line
-			// would otherwise carry the message text it was writing.
-			ParameterizedQueries: true,
 		},
 	)
 

docker-compose.yml

@@ -13,9 +13,6 @@ services:
     # Optional: Environment variables (can be overridden with .env file)
     # environment:
       # - BOT_LOG_LEVEL=info
-      # GORM SQL logging: silent | error | warn (default) | info.
-      # "info" logs every statement, including message text.
-      # - DB_LOG_LEVEL=warn
     
     # Volume mounts
     volumes:

draft.go

@@ -1,288 +0,0 @@
-package main
-
-import (
-	"context"
-	"strings"
-	"sync"
-	"time"
-
-	"github.com/go-telegram/bot"
-)
-
-const (
-	// draftUpdateInterval throttles sendMessageDraft. Telegram documents no
-	// limit for drafts specifically, so this matches its general guidance of
-	// about one send per second per chat before 429s start.
-	draftUpdateInterval = 1 * time.Second
-	// draftKeepaliveInterval re-sends the current draft during silence, such
-	// as a server-side web search. The preview lives roughly 30 seconds when
-	// untouched; whether updates extend it is undocumented, so this keeps
-	// well inside that either way.
-	draftKeepaliveInterval = 10 * time.Second
-	// telegramMessageLimit is the text cap, in characters, for both drafts
-	// and sent messages.
-	telegramMessageLimit = 4096
-)
-
-// draftEligible reports whether a turn may stream into a Telegram draft.
-// Telegram documents drafts for private chats only, and its reference lists
-// no business_connection_id parameter for sendMessageDraft (the Go library's
-// struct carries the field, but nothing says the server honours it), so
-// business traffic keeps the segment path until that is verified live.
-func (b *Bot) draftEligible(chatID int64, businessConnectionID string) bool {
-	return b.config.StreamDrafts && businessConnectionID == "" && chatID > 0
-}
-
-// draftStream is one growing Telegram draft for one turn.
-//
-// A single sender goroutine owns every sendMessageDraft call, so sends are
-// strictly sequential, never run on the model's stream goroutine, and cannot
-// land at Telegram out of order. The stream goroutine only records the newest
-// text and signals; the sender throttles, keeps the draft alive through silent
-// stretches, and goes quiet after the first failure so a broken draft degrades
-// to a plain final message.
-type draftStream struct {
-	b         *Bot
-	chatID    int64
-	draftID   string
-	interval  time.Duration
-	keepalive time.Duration
-	onFirst   func()
-
-	ctx       context.Context
-	cancel    context.CancelFunc
-	kick      chan struct{}
-	done      chan struct{}
-	closeOnce sync.Once
-
-	mu     sync.Mutex
-	latest string // newest text from the model
-	shown  string // last text Telegram accepted
-	failed bool
-}
-
-// newDraftStream builds the stream; start launches the sender. The two are
-// separate so tests can shorten the intervals before anything runs.
-func (b *Bot) newDraftStream(parent context.Context, chatID int64, draftID string, onFirst func()) *draftStream {
-	ctx, cancel := context.WithCancel(parent)
-	return &draftStream{
-		b:         b,
-		chatID:    chatID,
-		draftID:   draftID,
-		interval:  draftUpdateInterval,
-		keepalive: draftKeepaliveInterval,
-		onFirst:   onFirst,
-		ctx:       ctx,
-		cancel:    cancel,
-		kick:      make(chan struct{}, 1),
-		done:      make(chan struct{}),
-	}
-}
-
-func (d *draftStream) start() { go d.run() }
-
-// update records the full reply text so far and wakes the sender. It never
-// blocks on Telegram.
-func (d *draftStream) update(text string) {
-	d.mu.Lock()
-	d.latest = text
-	d.mu.Unlock()
-	select {
-	case d.kick <- struct{}{}:
-	default:
-	}
-}
-
-// run is the sender loop: wake on new text or on the keepalive deadline,
-// respect the throttle, send the newest text, repeat until closed or failed.
-func (d *draftStream) run() {
-	defer close(d.done)
-
-	var lastSend time.Time
-	for {
-		var keepalive <-chan time.Time
-		if !lastSend.IsZero() {
-			keepalive = time.After(d.keepalive - time.Since(lastSend))
-		}
-		woken := false
-		select {
-		case <-d.ctx.Done():
-			return
-		case <-d.kick:
-		case <-keepalive:
-			woken = true
-		}
-
-		if wait := d.interval - time.Since(lastSend); !lastSend.IsZero() && wait > 0 {
-			select {
-			case <-d.ctx.Done():
-				return
-			case <-time.After(wait):
-			}
-		}
-
-		d.mu.Lock()
-		text, shown := d.latest, d.shown
-		d.mu.Unlock()
-		if text == "" || (text == shown && !woken) {
-			continue
-		}
-		if !d.send(text) {
-			return
-		}
-		lastSend = time.Now()
-	}
-}
-
-// send performs one sendMessageDraft call and reports whether the draft is
-// still usable afterwards.
-func (d *draftStream) send(text string) bool {
-	text = clampRunes(text, telegramMessageLimit)
-	_, err := d.b.tgBot.SendMessageDraft(d.ctx, &bot.SendMessageDraftParams{
-		ChatID:     d.chatID,
-		DraftID:    d.draftID,
-		Text:       text,
-		CanStop:    true,
-		KeepOnStop: true,
-	})
-	if err != nil {
-		d.mu.Lock()
-		d.failed = true
-		d.mu.Unlock()
-		if d.ctx.Err() == nil {
-			// Cosmetic from here on: the final message is still sent.
-			InfoLogger.Printf("[%s] draft %s for chat %d failed, falling back to final message only: %v",
-				d.b.config.ID, d.draftID, d.chatID, err)
-		}
-		return false
-	}
-
-	d.mu.Lock()
-	first := d.shown == ""
-	d.shown = text
-	d.mu.Unlock()
-	if first && d.onFirst != nil {
-		d.onFirst()
-	}
-	return true
-}
-
-// current returns the text the user has actually seen.
-func (d *draftStream) current() string {
-	d.mu.Lock()
-	defer d.mu.Unlock()
-	return d.shown
-}
-
-// close stops the sender and waits for any send in flight, so nothing lands
-// on the chat after the final message. Telegram removes the draft itself once
-// the bot sends a message.
-func (d *draftStream) close() {
-	d.closeOnce.Do(func() {
-		d.cancel()
-		<-d.done
-	})
-}
-
-// replyOutput is the sink a turn writes into. Two implementations: a draft
-// that grows in place and becomes one final message, or the original
-// one-message-per-text-block segments.
-type replyOutput struct {
-	onSegment  func(string) error
-	onProgress func(string)
-	finish     func(ctx context.Context, joined string) error
-	partial    func() string
-	close      func()
-}
-
-func (b *Bot) newReplyOutput(turnCtx context.Context, turn *chatTurn, chatID int64, businessConnectionID string) replyOutput {
-	if b.draftEligible(chatID, businessConnectionID) {
-		draftID := turn.draftIDString()
-		d := b.newDraftStream(turnCtx, chatID, draftID, func() { b.markCommitted(turn) })
-		b.setTurnDraft(turn, draftID)
-		d.start()
-		return replyOutput{
-			onProgress: d.update,
-			finish: func(ctx context.Context, joined string) error {
-				d.close()
-				return b.sendLongMessage(ctx, chatID, joined, businessConnectionID)
-			},
-			partial: d.current,
-			close:   d.close,
-		}
-	}
-
-	return replyOutput{
-		onSegment: func(seg string) error {
-			b.markCommitted(turn)
-			return b.sendOneSegment(turnCtx, chatID, seg, businessConnectionID)
-		},
-		finish:  func(context.Context, string) error { return nil },
-		partial: func() string { return "" },
-		close:   func() {},
-	}
-}
-
-// sendLongMessage delivers text as one message, or as few as the character
-// cap allows, splitting on paragraph and line boundaries before cutting words.
-func (b *Bot) sendLongMessage(ctx context.Context, chatID int64, text, businessConnectionID string) error {
-	for _, chunk := range splitTelegramMessage(text, telegramMessageLimit) {
-		if err := b.sendOneSegment(ctx, chatID, chunk, businessConnectionID); err != nil {
-			return err
-		}
-	}
-	return nil
-}
-
-// splitTelegramMessage splits text into pieces of at most limit runes,
-// preferring paragraph breaks, then line breaks, then spaces.
-func splitTelegramMessage(text string, limit int) []string {
-	text = strings.TrimSpace(text)
-	if text == "" {
-		return nil
-	}
-	var out []string
-	for {
-		off := runeOffset(text, limit)
-		if off < 0 {
-			return append(out, text)
-		}
-		window := text[:off]
-		cut := -1
-		for _, sep := range []string{"\n\n", "\n", " "} {
-			if i := strings.LastIndex(window, sep); i > 0 {
-				cut = i
-				break
-			}
-		}
-		if cut < 0 {
-			cut = off
-		}
-		out = append(out, strings.TrimSpace(text[:cut]))
-		text = strings.TrimSpace(text[cut:])
-		if text == "" {
-			return out
-		}
-	}
-}
-
-// runeOffset returns the byte offset just past the n-th rune of s, or -1 when
-// s holds at most n runes.
-func runeOffset(s string, n int) int {
-	count := 0
-	for i := range s {
-		if count == n {
-			return i
-		}
-		count++
-	}
-	return -1
-}
-
-// clampRunes truncates s to at most n runes.
-func clampRunes(s string, n int) string {
-	if off := runeOffset(s, n); off >= 0 {
-		return s[:off]
-	}
-	return s
-}

draft_test.go

@@ -1,217 +0,0 @@
-package main
-
-import (
-	"context"
-	"errors"
-	"strings"
-	"sync"
-	"testing"
-	"time"
-
-	"github.com/go-telegram/bot"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-type draftRecorder struct {
-	mu    sync.Mutex
-	calls []bot.SendMessageDraftParams
-	err   error
-	delay time.Duration
-}
-
-func (r *draftRecorder) send(_ context.Context, p *bot.SendMessageDraftParams) (bool, error) {
-	time.Sleep(r.delay)
-	r.mu.Lock()
-	defer r.mu.Unlock()
-	r.calls = append(r.calls, *p)
-	return r.err == nil, r.err
-}
-
-func (r *draftRecorder) texts() []string {
-	r.mu.Lock()
-	defer r.mu.Unlock()
-	out := make([]string, len(r.calls))
-	for i, c := range r.calls {
-		out[i] = c.Text
-	}
-	return out
-}
-
-func (r *draftRecorder) count() int { return len(r.texts()) }
-
-const testDraftInterval = 40 * time.Millisecond
-
-func newTestDraft(t *testing.T, rec *draftRecorder, onFirst func()) *draftStream {
-	t.Helper()
-	b, mockTg := setupBotForTest(t, 123)
-	mockTg.SendMessageDraftFunc = rec.send
-	d := b.newDraftStream(context.Background(), 123, "7", onFirst)
-	d.interval = testDraftInterval
-	d.keepalive = time.Hour
-	d.start()
-	t.Cleanup(d.close)
-	return d
-}
-
-func eventuallyCalls(t *testing.T, rec *draftRecorder, n int) {
-	t.Helper()
-	require.Eventually(t, func() bool { return rec.count() == n }, time.Second, 2*time.Millisecond)
-}
-
-func TestDraftStream_FirstUpdateSendsPromptlyThenThrottles(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	rec := &draftRecorder{}
-	d := newTestDraft(t, rec, nil)
-
-	d.update("Hel")
-	eventuallyCalls(t, rec, 1)
-	assert.Equal(t, "Hel", rec.texts()[0], "the first text goes out without waiting for the window")
-
-	d.update("Hello")
-	d.update("Hello, wor")
-	time.Sleep(testDraftInterval / 4)
-	assert.Equal(t, 1, rec.count(), "inside the window nothing is sent")
-
-	eventuallyCalls(t, rec, 2)
-	assert.Equal(t, "Hello, wor", rec.texts()[1], "only the newest text is sent when the window elapses")
-}
-
-func TestDraftStream_SendsAreSequential(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	rec := &draftRecorder{delay: 3 * testDraftInterval}
-	d := newTestDraft(t, rec, nil)
-
-	d.update("a")
-	time.Sleep(testDraftInterval)
-	d.update("ab")
-	time.Sleep(testDraftInterval)
-	d.update("abc")
-
-	eventuallyCalls(t, rec, 2)
-	time.Sleep(2 * testDraftInterval)
-	assert.Equal(t, []string{"a", "abc"}, rec.texts(), "a slow send never overlaps the next; the newest text wins")
-}
-
-func TestDraftStream_ParamsCarryStopButton(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	rec := &draftRecorder{}
-	d := newTestDraft(t, rec, nil)
-
-	d.update("x")
-	eventuallyCalls(t, rec, 1)
-
-	p := rec.calls[0]
-	assert.Equal(t, "7", p.DraftID)
-	assert.Equal(t, int64(123), p.ChatID)
-	assert.True(t, p.CanStop)
-	assert.True(t, p.KeepOnStop)
-}
-
-func TestDraftStream_OnFirstFiresOnce(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	rec := &draftRecorder{}
-	var mu sync.Mutex
-	fired := 0
-	d := newTestDraft(t, rec, func() { mu.Lock(); fired++; mu.Unlock() })
-
-	d.update("a")
-	eventuallyCalls(t, rec, 1)
-	d.update("ab")
-	eventuallyCalls(t, rec, 2)
-
-	mu.Lock()
-	defer mu.Unlock()
-	assert.Equal(t, 1, fired)
-}
-
-func TestDraftStream_FailureStopsFurtherSends(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	rec := &draftRecorder{err: errors.New("TEXTDRAFT_PEER_INVALID")}
-	fired := false
-	d := newTestDraft(t, rec, func() { fired = true })
-
-	d.update("a")
-	eventuallyCalls(t, rec, 1)
-	d.update("ab")
-	time.Sleep(3 * testDraftInterval)
-
-	assert.Equal(t, 1, rec.count(), "after a failure the draft goes quiet")
-	assert.False(t, fired, "a rejected draft never commits the turn")
-	assert.Equal(t, "", d.current(), "nothing was shown to the user")
-}
-
-func TestDraftStream_ClampsToTelegramLimit(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	rec := &draftRecorder{}
-	d := newTestDraft(t, rec, nil)
-
-	d.update(strings.Repeat("é", telegramMessageLimit+500))
-	eventuallyCalls(t, rec, 1)
-
-	assert.Equal(t, telegramMessageLimit, len([]rune(rec.texts()[0])), "a draft over the cap is cut, not rejected")
-	assert.Eventually(t, func() bool { return d.current() != "" }, time.Second, 2*time.Millisecond)
-}
-
-func TestDraftStream_KeepaliveResendsDuringSilence(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	rec := &draftRecorder{}
-	d := newTestDraft(t, rec, nil)
-	d.keepalive = 2 * testDraftInterval
-
-	d.update("thinking")
-	eventuallyCalls(t, rec, 1)
-
-	require.Eventually(t, func() bool { return rec.count() >= 2 }, time.Second, 2*time.Millisecond)
-	assert.Equal(t, "thinking", rec.texts()[1], "silence re-sends the same text to keep the draft alive")
-}
-
-func TestDraftStream_CloseWaitsAndSilences(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	rec := &draftRecorder{delay: 2 * testDraftInterval}
-	d := newTestDraft(t, rec, nil)
-
-	d.update("a")
-	time.Sleep(testDraftInterval / 4) // the send is now in flight
-	d.close()
-
-	n := rec.count()
-	assert.LessOrEqual(t, n, 1)
-	d.update("after close")
-	time.Sleep(3 * testDraftInterval)
-	assert.Equal(t, n, rec.count(), "nothing is sent after close returns")
-}
-
-func TestDraftEligible(t *testing.T) {
-	b, _ := setupBotForTest(t, 123)
-
-	assert.False(t, b.draftEligible(123, ""), "off by default")
-	b.config.StreamDrafts = true
-	assert.True(t, b.draftEligible(123, ""))
-	assert.False(t, b.draftEligible(123, "biz-1"), "business connections keep segments")
-	assert.False(t, b.draftEligible(-100123, ""), "groups keep segments")
-}
-
-func TestSplitTelegramMessage(t *testing.T) {
-	assert.Nil(t, splitTelegramMessage("   ", 10))
-	assert.Equal(t, []string{"short"}, splitTelegramMessage("short", 10))
-
-	parts := splitTelegramMessage("para one\n\npara two\n\npara three", 20)
-	assert.Equal(t, []string{"para one\n\npara two", "para three"}, parts, "prefers paragraph breaks")
-
-	words := splitTelegramMessage("one two three four", 9)
-	assert.Equal(t, []string{"one two", "three", "four"}, words, "falls back to spaces")
-
-	for _, p := range splitTelegramMessage(strings.Repeat("a", 25), 10) {
-		assert.LessOrEqual(t, len([]rune(p)), 10, "hard cut when there is no separator")
-	}
-
-	multibyte := splitTelegramMessage(strings.Repeat("é", 12), 5)
-	assert.Len(t, multibyte, 3, "limits count runes, not bytes")
-}
-
-func TestClampRunes(t *testing.T) {
-	assert.Equal(t, "abc", clampRunes("abc", 5))
-	assert.Equal(t, "abc", clampRunes("abcde", 3))
-	assert.Equal(t, "éé", clampRunes("ééé", 2))
-	assert.Equal(t, "", clampRunes("x", 0))
-}
-
-func TestJoinNonEmpty(t *testing.T) {
-	assert.Equal(t, "", joinNonEmpty("", ""))
-	assert.Equal(t, "a", joinNonEmpty("a", ""))
-	assert.Equal(t, "b", joinNonEmpty("", "b"))
-	assert.Equal(t, "a\n\nb", joinNonEmpty("a", "b"))
-}

edit_stop_test.go

@@ -1,181 +0,0 @@
-package main
-
-import (
-	"context"
-	"testing"
-	"time"
-
-	"github.com/go-telegram/bot"
-	"github.com/go-telegram/bot/models"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-// storeUserMessage puts a user message with a Telegram id into DB and memory,
-// the way screenIncomingMessage does, without going through handleUpdate.
-func storeUserMessage(t *testing.T, b *Bot, chatID int64, telegramID int, text string) Message {
-	t.Helper()
-	msg := b.createMessage(chatID, chatID, "owner", "user", text, true)
-	msg.TelegramMessageID = telegramID
-	require.NoError(t, b.storeMessage(&msg))
-	b.addMessageToChatMemory(b.getOrCreateChatMemory(chatID), msg)
-	return msg
-}
-
-func editUpdate(chatID int64, telegramID int, text string) *models.Update {
-	return &models.Update{EditedMessage: &models.Message{
-		ID: telegramID, Chat: models.Chat{ID: chatID}, From: &models.User{ID: chatID, Username: "owner"}, Text: text,
-	}}
-}
-
-func TestHandleUpdate_EditedMessageUpdatesTextAndRearmsWindow(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	b.config.DebounceMs = int(bufferOnly / time.Millisecond)
-	const chatID int64 = 123
-	from := &models.User{ID: chatID, Username: "owner"}
-
-	b.handleUpdate(context.Background(), nil, &models.Update{Message: &models.Message{
-		ID: 77, Chat: models.Chat{ID: chatID}, From: from, Text: "How to use libtik",
-	}})
-	require.True(t, b.hasPendingIntake(chatID))
-	b.intakeBuffersMu.Lock()
-	seqBefore := b.intakeBuffers[chatID].seq
-	timerBefore := b.intakeBuffers[chatID].timer
-	b.intakeBuffersMu.Unlock()
-
-	b.handleUpdate(context.Background(), nil, editUpdate(chatID, 77, "How to use libtibik"))
-
-	var stored Message
-	require.NoError(t, b.db.Where("chat_id = ? AND telegram_message_id = ?", chatID, 77).First(&stored).Error)
-	assert.Equal(t, "How to use libtibik", stored.Text, "database copy follows the edit")
-
-	mem := b.getOrCreateChatMemory(chatID)
-	assert.Equal(t, "How to use libtibik", mem.Messages[len(mem.Messages)-1].Text, "memory copy follows the edit")
-
-	require.True(t, b.hasPendingIntake(chatID), "an edit must not flush or drop the batch")
-	b.intakeBuffersMu.Lock()
-	defer b.intakeBuffersMu.Unlock()
-	assert.Equal(t, seqBefore, b.intakeBuffers[chatID].seq, "same batch")
-	assert.NotSame(t, timerBefore, b.intakeBuffers[chatID].timer, "the window restarted")
-	assert.Equal(t, 1, b.intakeBuffers[chatID].count, "an edit is not a new message")
-}
-
-func TestHandleUpdate_EditFromEmojiToTextClearsEmojiFlag(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	b.config.DebounceMs = int(bufferOnly / time.Millisecond)
-	const chatID int64 = 123
-
-	b.handleUpdate(context.Background(), nil, &models.Update{Message: &models.Message{
-		ID: 78, Chat: models.Chat{ID: chatID}, From: &models.User{ID: chatID}, Text: "👍",
-	}})
-	b.intakeBuffersMu.Lock()
-	require.True(t, b.intakeBuffers[chatID].allEmojiOnly)
-	b.intakeBuffersMu.Unlock()
-
-	b.handleUpdate(context.Background(), nil, editUpdate(chatID, 78, "can you explain that again?"))
-
-	b.intakeBuffersMu.Lock()
-	defer b.intakeBuffersMu.Unlock()
-	assert.False(t, b.intakeBuffers[chatID].allEmojiOnly, "a real question must not get an emoji-only reply")
-}
-
-func TestHandleUpdate_CaptionEditApplies(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	const chatID int64 = 123
-	storeUserMessage(t, b, chatID, 79, "old caption")
-
-	b.handleUpdate(context.Background(), nil, &models.Update{EditedMessage: &models.Message{
-		ID: 79, Chat: models.Chat{ID: chatID}, From: &models.User{ID: chatID}, Caption: "new caption",
-	}})
-
-	var stored Message
-	require.NoError(t, b.db.Where("chat_id = ? AND telegram_message_id = ?", chatID, 79).First(&stored).Error)
-	assert.Equal(t, "new caption", stored.Text)
-}
-
-func TestHandleUpdate_EditOfAnsweredMessageOnlyCorrectsRecord(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	b.config.DebounceMs = int(bufferOnly / time.Millisecond)
-	const chatID int64 = 123
-	storeUserMessage(t, b, chatID, 80, "first question")
-	_, err := b.screenOutgoingMessage(chatID, "first answer")
-	require.NoError(t, err)
-
-	// A newer, unrelated message is waiting in its window.
-	b.handleUpdate(context.Background(), nil, &models.Update{Message: &models.Message{
-		ID: 81, Chat: models.Chat{ID: chatID}, From: &models.User{ID: chatID}, Text: "second question",
-	}})
-	b.intakeBuffersMu.Lock()
-	timerBefore := b.intakeBuffers[chatID].timer
-	b.intakeBuffersMu.Unlock()
-
-	b.handleUpdate(context.Background(), nil, editUpdate(chatID, 80, "first question, corrected"))
-
-	var stored Message
-	require.NoError(t, b.db.Where("telegram_message_id = ?", 80).First(&stored).Error)
-	assert.Equal(t, "first question, corrected", stored.Text, "the record is corrected")
-	b.intakeBuffersMu.Lock()
-	defer b.intakeBuffersMu.Unlock()
-	assert.Same(t, timerBefore, b.intakeBuffers[chatID].timer, "an answered message's edit must not touch another batch's window")
-}
-
-func TestHandleUpdate_EditRestartsUncommittedTurn(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	const chatID int64 = 123
-	msg := storeUserMessage(t, b, chatID, 82, "how do I use libtik")
-	turn, ctx, ok := b.beginTextTurn(textReq(chatID), msg.ID)
-	require.True(t, ok)
-
-	b.handleUpdate(context.Background(), nil, editUpdate(chatID, 82, "how do I use libtibik"))
-
-	assert.ErrorIs(t, ctx.Err(), context.Canceled, "the model was reading the typo; rerun with the fix")
-	assert.Equal(t, cancelRestart, b.turnReason(turn))
-	next := b.finishTurn(turn)
-	require.NotNil(t, next)
-	assert.Equal(t, chatID, next.chatID)
-}
-
-func TestHandleUpdate_EditOfUnknownMessageIsIgnored(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-
-	b.handleUpdate(context.Background(), nil, editUpdate(123, 999, "never stored"))
-
-	assert.False(t, b.hasPendingIntake(123))
-	var count int64
-	b.db.Model(&Message{}).Where("chat_id = ?", 123).Count(&count)
-	assert.Zero(t, count, "an edit never creates a message")
-}
-
-func TestHandleUpdate_StopGenerationCancelsMatchingTurn(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	turn, ctx, ok := b.beginTextTurn(textReq(123), 1)
-	require.True(t, ok)
-	b.setTurnDraft(turn, turn.draftIDString())
-
-	b.handleUpdate(context.Background(), nil, &models.Update{
-		StoppedMessageGeneration: &models.MessageGenerationStopped{
-			Chat: models.Chat{ID: 123}, DraftID: int(turn.seq),
-		},
-	})
-
-	assert.ErrorIs(t, ctx.Err(), context.Canceled)
-	assert.Equal(t, cancelStop, b.turnReason(turn))
-	b.finishTurn(turn)
-}
-
-func TestClearChatHistory_DiscardsCollectedFlush(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, mockTg := setupBotForTest(t, 123)
-	mockTg.SendMessageFunc = func(_ context.Context, _ *bot.SendMessageParams) (*models.Message, error) {
-		return &models.Message{}, nil
-	}
-	const chatID int64 = 123
-	turn, _, ok := b.beginTextTurn(textReq(chatID), 5)
-	require.True(t, ok)
-	b.markCommitted(turn)
-	_, _, ok = b.beginTextTurn(textReq(chatID), 7)
-	require.False(t, ok, "collected behind the running turn")
-
-	b.clearChatHistory(context.Background(), chatID, chatID, 0, 0, "", false)
-
-	assert.Nil(t, b.finishTurn(turn), "the collected flush must not rerun against the emptied chat")
-}

elevenlabs.go

@@ -56,7 +56,7 @@ func (b *Bot) transcribeVoice(ctx context.Context, fileID string) (string, error
 
 	var buf bytes.Buffer
 	mw := multipart.NewWriter(&buf)
-	if err := mw.WriteField("model_id", "scribe_v2"); err != nil {
+	if err := mw.WriteField("model_id", "scribe_v1"); err != nil {
 		return "", fmt.Errorf("multipart write error: %w", err)
 	}
 	part, err := mw.CreateFormFile("file", "audio.ogg")

go.mod

@@ -1,40 +1,39 @@
 module github.com/HugeFrog24/go-telegram-bot
 
-go 1.27.0
+go 1.26.0
 
 require (
-	github.com/anthropics/anthropic-sdk-go/v2 v2.3.0
-	github.com/go-telegram/bot v1.27.0
-	github.com/stretchr/testify v1.12.1
-	golang.org/x/sync v0.24.0
-	golang.org/x/time v0.16.0
+	github.com/anthropics/anthropic-sdk-go v1.57.0
+	github.com/go-telegram/bot v1.22.0
+	github.com/stretchr/testify v1.11.1
+	golang.org/x/sync v0.22.0
+	golang.org/x/time v0.15.0
 	gorm.io/driver/sqlite v1.6.0
 	gorm.io/gorm v1.31.2
 )
 
 require (
 	github.com/bahlo/generic-list-go v0.2.0 // indirect
-	github.com/buger/jsonparser v1.6.1 // indirect
-	github.com/cespare/xxhash/v2 v2.3.0 // indirect
-	github.com/go-logr/logr v1.4.4 // indirect
-	github.com/go-logr/stdr v1.2.2 // indirect
+	github.com/buger/jsonparser v1.2.0 // indirect
+	github.com/davecgh/go-spew v1.1.1 // indirect
 	github.com/invopop/jsonschema v0.14.0 // indirect
 	github.com/jinzhu/inflection v1.0.0 // indirect
 	github.com/jinzhu/now v1.1.5 // indirect
-	github.com/mattn/go-sqlite3 v1.14.52 // indirect
-	github.com/pb33f/go-yaml v0.1.1 // indirect
-	github.com/pb33f/ordered-map/v2 v2.3.2 // indirect
+	github.com/kr/pretty v0.3.1 // indirect
+	github.com/mailru/easyjson v0.9.2 // indirect
+	github.com/mattn/go-sqlite3 v1.14.48 // indirect
+	github.com/pb33f/ordered-map/v2 v2.3.1 // indirect
+	github.com/pmezard/go-difflib v1.0.0 // indirect
+	github.com/rogpeppe/go-internal v1.14.1 // indirect
 	github.com/standard-webhooks/standard-webhooks/libraries v0.0.1 // indirect
-	github.com/stretchr/objx v0.8.0 // indirect
-	github.com/tidwall/gjson v1.20.0 // indirect
+	github.com/stretchr/objx v0.5.3 // indirect
+	github.com/tidwall/gjson v1.19.0 // indirect
 	github.com/tidwall/match v1.2.0 // indirect
-	github.com/tidwall/pretty v1.2.2 // indirect
+	github.com/tidwall/pretty v1.2.1 // indirect
 	github.com/tidwall/sjson v1.2.5 // indirect
-	go.opentelemetry.io/auto/sdk v1.2.1 // indirect
-	go.opentelemetry.io/otel v1.47.0 // indirect
-	go.opentelemetry.io/otel/log v1.47.0 // indirect
-	go.opentelemetry.io/otel/metric v1.47.0 // indirect
-	go.opentelemetry.io/otel/trace v1.47.0 // indirect
-	go.yaml.in/yaml/v3 v3.0.5 // indirect
-	golang.org/x/text v0.43.0 // indirect
+	github.com/wk8/go-ordered-map/v2 v2.1.8 // indirect
+	go.yaml.in/yaml/v4 v4.0.0-rc.6 // indirect
+	golang.org/x/text v0.40.0 // indirect
+	gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c // indirect
+	gopkg.in/yaml.v3 v3.0.1 // indirect
 )

go.sum

@@ -1,152 +1,100 @@
-cloud.google.com/go/auth v0.7.2 h1:uiha352VrCDMXg+yoBtaD0tUF4Kv9vrtrWPYXwutnDE=
-cloud.google.com/go/auth v0.7.2/go.mod h1:VEc4p5NNxycWQTMQEDQF0bd6aTMb6VgYDXEwiJJQAbs=
-cloud.google.com/go/auth/oauth2adapt v0.2.3 h1:MlxF+Pd3OmSudg/b1yZ5lJwoXCEaeedAguodky1PcKI=
-cloud.google.com/go/auth/oauth2adapt v0.2.3/go.mod h1:tMQXOfZzFuNuUxOypHlQEXgdfX5cuhwU+ffUuXRJE8I=
-cloud.google.com/go/compute/metadata v0.5.0 h1:Zr0eK8JbFv6+Wi4ilXAR8FJ3wyNdpxHKJNPos6LTZOY=
-cloud.google.com/go/compute/metadata v0.5.0/go.mod h1:aHnloV2TPI38yx4s9+wAZhHykWvVCfu7hQbF+9CWoiY=
-github.com/anthropics/anthropic-sdk-go/v2 v2.3.0 h1:wEaN13ToNCcjdqP+47ALC78Jt+DkGMmwqL5zzPQTGkQ=
-github.com/anthropics/anthropic-sdk-go/v2 v2.3.0/go.mod h1:LAOInMKgn6reTIrPvKXa5weyqzxsJkxSBuzhnxzukpQ=
-github.com/aws/aws-sdk-go-v2 v1.38.0 h1:UCRQ5mlqcFk9HJDIqENSLR3wiG1VTWlyUfLDEvY7RxU=
-github.com/aws/aws-sdk-go-v2 v1.38.0/go.mod h1:9Q0OoGQoboYIAJyslFyF1f5K1Ryddop8gqMhWx/n4Wg=
-github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.6.3 h1:tW1/Rkad38LA15X4UQtjXZXNKsCgkshC3EbmcUmghTg=
-github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.6.3/go.mod h1:UbnqO+zjqk3uIt9yCACHJ9IVNhyhOCnYk8yA19SAWrM=
-github.com/aws/aws-sdk-go-v2/config v1.27.27 h1:HdqgGt1OAP0HkEDDShEl0oSYa9ZZBSOmKpdpsDMdO90=
-github.com/aws/aws-sdk-go-v2/config v1.27.27/go.mod h1:MVYamCg76dFNINkZFu4n4RjDixhVr51HLj4ErWzrVwg=
-github.com/aws/aws-sdk-go-v2/credentials v1.17.27 h1:2raNba6gr2IfA0eqqiP2XiQ0UVOpGPgDSi0I9iAP+UI=
-github.com/aws/aws-sdk-go-v2/credentials v1.17.27/go.mod h1:gniiwbGahQByxan6YjQUMcW4Aov6bLC3m+evgcoN4r4=
-github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.16.11 h1:KreluoV8FZDEtI6Co2xuNk/UqI9iwMrOx/87PBNIKqw=
-github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.16.11/go.mod h1:SeSUYBLsMYFoRvHE0Tjvn7kbxaUhl75CJi1sbfhMxkU=
-github.com/aws/aws-sdk-go-v2/internal/configsources v1.3.15 h1:SoNJ4RlFEQEbtDcCEt+QG56MY4fm4W8rYirAmq+/DdU=
-github.com/aws/aws-sdk-go-v2/internal/configsources v1.3.15/go.mod h1:U9ke74k1n2bf+RIgoX1SXFed1HLs51OgUSs+Ph0KJP8=
-github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.6.15 h1:C6WHdGnTDIYETAm5iErQUiVNsclNx9qbJVPIt03B6bI=
-github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.6.15/go.mod h1:ZQLZqhcu+JhSrA9/NXRm8SkDvsycE+JkV3WGY41e+IM=
-github.com/aws/aws-sdk-go-v2/internal/ini v1.8.0 h1:hT8rVHwugYE2lEfdFE0QWVo81lF7jMrYJVDWI+f+VxU=
-github.com/aws/aws-sdk-go-v2/internal/ini v1.8.0/go.mod h1:8tu/lYfQfFe6IGnaOdrpVgEL2IrrDOf6/m9RQum4NkY=
-github.com/aws/aws-sdk-go-v2/service/internal/accept-encoding v1.11.3 h1:dT3MqvGhSoaIhRseqw2I0yH81l7wiR2vjs57O51EAm8=
-github.com/aws/aws-sdk-go-v2/service/internal/accept-encoding v1.11.3/go.mod h1:GlAeCkHwugxdHaueRr4nhPuY+WW+gR8UjlcqzPr1SPI=
-github.com/aws/aws-sdk-go-v2/service/internal/presigned-url v1.11.17 h1:HGErhhrxZlQ044RiM+WdoZxp0p+EGM62y3L6pwA4olE=
-github.com/aws/aws-sdk-go-v2/service/internal/presigned-url v1.11.17/go.mod h1:RkZEx4l0EHYDJpWppMJ3nD9wZJAa8/0lq9aVC+r2UII=
-github.com/aws/aws-sdk-go-v2/service/sso v1.22.4 h1:BXx0ZIxvrJdSgSvKTZ+yRBeSqqgPM89VPlulEcl37tM=
-github.com/aws/aws-sdk-go-v2/service/sso v1.22.4/go.mod h1:ooyCOXjvJEsUw7x+ZDHeISPMhtwI3ZCB7ggFMcFfWLU=
-github.com/aws/aws-sdk-go-v2/service/ssooidc v1.26.4 h1:yiwVzJW2ZxZTurVbYWA7QOrAaCYQR72t0wrSBfoesUE=
-github.com/aws/aws-sdk-go-v2/service/ssooidc v1.26.4/go.mod h1:0oxfLkpz3rQ/CHlx5hB7H69YUpFiI1tql6Q6Ne+1bCw=
-github.com/aws/aws-sdk-go-v2/service/sts v1.30.3 h1:ZsDKRLXGWHk8WdtyYMoGNO7bTudrvuKpDKgMVRlepGE=
-github.com/aws/aws-sdk-go-v2/service/sts v1.30.3/go.mod h1:zwySh8fpFyXp9yOr/KVzxOl8SRqgf/IDw5aUt9UKFcQ=
-github.com/aws/smithy-go v1.22.5 h1:P9ATCXPMb2mPjYBgueqJNCA5S9UfktsW0tTxi+a7eqw=
-github.com/aws/smithy-go v1.22.5/go.mod h1:t1ufH5HMublsJYulve2RKmHDC15xu1f26kHCp/HgceI=
+github.com/anthropics/anthropic-sdk-go v1.45.0 h1:rWnpyBpm9OAm97jyH5bi6W4SRCwJeNY/RyhaJ7CHSUI=
+github.com/anthropics/anthropic-sdk-go v1.45.0/go.mod h1:bx5vWuHFuGPkELH8Z4KUiNSohFnUwScdpTyr+50myPo=
+github.com/anthropics/anthropic-sdk-go v1.52.0 h1:1TB9jt4DN87VMwS/hB1VK26tYzK0ipEOtqPaPGFtJQg=
+github.com/anthropics/anthropic-sdk-go v1.52.0/go.mod h1:3EfIfmFqxH6rbiLcIP4tPFyXL/IHakx2wDG4OU+TIEI=
+github.com/anthropics/anthropic-sdk-go v1.57.0 h1:iEAcPbUKfJ2Iqz9uN/jEndCNW2+x7OYLHDidXDhPjI0=
+github.com/anthropics/anthropic-sdk-go v1.57.0/go.mod h1:3EfIfmFqxH6rbiLcIP4tPFyXL/IHakx2wDG4OU+TIEI=
 github.com/bahlo/generic-list-go v0.2.0 h1:5sz/EEAK+ls5wF+NeqDpk5+iNdMDXrh3z3nPnH1Wvgk=
 github.com/bahlo/generic-list-go v0.2.0/go.mod h1:2KvAjgMlE5NNynlg/5iLrrCCZ2+5xWbdbCW3pNTGyYg=
-github.com/buger/jsonparser v1.6.1 h1:I0phFv0PlbLHnM7TZAVjZ2MJ2/eWRTDyuO7GLR98IEs=
-github.com/buger/jsonparser v1.6.1/go.mod h1:6RYKKt7H4d4+iWqouImQ9R2FZql3VbhNgx27UK13J/0=
-github.com/cespare/xxhash/v2 v2.3.0 h1:UL815xU9SqsFlibzuggzjXhog7bL6oX9BbNZnL2UFvs=
-github.com/cespare/xxhash/v2 v2.3.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
+github.com/buger/jsonparser v1.2.0 h1:4EFcvK1kD4jyj6YqNK6skK6w+y7FHHBR+XBCtxwu/6g=
+github.com/buger/jsonparser v1.2.0/go.mod h1:6RYKKt7H4d4+iWqouImQ9R2FZql3VbhNgx27UK13J/0=
+github.com/creack/pty v1.1.9/go.mod h1:oKZEueFk5CKHvIhNR5MUki03XCEU+Q6VDXinZuGJ33E=
+github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
+github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/dnaeon/go-vcr v1.2.0 h1:zHCHvJYTMh1N7xnV7zf1m1GPBF9Ad0Jk/whtQ1663qI=
 github.com/dnaeon/go-vcr v1.2.0/go.mod h1:R4UdLID7HZT3taECzJs4YgbbH6PIGXB6W/sc5OLb6RQ=
-github.com/felixge/httpsnoop v1.0.4 h1:NFTV2Zj1bL4mc9sqWACXbQFVBBg2W3GPvqp8/ESS2Wg=
-github.com/felixge/httpsnoop v1.0.4/go.mod h1:m8KPJKqk1gH5J9DgRY2ASl2lWCfGKXixSwevea8zH2U=
-github.com/go-logr/logr v1.2.2/go.mod h1:jdQByPbusPIv2/zmleS9BjJVeZ6kBagPoEUsqbVz/1A=
-github.com/go-logr/logr v1.4.4 h1:tG4xh9yMsRCAiodLVTxyrkzSZ9+o0L1Kg/+cPVcbP/8=
-github.com/go-logr/logr v1.4.4/go.mod h1:9T104GzyrTigFIr8wt5mBrctHMim0Nb2HLGrmQ40KvY=
-github.com/go-logr/stdr v1.2.2 h1:hSWxHoqTgW2S2qGc0LTAI563KZ5YKYRhT3MFKZMbjag=
-github.com/go-logr/stdr v1.2.2/go.mod h1:mMo/vtBO5dYbehREoey6XUKy/eSumjCCveDpRre4VKE=
-github.com/go-telegram/bot v1.27.0 h1:PG9yBDzM8O26oUerE9SZwWlHon4RpH8WcNslZ7CEZmE=
-github.com/go-telegram/bot v1.27.0/go.mod h1:i2TRs7fXWIeaceF3z7KzsMt/he0TwkVC680mvdTFYeM=
-github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da h1:oI5xCqsCo564l8iNU+DwB5epxmsaqB+rhGL0m5jtYqE=
-github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
-github.com/golang/protobuf v1.5.4 h1:i7eJL8qZTpSEXOPTxNKhASYpMn+8e5Q6AdndVa1dWek=
-github.com/golang/protobuf v1.5.4/go.mod h1:lnTiLA8Wa4RWRcIUkrtSVa5nRhsEGBg48fD6rSs7xps=
-github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
-github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
-github.com/google/jsonschema-go v0.4.2 h1:tmrUohrwoLZZS/P3x7ex0WAVknEkBZM46iALbcqoRA8=
-github.com/google/jsonschema-go v0.4.2/go.mod h1:r5quNTdLOYEz95Ru18zA0ydNbBuYoo9tgaYcxEYhJVE=
-github.com/google/s2a-go v0.1.7 h1:60BLSyTrOV4/haCDW4zb1guZItoSq8foHCXrAnjBo/o=
-github.com/google/s2a-go v0.1.7/go.mod h1:50CgR4k1jNlWBu4UfS4AcfhVe1r6pdZPygJ3R8F0Qdw=
-github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
-github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
-github.com/googleapis/enterprise-certificate-proxy v0.3.2 h1:Vie5ybvEvT75RniqhfFxPRy3Bf7vr3h0cechB90XaQs=
-github.com/googleapis/enterprise-certificate-proxy v0.3.2/go.mod h1:VLSiSSBs/ksPL8kq3OBOQ6WRI2QnaFynd1DCjZ62+V0=
+github.com/go-telegram/bot v1.20.0 h1:4Pea/qTidSspr4WBJw9FbHUMNhYeqszBqQUfsQEyFbc=
+github.com/go-telegram/bot v1.20.0/go.mod h1:i2TRs7fXWIeaceF3z7KzsMt/he0TwkVC680mvdTFYeM=
+github.com/go-telegram/bot v1.21.0 h1:Va/PbGc2vBDdv57GCUEEVV6ROlHWiC6SklJY9Hvhzps=
+github.com/go-telegram/bot v1.21.0/go.mod h1:i2TRs7fXWIeaceF3z7KzsMt/he0TwkVC680mvdTFYeM=
+github.com/go-telegram/bot v1.22.0 h1:zK29OoTYMmR5emJrCtGa2SjaGleeZiUB/C1i7kc2lXE=
+github.com/go-telegram/bot v1.22.0/go.mod h1:i2TRs7fXWIeaceF3z7KzsMt/he0TwkVC680mvdTFYeM=
+github.com/invopop/jsonschema v0.13.0 h1:KvpoAJWEjR3uD9Kbm2HWJmqsEaHt8lBUpd0qHcIi21E=
+github.com/invopop/jsonschema v0.13.0/go.mod h1:ffZ5Km5SWWRAIN6wbDXItl95euhFz2uON45H2qjYt+0=
 github.com/invopop/jsonschema v0.14.0 h1:MHQqLhvpNUZfw+hM3AZDYK7jxO8FZoQeQM77g8iyZjg=
 github.com/invopop/jsonschema v0.14.0/go.mod h1:ygm6C2EaVNMBDPpaPlnOA2pFAxBnxGjFlMZABxm9n2I=
 github.com/jinzhu/inflection v1.0.0 h1:K317FqzuhWc8YvSVlFMCCUb36O/S9MCKRDI7QkRKD/E=
 github.com/jinzhu/inflection v1.0.0/go.mod h1:h+uFLlag+Qp1Va5pdKtLDYj+kHp5pxUVkryuEj+Srlc=
 github.com/jinzhu/now v1.1.5 h1:/o9tlHleP7gOFmsnYNz3RGnqzefHA47wQpKrrdTIwXQ=
 github.com/jinzhu/now v1.1.5/go.mod h1:d3SSVoowX0Lcu0IBviAWJpolVfI5UJVZZ7cO71lE/z8=
-github.com/mattn/go-sqlite3 v1.14.52 h1:wVbm2Qnf4OXkqhBTSPuCRZDRnxfbVrrmiCEroVdog8U=
-github.com/mattn/go-sqlite3 v1.14.52/go.mod h1:6JTjA44L93a0QCyJef5YvlPoKXntQPjzWv5gtm9sB6w=
-github.com/modelcontextprotocol/go-sdk v1.3.1 h1:TfqtNKOIWN4Z1oqmPAiWDC2Jq7K9OdJaooe0teoXASI=
-github.com/modelcontextprotocol/go-sdk v1.3.1/go.mod h1:DgVX498dMD8UJlseK1S5i1T4tFz2fkBk4xogC3D15nw=
-github.com/pb33f/go-yaml v0.1.1 h1:yWGmhVdwzionRnSc/hEJwcfq6PaZL2QdYzyReXBbX7Q=
-github.com/pb33f/go-yaml v0.1.1/go.mod h1:QvWdkHP5VjFYAeHMQUkaPQqREaQnBNOXmWzAjNrnSdo=
-github.com/pb33f/ordered-map/v2 v2.3.2 h1:wDyaZ2Pv9QLh64X4utCeD5Zoi9nIv2aW3lwv172O5PQ=
-github.com/pb33f/ordered-map/v2 v2.3.2/go.mod h1:1OhFrXu3OYw3kM+FXF+Ug3ugmmZ9LE8z0JfQMLvtV0w=
-github.com/segmentio/asm v1.1.3 h1:WM03sfUOENvvKexOLp+pCqgb/WDjsi7EK8gIsICtzhc=
-github.com/segmentio/asm v1.1.3/go.mod h1:Ld3L4ZXGNcSLRg4JBsZ3//1+f/TjYl0Mzen/DQy1EJg=
-github.com/segmentio/encoding v0.5.4 h1:OW1VRern8Nw6ITAtwSZ7Idrl3MXCFwXHPgqESYfvNt0=
-github.com/segmentio/encoding v0.5.4/go.mod h1:HS1ZKa3kSN32ZHVZ7ZLPLXWvOVIiZtyJnO1gPH1sKt0=
+github.com/kr/pretty v0.2.1/go.mod h1:ipq/a2n7PKx3OHsz4KJII5eveXtPO4qwEXGdVfWzfnI=
+github.com/kr/pretty v0.3.1 h1:flRD4NNwYAUpkphVc1HcthR4KEIFJ65n8Mw5qdRn3LE=
+github.com/kr/pretty v0.3.1/go.mod h1:hoEshYVHaxMs3cyo3Yncou5ZscifuDolrwPKZanG3xk=
+github.com/kr/pty v1.1.1/go.mod h1:pFQYn66WHrOpPYNljwOMqo10TkYh1fy3cYio2l3bCsQ=
+github.com/kr/text v0.1.0/go.mod h1:4Jbv+DJW3UT/LiOwJeYQe1efqtUx/iVham/4vfdArNI=
+github.com/kr/text v0.2.0 h1:5Nx0Ya0ZqY2ygV366QzturHI13Jq95ApcVaJBhpS+AY=
+github.com/kr/text v0.2.0/go.mod h1:eLer722TekiGuMkidMxC/pM04lWEeraHUUmBw8l2grE=
+github.com/mailru/easyjson v0.9.2 h1:dX8U45hQsZpxd80nLvDGihsQ/OxlvTkVUXH2r/8cb2M=
+github.com/mailru/easyjson v0.9.2/go.mod h1:1+xMtQp2MRNVL/V1bOzuP3aP8VNwRW55fQUto+XFtTU=
+github.com/mattn/go-sqlite3 v1.14.44 h1:3VSe+xafpbzsLbdr2AWlAZk9yRHiBhTBakioXaCKTF8=
+github.com/mattn/go-sqlite3 v1.14.44/go.mod h1:pjEuOr8IwzLJP2MfGeTb0A35jauH+C2kbHKBr7yXKVQ=
+github.com/mattn/go-sqlite3 v1.14.47 h1:jOBI62gS7nKeZv+as1oGEy0+1qISgXwH/QBlR6KbfIo=
+github.com/mattn/go-sqlite3 v1.14.47/go.mod h1:6JTjA44L93a0QCyJef5YvlPoKXntQPjzWv5gtm9sB6w=
+github.com/mattn/go-sqlite3 v1.14.48 h1:7XHIgl0a8HwOaiK4E47ozLkST78rR9+OtNGx27D/TFs=
+github.com/mattn/go-sqlite3 v1.14.48/go.mod h1:6JTjA44L93a0QCyJef5YvlPoKXntQPjzWv5gtm9sB6w=
+github.com/pb33f/ordered-map/v2 v2.3.1 h1:5319HDO0aw4DA4gzi+zv4FXU9UlSs3xGZ40wcP1nBjY=
+github.com/pb33f/ordered-map/v2 v2.3.1/go.mod h1:qxFQgd0PkVUtOMCkTapqotNgzRhMPL7VvaHKbd1HnmQ=
+github.com/pkg/diff v0.0.0-20210226163009-20ebb0f2a09e/go.mod h1:pJLUxLENpZxwdsKMEsNbx1VGcRFpLqf3715MtcvvzbA=
+github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
+github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
+github.com/rogpeppe/go-internal v1.9.0/go.mod h1:WtVeX8xhTBvf0smdhujwtBcq4Qrzq/fJaraNFVN+nFs=
+github.com/rogpeppe/go-internal v1.14.1 h1:UQB4HGPB6osV0SQTLymcB4TgvyWu6ZyliaW0tI/otEQ=
+github.com/rogpeppe/go-internal v1.14.1/go.mod h1:MaRKkUm5W0goXpeCfT7UZI6fk/L7L7so1lCWt35ZSgc=
 github.com/standard-webhooks/standard-webhooks/libraries v0.0.1 h1:uOfcYT+3QungH6tIGSVCR/Y3KJmgJiHcojJbMTPDZAI=
 github.com/standard-webhooks/standard-webhooks/libraries v0.0.1/go.mod h1:L1MQhA6x4dn9r007T033lsaZMv9EmBAdXyU/+EF40fo=
-github.com/stretchr/objx v0.8.0 h1:rm9oMkeprG/wHifkZP4EuDww9S0mvEaQ/p9cXTNT4cw=
-github.com/stretchr/objx v0.8.0/go.mod h1:xSzHnPHkdEv7vtgdQej8atlCuU1mdX+ELge8bn0hCsE=
-github.com/stretchr/testify v1.12.1 h1:EuwCh5fleGS7H32xRwO3wRGT7DxrDhLAT6FF8MpWDWE=
-github.com/stretchr/testify v1.12.1/go.mod h1:MDEgiDPPsNp5cuIrHPPCyornHKgEVbtFUmoNlxoYthg=
+github.com/stretchr/objx v0.5.3 h1:jmXUvGomnU1o3W/V5h2VEradbpJDwGrzugQQvL0POH4=
+github.com/stretchr/objx v0.5.3/go.mod h1:rDQraq+vQZU7Fde9LOZLr8Tax6zZvy4kuNKF+QYS+U0=
+github.com/stretchr/testify v1.11.1 h1:7s2iGBzp5EwR7/aIZr8ao5+dra3wiQyKjjFuvgVKu7U=
+github.com/stretchr/testify v1.11.1/go.mod h1:wZwfW3scLgRK+23gO65QZefKpKQRnfz6sD981Nm4B6U=
 github.com/tidwall/gjson v1.14.2/go.mod h1:/wbyibRr2FHMks5tjHJ5F8dMZh3AcwJEMf5vlfC0lxk=
-github.com/tidwall/gjson v1.20.0 h1:+agJ3rEzKcCXKDKo0ml26UROcpcAlnZyjq+TjbY5Fto=
-github.com/tidwall/gjson v1.20.0/go.mod h1:V37/opeE/JbLUOfH0QTXiNez2l0RUjYUhpT4szFQAfc=
+github.com/tidwall/gjson v1.19.0 h1:xwxm7n691Uf3u5OFjzngavjGTh55KX5q/9w9xHW88JU=
+github.com/tidwall/gjson v1.19.0/go.mod h1:V37/opeE/JbLUOfH0QTXiNez2l0RUjYUhpT4szFQAfc=
 github.com/tidwall/match v1.1.1/go.mod h1:eRSPERbgtNPcGhD8UCthc6PmLEQXEWd3PRB5JTxsfmM=
 github.com/tidwall/match v1.2.0 h1:0pt8FlkOwjN2fPt4bIl4BoNxb98gGHN2ObFEDkrfZnM=
 github.com/tidwall/match v1.2.0/go.mod h1:eRSPERbgtNPcGhD8UCthc6PmLEQXEWd3PRB5JTxsfmM=
 github.com/tidwall/pretty v1.2.0/go.mod h1:ITEVvHYasfjBbM0u2Pg8T2nJnzm8xPwvNhhsoaGGjNU=
-github.com/tidwall/pretty v1.2.2 h1:dz1jrRuE7or/74V490B4/GP1pZm5WKlt2bgCP5A83w8=
-github.com/tidwall/pretty v1.2.2/go.mod h1:ITEVvHYasfjBbM0u2Pg8T2nJnzm8xPwvNhhsoaGGjNU=
+github.com/tidwall/pretty v1.2.1 h1:qjsOFOWWQl+N3RsoF5/ssm1pHmJJwhjlSbZ51I6wMl4=
+github.com/tidwall/pretty v1.2.1/go.mod h1:ITEVvHYasfjBbM0u2Pg8T2nJnzm8xPwvNhhsoaGGjNU=
 github.com/tidwall/sjson v1.2.5 h1:kLy8mja+1c9jlljvWTlSazM7cKDRfJuR/bOJhcY5NcY=
 github.com/tidwall/sjson v1.2.5/go.mod h1:Fvgq9kS/6ociJEDnK0Fk1cpYF4FIW6ZF7LAe+6jwd28=
-github.com/yosida95/uritemplate/v3 v3.0.2 h1:Ed3Oyj9yrmi9087+NczuL5BwkIc4wvTb5zIM+UJPGz4=
-github.com/yosida95/uritemplate/v3 v3.0.2/go.mod h1:ILOh0sOhIJR3+L/8afwt/kE++YT040gmv5BQTMR2HP4=
-go.opencensus.io v0.24.0 h1:y73uSU6J157QMP2kn2r30vwW1A2W2WFwSCGnAVxeaD0=
-go.opencensus.io v0.24.0/go.mod h1:vNK8G9p7aAivkbmorf4v+7Hgx+Zs0yY+0fOtgBfjQKo=
-go.opentelemetry.io/auto/sdk v1.2.1 h1:jXsnJ4Lmnqd11kwkBV2LgLoFMZKizbCi5fNZ/ipaZ64=
-go.opentelemetry.io/auto/sdk v1.2.1/go.mod h1:KRTj+aOaElaLi+wW1kO/DZRXwkF4C5xPbEe3ZiIhN7Y=
-go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.49.0 h1:4Pp6oUg3+e/6M4C0A/3kJ2VYa++dsWVTtGgLVj5xtHg=
-go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.49.0/go.mod h1:Mjt1i1INqiaoZOMGR1RIUJN+i3ChKoFRqzrRQhlkbs0=
-go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0 h1:jq9TW8u3so/bN+JPT166wjOI6/vQPF6Xe7nMNIltagk=
-go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0/go.mod h1:p8pYQP+m5XfbZm9fxtSKAbM6oIllS7s2AfxrChvc7iw=
-go.opentelemetry.io/otel v1.47.0 h1:j7ALJ/zgkS7Z6aeJW09p8VC9804bC+PpeTfCD4XPnOM=
-go.opentelemetry.io/otel v1.47.0/go.mod h1:8wS9O2qfXrYrzp6hIF/HOYJJf/wIhFPhR2xLuP+iXQU=
-go.opentelemetry.io/otel/log v1.47.0 h1:cOTS1CcLbSQeZKanGJ+0JpF/+t4PELi3O3bbl2lqCcI=
-go.opentelemetry.io/otel/log v1.47.0/go.mod h1:9byitSQ5pLC6PpqwGXjqdMKya6ZTswHRZh2vvXT33nw=
-go.opentelemetry.io/otel/metric v1.47.0 h1:4PptaldXx3Eat1XjMZ68pPJEs5wrhlemctZE9a3UdWY=
-go.opentelemetry.io/otel/metric v1.47.0/go.mod h1:ADGSXxRrXM6bjbvLo535EstVFlPpPYZm4LBKixjDHwU=
-go.opentelemetry.io/otel/sdk v1.33.0 h1:iax7M131HuAm9QkZotNHEfstof92xM+N8sr3uHXc2IM=
-go.opentelemetry.io/otel/sdk v1.33.0/go.mod h1:A1Q5oi7/9XaMlIWzPSxLRWOI8nG3FnzHJNbiENQuihM=
-go.opentelemetry.io/otel/trace v1.47.0 h1:JOjX/Oci8K94QHddo+bbfya/Ai/nf6/dt9ZfrFNWSrM=
-go.opentelemetry.io/otel/trace v1.47.0/go.mod h1:jNaSLa2PZEYFG6fRjJABAu+bw4FS08uDmPg28lTghu0=
-go.yaml.in/yaml/v3 v3.0.5 h1:N6y/pJk8buWs9NY5ERU2HSMfm+IuD/OtfdAnq6kESPw=
-go.yaml.in/yaml/v3 v3.0.5/go.mod h1:HVTZu1O7/Vkt2N+BFy8Zza+lnLsABggaTM2ZpNIGuKg=
-golang.org/x/crypto v0.40.0 h1:r4x+VvoG5Fm+eJcxMaY8CQM7Lb0l1lsmjGBQ6s8BfKM=
-golang.org/x/crypto v0.40.0/go.mod h1:Qr1vMER5WyS2dfPHAlsOj01wgLbsyWtFn/aY+5+ZdxY=
-golang.org/x/net v0.41.0 h1:vBTly1HeNPEn3wtREYfy4GZ/NECgw2Cnl+nK6Nz3uvw=
-golang.org/x/net v0.41.0/go.mod h1:B/K4NNqkfmg07DQYrbwvSluqCJOOXwUjeb/5lOisjbA=
-golang.org/x/oauth2 v0.30.0 h1:dnDm7JmhM45NNpd8FDDeLhK6FwqbOf4MLCM9zb1BOHI=
-golang.org/x/oauth2 v0.30.0/go.mod h1:B++QgG3ZKulg6sRPGD/mqlHQs5rB3Ml9erfeDY7xKlU=
-golang.org/x/sync v0.24.0 h1:iqiDrGoDgyuvAKrhLQthUEzU/R378N78gXbBjxed7oA=
-golang.org/x/sync v0.24.0/go.mod h1:sUUOizhqBxiL6pEWpqNLUiaJn1ShEbZ6BBqskPbjZm0=
-golang.org/x/sys v0.35.0 h1:vz1N37gP5bs89s7He8XuIYXpyY0+QlsKmzipCbUtyxI=
-golang.org/x/sys v0.35.0/go.mod h1:BJP2sWEmIv4KK5OTEluFJCKSidICx8ciO85XgH3Ak8k=
-golang.org/x/text v0.43.0 h1:1QivrlhwAsnMOcqgOdl9my7s9OyZl8O5/UGUTJVXoWc=
-golang.org/x/text v0.43.0/go.mod h1:hbSIYA/amXcRXTFSZqh8tWHEoqRvvbKPJ2uz/KU+a+s=
-golang.org/x/time v0.16.0 h1:vMb6ptszcQMkcwiRTAuNNU50gom6++Q/6gY2hDM6VDE=
-golang.org/x/time v0.16.0/go.mod h1:rVKOqvZeKvrDKTQiAHJ7wmwP0RzleSphoEA9RcdLA0s=
-google.golang.org/api v0.189.0 h1:equMo30LypAkdkLMBqfeIqtyAnlyig1JSZArl4XPwdI=
-google.golang.org/api v0.189.0/go.mod h1:FLWGJKb0hb+pU2j+rJqwbnsF+ym+fQs73rbJ+KAUgy8=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20240722135656-d784300faade h1:oCRSWfwGXQsqlVdErcyTt4A93Y8fo0/9D4b1gnI++qo=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20240722135656-d784300faade/go.mod h1:Ue6ibwXGpU+dqIcODieyLOcgj7z8+IcskoNIgZxtrFY=
-google.golang.org/grpc v1.64.1 h1:LKtvyfbX3UGVPFcGqJ9ItpVWW6oN/2XqTxfAnwRRXiA=
-google.golang.org/grpc v1.64.1/go.mod h1:hiQF4LFZelK2WKaP6W0L92zGHtiQdZxk8CrSdvyjeP0=
-google.golang.org/protobuf v1.34.2 h1:6xV6lTsCfpGD21XK49h7MhtcApnLqkfYgPcdHftf6hg=
-google.golang.org/protobuf v1.34.2/go.mod h1:qYOHts0dSfpeUzUFpOMr/WGzszTmLH+DiWniOlNbLDw=
+github.com/wk8/go-ordered-map/v2 v2.1.8 h1:5h/BUHu93oj4gIdvHHHGsScSTMijfx5PeYkE/fJgbpc=
+github.com/wk8/go-ordered-map/v2 v2.1.8/go.mod h1:5nJHM5DyteebpVlHnWMV0rPz6Zp7+xBAnxjb1X5vnTw=
+go.yaml.in/yaml/v4 v4.0.0-rc.6 h1:1h7H1ohdUh93/FyE4YaDa1Zh64K6VVbjF4K6WUxMtH4=
+go.yaml.in/yaml/v4 v4.0.0-rc.6/go.mod h1:aZqd9kCMsGL7AuUv/m/PvWLdg5sjJsZ4oHDEnfPPfY0=
+golang.org/x/sync v0.20.0 h1:e0PTpb7pjO8GAtTs2dQ6jYa5BWYlMuX047Dco/pItO4=
+golang.org/x/sync v0.20.0/go.mod h1:9xrNwdLfx4jkKbNva9FpL6vEN7evnE43NNNJQ2LF3+0=
+golang.org/x/sync v0.21.0 h1:HLII4xRRTtCRkxYp4HNFF0Js/Og6q2i++KXbg0gHCwM=
+golang.org/x/sync v0.21.0/go.mod h1:9xrNwdLfx4jkKbNva9FpL6vEN7evnE43NNNJQ2LF3+0=
+golang.org/x/sync v0.22.0 h1:SZjpbeLmrCk4xhRSZFNZW5gFUeCeFgjekvI/+gfScek=
+golang.org/x/sync v0.22.0/go.mod h1:9xrNwdLfx4jkKbNva9FpL6vEN7evnE43NNNJQ2LF3+0=
+golang.org/x/text v0.37.0 h1:Cqjiwd9eSg8e0QAkyCaQTNHFIIzWtidPahFWR83rTrc=
+golang.org/x/text v0.37.0/go.mod h1:a5sjxXGs9hsn/AJVwuElvCAo9v8QYLzvavO5z2PiM38=
+golang.org/x/text v0.38.0 h1:sXmwo9DwP3OK9EZ7PqAdaooSGozfl/3a6/xJcbzPRhE=
+golang.org/x/text v0.38.0/go.mod h1:YXZt3QhHUKYT53r2lLKFIVi6Ao1jdzrTR/KQ09qyxF4=
+golang.org/x/text v0.40.0 h1:Ub2Z6/xjgF1WrYQz2nuITOEegKFtiIy+rieRJ5lHZKs=
+golang.org/x/text v0.40.0/go.mod h1:hpnzDAfGV753zIKo+wk3u1bVKCGPbrnF7+7LBF/UHVY=
+golang.org/x/time v0.15.0 h1:bbrp8t3bGUeFOx08pvsMYRTCVSMk89u4tKbNOZbp88U=
+golang.org/x/time v0.15.0/go.mod h1:Y4YMaQmXwGQZoFaVFk4YpCt4FLQMYKZe9oeV/f4MSno=
+gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
+gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=
 gopkg.in/yaml.v2 v2.2.8 h1:obN1ZagJSUGI0Ek/LBmuj4SNLPfIny3KsKFopxRdj10=
 gopkg.in/yaml.v2 v2.2.8/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
+gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
 gorm.io/driver/sqlite v1.6.0 h1:WHRRrIiulaPiPFmDcod6prc4l2VGVWHz80KspNsxSfQ=
 gorm.io/driver/sqlite v1.6.0/go.mod h1:AO9V1qIQddBESngQUKWL9yoH93HIeA1X6V633rBwyT8=
+gorm.io/gorm v1.31.1 h1:7CA8FTFz/gRfgqgpeKIBcervUn3xSyPUmr6B2WXJ7kg=
+gorm.io/gorm v1.31.1/go.mod h1:XyQVbO2k6YkOis7C2437jSit3SsDK72s7n7rsSHd+Gs=
 gorm.io/gorm v1.31.2 h1:3o8FXNo9v9S858gil+3LlZA1LkCOzgb4g5BL64FgaCo=
 gorm.io/gorm v1.31.2/go.mod h1:XyQVbO2k6YkOis7C2437jSit3SsDK72s7n7rsSHd+Gs=

handlers.go

@@ -7,7 +7,7 @@ import (
 	"strconv"
 	"strings"
 
-	"github.com/anthropics/anthropic-sdk-go/v2"
+	"github.com/anthropics/anthropic-sdk-go"
 	"github.com/go-telegram/bot"
 	"github.com/go-telegram/bot/models"
 	"golang.org/x/sync/errgroup"
@@ -28,15 +28,6 @@ func (b *Bot) handleVoiceMessage(ctx context.Context, message *models.Message, u
 		return
 	}
 
-	// Voice is a committed turn: it waits for any running turn and is never
-	// restarted, so a text sent right after a voice note is answered afterwards.
-	turn, turnCtx := b.beginTurn(ctx, chatID)
-	defer b.endTurn(turn)
-	ctx = turnCtx
-
-	stopTyping := b.startChatAction(ctx, chatID, businessConnectionID, models.ChatActionTyping)
-	defer stopTyping()
-
 	transcript, err := b.transcribeVoice(ctx, message.Voice.FileID)
 	if err != nil {
 		ErrorLogger.Printf("Error transcribing voice message from user %d: %v", userID, err)
@@ -60,8 +51,9 @@ func (b *Bot) handleVoiceMessage(ctx context.Context, message *models.Message, u
 	}
 	b.chatMemoriesMu.Unlock()
 
-	contextMessages := b.snapshotForTurn(turn, b.getOrCreateChatMemory(chatID))
-	response, err := b.getAnthropicResponse(ctx, chatID, contextMessages, false, username, firstName, lastName, isPremium, languageCode, messageTime, nil, nil)
+	chatMemory := b.getOrCreateChatMemory(chatID)
+	contextMessages := b.prepareContextMessages(chatMemory)
+	response, err := b.getAnthropicResponse(ctx, chatID, contextMessages, false, username, firstName, lastName, isPremium, languageCode, messageTime, nil)
 	if err != nil {
 		ErrorLogger.Printf("Error getting Anthropic response for voice: %v", err)
 		if err := b.sendResponse(ctx, chatID, b.anthropicErrorResponse(err, userID), businessConnectionID); err != nil {
@@ -70,12 +62,6 @@ func (b *Bot) handleVoiceMessage(ctx context.Context, message *models.Message, u
 		return
 	}
 
-	// Switch the indicator once the model is done and synthesis begins, so the
-	// client shows "recording audio" rather than "typing" for a voice reply.
-	stopTyping()
-	stopRecording := b.startChatAction(ctx, chatID, businessConnectionID, models.ChatActionUploadVoice)
-	defer stopRecording()
-
 	audioReader, err := b.generateSpeech(ctx, response)
 	if err != nil {
 		ErrorLogger.Printf("Error generating speech, falling back to text: %v", err)
@@ -125,53 +111,6 @@ func (b *Bot) handlePhotoMessage(
 		return
 	}
 
-	// The Files API uploads depend on nothing the chat's running turn is doing,
-	// so they overlap it instead of queueing behind it. On an album this is the
-	// longest wait in the bot, and the typing indicator covers it.
-	stopTyping := b.startChatAction(ctx, chatID, businessConnectionID, models.ChatActionTyping)
-	finalUploaded, caption, err := b.uploadPhotoItems(ctx, items, chatID, businessConnectionID)
-	stopTyping()
-	if err != nil || len(finalUploaded) == 0 {
-		return
-	}
-
-	req := turnRequest{
-		ctx: ctx, chatID: chatID, userID: userID,
-		username: username, firstName: firstName, lastName: lastName,
-		languageCode: languageCode, isPremium: isPremium, messageTime: messageTime,
-		businessConnectionID: businessConnectionID,
-	}
-	// Photos wait for any running turn and are never restarted.
-	turn, turnCtx := b.beginTurn(ctx, chatID)
-
-	b.runTurn(turn, turnCtx, req, func(ctx context.Context, out replyOutput) (string, error) {
-		chatMemory := b.getOrCreateChatMemory(chatID)
-		userMessage := b.createMessage(chatID, userID, username, "user", caption, true)
-		userMessage.ImageFileIDs = finalUploaded
-		userMessage.TelegramMessageID = items[0].ID
-		if err := b.storeMessage(&userMessage); err != nil {
-			b.compensatingDelete(ctx, finalUploaded)
-			ErrorLogger.Printf("[%s] store photo message failed: %v", b.config.ID, err)
-			if sendErr := b.sendResponse(ctx, chatID, "Sorry, I had trouble saving your message.", businessConnectionID); sendErr != nil {
-				ErrorLogger.Printf("Error sending store failure message: %v", sendErr)
-			}
-			return "", errTurnSilent
-		}
-		b.addMessageToChatMemory(chatMemory, userMessage)
-
-		contextMessages := b.snapshotForTurn(turn, chatMemory)
-		return b.getAnthropicResponse(
-			ctx, chatID, contextMessages, false,
-			username, firstName, lastName, isPremium, languageCode, messageTime,
-			out.onSegment, out.onProgress,
-		)
-	})
-}
-
-// uploadPhotoItems uploads every photo in items to the Files API in parallel
-// and returns their file ids with the album caption. On any failure it deletes
-// the uploads that did succeed, tells the user, and returns errTurnSilent.
-func (b *Bot) uploadPhotoItems(ctx context.Context, items []*models.Message, chatID int64, businessConnectionID string) ([]string, string, error) {
 	uploaded := make([]string, len(items))
 	caption := ""
 	g, gctx := errgroup.WithContext(ctx)
@@ -204,7 +143,7 @@ func (b *Bot) uploadPhotoItems(ctx context.Context, items []*models.Message, cha
 		if sendErr := b.sendResponse(ctx, chatID, "Sorry, I couldn't process one of your images.", businessConnectionID); sendErr != nil {
 			ErrorLogger.Printf("Error sending photo failure message: %v", sendErr)
 		}
-		return nil, "", errTurnSilent
+		return
 	}
 	finalUploaded := make([]string, 0, len(uploaded))
 	for _, fid := range uploaded {
@@ -212,38 +151,41 @@ func (b *Bot) uploadPhotoItems(ctx context.Context, items []*models.Message, cha
 			finalUploaded = append(finalUploaded, fid)
 		}
 	}
-	return finalUploaded, caption, nil
-}
-
-// respondToChat runs one assistant turn against the chat's current memory and
-// streams the reply back. Both the immediate path and the debounced flush go
-// through here, so a coalesced turn is byte-for-byte the same request as a
-// single-message one: the messages were already written to memory at intake, and
-// the model simply sees more of them.
-//
-// If a turn is already running for the chat, beginTextTurn decides whether this
-// flush is dropped, restarts it, or runs after it; see turn.go.
-func (b *Bot) respondToChat(req turnRequest) {
-	wantID := b.newestUserMessageIDInChat(req.chatID)
-	if wantID == 0 {
-		// Nothing of the user's is in memory, for example because /clear
-		// landed after this flush was collected. There is nothing to answer.
-		InfoLogger.Printf("[%s] turn: chat %d has no user message to answer; skipping", b.config.ID, req.chatID)
-		return
-	}
-	turn, turnCtx, ok := b.beginTextTurn(req, wantID)
-	if !ok {
+	if len(finalUploaded) == 0 {
 		return
 	}
 
-	b.runTurn(turn, turnCtx, req, func(ctx context.Context, out replyOutput) (string, error) {
-		contextMessages := b.snapshotForTurn(turn, b.getOrCreateChatMemory(req.chatID))
-		return b.getAnthropicResponse(
-			ctx, req.chatID, contextMessages, req.isEmojiOnly,
-			req.username, req.firstName, req.lastName, req.isPremium, req.languageCode, req.messageTime,
-			out.onSegment, out.onProgress,
+	chatMemory := b.getOrCreateChatMemory(chatID)
+	userMessage := b.createMessage(chatID, userID, username, "user", caption, true)
+	userMessage.ImageFileIDs = finalUploaded
+	if err := b.storeMessage(&userMessage); err != nil {
+		b.compensatingDelete(ctx, finalUploaded)
+		ErrorLogger.Printf("[%s] store photo message failed: %v", b.config.ID, err)
+		if sendErr := b.sendResponse(ctx, chatID, "Sorry, I had trouble saving your message.", businessConnectionID); sendErr != nil {
+			ErrorLogger.Printf("Error sending store failure message: %v", sendErr)
+		}
+		return
+	}
+	b.addMessageToChatMemory(chatMemory, userMessage)
+
+	contextMessages := b.prepareContextMessages(chatMemory)
+	joined, err := b.getAnthropicResponse(
+		ctx, chatID, contextMessages, false,
+		username, firstName, lastName, isPremium, languageCode, messageTime,
+		func(seg string) error {
+			return b.sendOneSegment(ctx, chatID, seg, businessConnectionID)
+		},
 	)
-	})
+	if err != nil {
+		ErrorLogger.Printf("Error getting Anthropic response for photo: %v", err)
+		if sendErr := b.sendResponse(ctx, chatID, b.anthropicErrorResponse(err, userID), businessConnectionID); sendErr != nil {
+			ErrorLogger.Printf("Error sending anthropic error response: %v", sendErr)
+		}
+		return
+	}
+	if _, storeErr := b.screenOutgoingMessage(chatID, joined); storeErr != nil {
+		ErrorLogger.Printf("Error recording assistant turn: %v", storeErr)
+	}
 }
 
 func (b *Bot) anthropicErrorResponse(err error, userID int64) string {
@@ -277,19 +219,6 @@ func (b *Bot) anthropicErrorResponse(err error, userID int64) string {
 }
 
 func (b *Bot) handleUpdate(ctx context.Context, tgBot *bot.Bot, update *models.Update) {
-	if stopped := update.StoppedMessageGeneration; stopped != nil {
-		b.handleStopGeneration(stopped.Chat.ID, stopped.DraftID)
-		return
-	}
-	if edited := update.EditedMessage; edited != nil {
-		b.handleEditedMessage(ctx, edited)
-		return
-	}
-	if edited := update.EditedBusinessMessage; edited != nil {
-		b.handleEditedMessage(ctx, edited)
-		return
-	}
-
 	var message *models.Message
 
 	if update.Message != nil {
@@ -339,17 +268,12 @@ func (b *Bot) handleUpdate(ctx context.Context, tgBot *bot.Bot, update *models.U
 		}
 	}
 
-	// Media never waits on the text debounce window. Cancelling here does not
-	// discard the buffered text: those messages are already in chat memory, so
-	// the turn this media triggers answers them too.
 	if message.MediaGroupID != "" && len(message.Photo) > 0 {
-		b.cancelIntake(chatID)
 		b.bufferAlbumItem(ctx, message, chatID, userID, username, firstName, lastName,
 			isPremium, languageCode, messageTime, businessConnectionID)
 		return
 	}
 	if len(message.Photo) > 0 {
-		b.cancelIntake(chatID)
 		if !b.checkRateLimits(userID) {
 			b.sendRateLimitExceededMessage(ctx, chatID, businessConnectionID)
 			return
@@ -498,17 +422,15 @@ func (b *Bot) handleUpdate(ctx context.Context, tgBot *bot.Bot, update *models.U
 	}
 
 	if message.Voice != nil {
-		b.cancelIntake(chatID)
 		b.handleVoiceMessage(ctx, message, userMsg, chatID, userID, username, firstName, lastName, isPremium, languageCode, messageTime, businessConnectionID)
 		return
 	}
 
+	chatMemory := b.getOrCreateChatMemory(chatID)
+	contextMessages := b.prepareContextMessages(chatMemory)
+
 	if message.Sticker != nil {
-		b.cancelIntake(chatID)
-		turn, turnCtx := b.beginTurn(ctx, chatID)
-		defer b.endTurn(turn)
-		contextMessages := b.snapshotForTurn(turn, b.getOrCreateChatMemory(chatID))
-		b.handleStickerMessage(turnCtx, chatID, userMsg, message, contextMessages, businessConnectionID)
+		b.handleStickerMessage(ctx, chatID, userMsg, message, contextMessages, businessConnectionID)
 		return
 	}
 
@@ -519,98 +441,24 @@ func (b *Bot) handleUpdate(ctx context.Context, tgBot *bot.Bot, update *models.U
 
 	isEmojiOnly := isOnlyEmojis(text)
 
-	// Plain text is the only thing that debounces: it is what users fragment
-	// across several sends, and it is the only kind whose meaning survives being
-	// read as one turn.
-	if b.config.DebounceWindow() > 0 {
-		b.bufferIntake(ctx, chatID, userID, username, firstName, lastName,
-			isPremium, languageCode, messageTime, businessConnectionID, isEmojiOnly)
+	joined, err := b.getAnthropicResponse(
+		ctx, chatID, contextMessages, isEmojiOnly,
+		username, firstName, lastName, isPremium, languageCode, messageTime,
+		func(seg string) error {
+			return b.sendOneSegment(ctx, chatID, seg, businessConnectionID)
+		},
+	)
+	if err != nil {
+		ErrorLogger.Printf("Error getting Anthropic response: %v", err)
+		if sendErr := b.sendResponse(ctx, chatID, b.anthropicErrorResponse(err, userID), businessConnectionID); sendErr != nil {
+			ErrorLogger.Printf("Error sending response: %v", sendErr)
+		}
 		return
 	}
 
-	b.respondToChat(turnRequest{
-		ctx: ctx, chatID: chatID, userID: userID, isEmojiOnly: isEmojiOnly,
-		username: username, firstName: firstName, lastName: lastName,
-		isPremium: isPremium, languageCode: languageCode, messageTime: messageTime,
-		businessConnectionID: businessConnectionID,
-	})
+	if _, storeErr := b.screenOutgoingMessage(chatID, joined); storeErr != nil {
+		ErrorLogger.Printf("Error recording assistant turn: %v", storeErr)
 	}
-
-// handleEditedMessage applies a Telegram edit to the stored copy of a user
-// message. Mobile clients deliver autocorrect fixes as edits within a few
-// seconds of the send, so the model should see the corrected text. What
-// happens next depends on where the message is:
-//
-//   - still in its quiet window: the window restarts as if it had just arrived;
-//   - being read by a turn that has shown nothing yet: that turn restarts;
-//   - already answered, or shown: only the record is corrected.
-func (b *Bot) handleEditedMessage(ctx context.Context, edited *models.Message) {
-	text := edited.Text
-	if text == "" {
-		text = edited.Caption
-	}
-	if edited.From == nil || text == "" {
-		return
-	}
-	chatID := edited.Chat.ID
-
-	// Memory is the gate: only a message still in memory can influence a turn,
-	// and the check is cheap. It also keeps edits from chats this bot never
-	// stored off the database; a business connection delivers every edit on
-	// the whole account.
-	var msgID uint
-	answered := false
-	b.chatMemoriesMu.Lock()
-	if mem, exists := b.chatMemories[chatID]; exists {
-		for i := len(mem.Messages) - 1; i >= 0; i-- {
-			m := &mem.Messages[i]
-			if m.IsUser && m.TelegramMessageID == edited.ID {
-				m.Text = text
-				msgID = m.ID
-				break
-			}
-			if !m.IsUser {
-				answered = true
-			}
-		}
-	}
-	b.chatMemoriesMu.Unlock()
-	if msgID == 0 {
-		return
-	}
-
-	if err := b.db.Model(&Message{}).Where("id = ?", msgID).Update("text", text).Error; err != nil {
-		ErrorLogger.Printf("[%s] applying edit to message %d in chat %d: %v", b.config.ID, edited.ID, chatID, err)
-	}
-	if answered {
-		return
-	}
-
-	isEmojiOnly := isOnlyEmojis(text)
-	if b.touchIntake(chatID, isEmojiOnly) {
-		InfoLogger.Printf("[%s] edit to message %d re-armed the quiet window for chat %d", b.config.ID, edited.ID, chatID)
-		return
-	}
-
-	req := turnRequest{
-		ctx: ctx, chatID: chatID, userID: edited.From.ID,
-		username: edited.From.Username, firstName: edited.From.FirstName, lastName: edited.From.LastName,
-		languageCode: edited.From.LanguageCode, isPremium: edited.From.IsPremium, messageTime: edited.Date,
-		businessConnectionID: edited.BusinessConnectionID, isEmojiOnly: isEmojiOnly,
-	}
-	if b.restartForEdit(req, msgID) {
-		InfoLogger.Printf("[%s] edit to message %d restarted the running turn for chat %d", b.config.ID, edited.ID, chatID)
-	}
-}
-
-// handleStopGeneration routes Telegram's stopped_message_generation update,
-// sent when the user taps Stop under a streaming draft.
-func (b *Bot) handleStopGeneration(chatID int64, draftID int) {
-	if b.stopTurn(chatID, draftID) {
-		InfoLogger.Printf("[%s] user stopped generation in chat %d (draft %d)", b.config.ID, chatID, draftID)
-		return
-	}
-	InfoLogger.Printf("[%s] stop for chat %d ignored: draft %d is not the running turn", b.config.ID, chatID, draftID)
 }
 
 func (b *Bot) sendRateLimitExceededMessage(ctx context.Context, chatID int64, businessConnectionID string) {
@@ -621,7 +469,7 @@ func (b *Bot) sendRateLimitExceededMessage(ctx context.Context, chatID int64, bu
 
 func (b *Bot) handleStickerMessage(ctx context.Context, chatID int64, userMessage Message, message *models.Message, contextMessages []anthropic.BetaMessageParam, businessConnectionID string) {
 
-	response, err := b.generateStickerResponse(ctx, userMessage, contextMessages, businessConnectionID)
+	response, err := b.generateStickerResponse(ctx, userMessage, contextMessages)
 	if err != nil {
 		ErrorLogger.Printf("Error generating sticker response: %v", err)
 		if message.Sticker.IsAnimated {
@@ -639,13 +487,10 @@ func (b *Bot) handleStickerMessage(ctx context.Context, chatID int64, userMessag
 	}
 }
 
-func (b *Bot) generateStickerResponse(ctx context.Context, message Message, contextMessages []anthropic.BetaMessageParam, businessConnectionID string) (string, error) {
-	stopTyping := b.startChatAction(ctx, message.ChatID, businessConnectionID, models.ChatActionTyping)
-	defer stopTyping()
-
+func (b *Bot) generateStickerResponse(ctx context.Context, message Message, contextMessages []anthropic.BetaMessageParam) (string, error) {
 	if message.StickerFileID != "" {
 		messageTime := int(message.Timestamp.Unix())
-		response, err := b.getAnthropicResponse(ctx, message.ChatID, contextMessages, true, message.Username, "", "", false, "", messageTime, nil, nil)
+		response, err := b.getAnthropicResponse(ctx, message.ChatID, contextMessages, true, message.Username, "", "", false, "", messageTime, nil)
 		if err != nil {
 			return "", err
 		}
@@ -724,25 +569,6 @@ func (b *Bot) clearChatHistory(ctx context.Context, chatID int64, currentUserID
 		return
 	}
 
-	// Drop any armed intake buffer for the same chat before clearing memory.
-	// Otherwise the debounce timer fires moments later and repopulates the chat
-	// with the very messages that were just deleted — the openclaw/openclaw#51046
-	// failure mode, but with a privacy consequence rather than a stray reply.
-	clearedChatID := chatID
-	if targetUserID != currentUserID {
-		clearedChatID = targetChatID
-		if clearedChatID == 0 {
-			clearedChatID = targetUserID
-		}
-	}
-	if discarded := b.cancelIntake(clearedChatID); discarded > 0 {
-		InfoLogger.Printf("[%s] discarded %d buffered message(s) for chat %d on history clear",
-			b.config.ID, discarded, clearedChatID)
-	}
-	// Likewise a flush collected behind a running turn: rerun against the
-	// emptied chat, it would call the model with no user message at all.
-	b.discardRerun(clearedChatID)
-
 	b.chatMemoriesMu.Lock()
 	if targetUserID == currentUserID {
 		delete(b.chatMemories, chatID)

handlers_test.go

@@ -593,7 +593,7 @@ func setupBotForTest(t *testing.T, ownerID int64) (*Bot, *MockTelegramClient) {
 		MessagePerHour:  5,
 		MessagePerDay:   10,
 		TempBanDuration: "1h",
-		Model:           "claude-haiku-4-5",
+		Model:           "claude-3-5-haiku-latest",
 		SystemPrompts:   make(map[string]string),
 		Active:          true,
 	}
@@ -632,7 +632,7 @@ func TestAnthropicErrorResponse(t *testing.T) { //NOSONAR go:S100 -- underscore
 		RoleID: userRole.ID, Role: userRole,
 	}).Error)
 
-	modelErr := fmt.Errorf("%w: claude-haiku-4-5", ErrModelNotFound)
+	modelErr := fmt.Errorf("%w: claude-3-5-haiku-latest", ErrModelNotFound)
 	otherErr := errors.New("network error")
 
 	tests := []struct {
@@ -696,7 +696,7 @@ func TestSetModelCommand(t *testing.T) { //NOSONAR go:S100 -- underscore separat
 	defer func() { _ = os.RemoveAll(tempDir) }()
 
 	configPath := filepath.Join(tempDir, "config.json")
-	initialJSON := `{"id":"test_bot","telegram_token":"test_token","model":"claude-haiku-4-5","messages_per_hour":5,"messages_per_day":10}`
+	initialJSON := `{"id":"test_bot","telegram_token":"test_token","model":"claude-3-5-haiku-latest","messages_per_hour":5,"messages_per_day":10}`
 	assert.NoError(t, os.WriteFile(configPath, []byte(initialJSON), 0600))
 	b.config.ConfigFilePath = configPath
 

intake_buffer.go

@@ -1,143 +0,0 @@
-package main
-
-import (
-	"context"
-	"time"
-)
-
-// pendingIntake holds a chat's coalescing window. The buffered message bodies
-// are deliberately absent: screenIncomingMessage has already persisted each
-// message to the database and to chat memory by the time it is buffered, so the
-// flushed turn picks them all up from memory. What is kept here is the request
-// the turn needs, always refreshed to the most recent message in the batch.
-type pendingIntake struct {
-	turnRequest
-	allEmojiOnly bool // every message in the batch was emoji only
-	count        int
-	seq          uint64
-	timer        *time.Timer
-}
-
-// bufferIntake holds a text message for the configured quiet window instead of
-// dispatching a turn immediately, resetting the window on each new message.
-// Rapid follow-ups therefore produce one reply rather than one per message.
-func (b *Bot) bufferIntake(
-	ctx context.Context,
-	chatID, userID int64,
-	username, firstName, lastName string,
-	isPremium bool,
-	languageCode string,
-	messageTime int,
-	businessConnectionID string,
-	isEmojiOnly bool,
-) {
-	window := b.config.DebounceWindow()
-
-	b.intakeBuffersMu.Lock()
-	defer b.intakeBuffersMu.Unlock()
-
-	pending, exists := b.intakeBuffers[chatID]
-	if !exists {
-		b.intakeSeq++
-		pending = &pendingIntake{seq: b.intakeSeq, allEmojiOnly: true}
-		b.intakeBuffers[chatID] = pending
-	}
-
-	// Reply metadata tracks the most recent message in the batch.
-	pending.turnRequest = turnRequest{
-		ctx: ctx, chatID: chatID, userID: userID,
-		username: username, firstName: firstName, lastName: lastName,
-		isPremium: isPremium, languageCode: languageCode, messageTime: messageTime,
-		businessConnectionID: businessConnectionID,
-	}
-	pending.allEmojiOnly = pending.allEmojiOnly && isEmojiOnly
-	pending.count++
-
-	b.armIntakeLocked(pending, window)
-}
-
-// armIntakeLocked starts or restarts the quiet-window timer. The caller holds
-// intakeBuffersMu.
-func (b *Bot) armIntakeLocked(pending *pendingIntake, window time.Duration) {
-	if pending.timer != nil {
-		pending.timer.Stop()
-	}
-	ctx, chatID, seq := pending.ctx, pending.chatID, pending.seq
-	pending.timer = time.AfterFunc(window, func() {
-		b.flushIntake(ctx, chatID, seq)
-	})
-}
-
-// touchIntake restarts the quiet window for a chat with a pending batch, as an
-// edit to one of its messages does, and reports whether there was one.
-// isEmojiOnly is the edited text's classification; an edit can only make the
-// batch less emoji-only, never more, since other messages are not re-read.
-func (b *Bot) touchIntake(chatID int64, isEmojiOnly bool) bool {
-	b.intakeBuffersMu.Lock()
-	defer b.intakeBuffersMu.Unlock()
-
-	pending, exists := b.intakeBuffers[chatID]
-	if !exists {
-		return false
-	}
-	pending.allEmojiOnly = pending.allEmojiOnly && isEmojiOnly
-	b.armIntakeLocked(pending, b.config.DebounceWindow())
-	return true
-}
-
-// flushIntake dispatches the coalesced turn for a chat. seq guards against a
-// timer that had already fired before its Stop call landed: a stale goroutine
-// would otherwise flush a buffer belonging to a later batch.
-func (b *Bot) flushIntake(ctx context.Context, chatID int64, seq uint64) {
-	b.intakeBuffersMu.Lock()
-	pending, exists := b.intakeBuffers[chatID]
-	if !exists || pending.seq != seq {
-		b.intakeBuffersMu.Unlock()
-		return
-	}
-	delete(b.intakeBuffers, chatID)
-	captured := *pending
-	b.intakeBuffersMu.Unlock()
-
-	if captured.count > 1 {
-		InfoLogger.Printf("[%s] intake flush: coalesced %d messages into one turn for chat %d",
-			b.config.ID, captured.count, chatID)
-	}
-
-	req := captured.turnRequest
-	req.ctx = ctx
-	req.isEmojiOnly = captured.allEmojiOnly
-	b.respondToChat(req)
-}
-
-// cancelIntake drops a chat's pending buffer without dispatching, returning how
-// many messages were discarded.
-//
-// This is the fix for the class of bug in openclaw/openclaw#51046, where a stop
-// command aborted the running turn but left the debounce buffer armed, so the
-// timer fired afterwards and started the very turn the user had just cancelled.
-// Here the stakes are higher than a stray turn: /clear and /clear_hard delete
-// chat memory, and a surviving buffer would repopulate it moments later with
-// content the user asked to have removed.
-func (b *Bot) cancelIntake(chatID int64) int {
-	b.intakeBuffersMu.Lock()
-	defer b.intakeBuffersMu.Unlock()
-
-	pending, exists := b.intakeBuffers[chatID]
-	if !exists {
-		return 0
-	}
-	if pending.timer != nil {
-		pending.timer.Stop()
-	}
-	delete(b.intakeBuffers, chatID)
-	return pending.count
-}
-
-// hasPendingIntake reports whether a chat currently holds a buffered batch.
-func (b *Bot) hasPendingIntake(chatID int64) bool {
-	b.intakeBuffersMu.Lock()
-	defer b.intakeBuffersMu.Unlock()
-	_, exists := b.intakeBuffers[chatID]
-	return exists
-}

intake_buffer_test.go

@@ -1,176 +0,0 @@
-package main
-
-import (
-	"context"
-	"testing"
-	"time"
-
-	"github.com/go-telegram/bot"
-	"github.com/go-telegram/bot/models"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-// bufferOnly parks a message in the intake buffer without letting the flush run,
-// by using a window long enough that no test waits it out.
-const bufferOnly = 10 * time.Second
-
-func bufferText(b *Bot, chatID int64, isEmojiOnly bool) {
-	b.bufferIntake(context.Background(), chatID, 555,
-		"tester", "Test", "User", false, "en", int(time.Now().Unix()), "", isEmojiOnly)
-}
-
-func TestBufferIntake_CoalescesIntoSingleTurn(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	b.config.DebounceMs = int(bufferOnly / time.Millisecond)
-
-	// The burst this whole feature exists for: one real question, then filler.
-	for i := 0; i < 5; i++ {
-		bufferText(b, 900, false)
-	}
-
-	b.intakeBuffersMu.Lock()
-	defer b.intakeBuffersMu.Unlock()
-	assert.Len(t, b.intakeBuffers, 1, "rapid messages must share one buffer entry")
-	assert.Equal(t, 5, b.intakeBuffers[900].count, "all five messages land in the same batch")
-}
-
-// Separate chats must not share a window; one user's burst cannot delay another's.
-func TestBufferIntake_IsolatesChats(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	b.config.DebounceMs = int(bufferOnly / time.Millisecond)
-
-	bufferText(b, 910, false)
-	bufferText(b, 911, false)
-	bufferText(b, 911, false)
-
-	b.intakeBuffersMu.Lock()
-	defer b.intakeBuffersMu.Unlock()
-	assert.Equal(t, 1, b.intakeBuffers[910].count)
-	assert.Equal(t, 2, b.intakeBuffers[911].count)
-}
-
-func TestBufferIntake_ResetsWindowAndKeepsLatestMetadata(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	b.config.DebounceMs = int(bufferOnly / time.Millisecond)
-
-	b.bufferIntake(context.Background(), 901, 1, "first", "First", "", false, "en", 1000, "", true)
-	b.bufferIntake(context.Background(), 901, 2, "second", "Second", "", true, "de", 2000, "biz-42", false)
-
-	b.intakeBuffersMu.Lock()
-	defer b.intakeBuffersMu.Unlock()
-	pending := b.intakeBuffers[901]
-	require.NotNil(t, pending)
-
-	assert.Equal(t, 2, pending.count)
-	// OpenClaw semantics: reply metadata follows the most recent message.
-	assert.Equal(t, "second", pending.username)
-	assert.Equal(t, int64(2), pending.userID)
-	assert.Equal(t, "de", pending.languageCode)
-	assert.Equal(t, 2000, pending.messageTime)
-	assert.Equal(t, "biz-42", pending.businessConnectionID)
-	assert.True(t, pending.isPremium)
-	// One non-emoji message makes the whole coalesced turn non-emoji.
-	assert.False(t, pending.allEmojiOnly)
-}
-
-func TestBufferIntake_AllEmojiOnlySurvivesWhenEveryMessageIsEmoji(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	b.config.DebounceMs = int(bufferOnly / time.Millisecond)
-
-	bufferText(b, 902, true)
-	bufferText(b, 902, true)
-
-	b.intakeBuffersMu.Lock()
-	defer b.intakeBuffersMu.Unlock()
-	assert.True(t, b.intakeBuffers[902].allEmojiOnly)
-}
-
-func TestCancelIntake_DiscardsPendingBatch(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	b.config.DebounceMs = int(bufferOnly / time.Millisecond)
-
-	bufferText(b, 903, false)
-	bufferText(b, 903, false)
-	require.True(t, b.hasPendingIntake(903))
-
-	assert.Equal(t, 2, b.cancelIntake(903), "cancel reports how many messages it discarded")
-	assert.False(t, b.hasPendingIntake(903))
-	assert.Equal(t, 0, b.cancelIntake(903), "cancelling an empty chat is a no-op")
-}
-
-// A cancelled buffer must never dispatch afterwards. This is the openclaw#51046
-// shape: the timer had already been armed when the cancel landed.
-func TestCancelIntake_TimerDoesNotFireAfterCancel(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	b.config.DebounceMs = 20
-
-	bufferText(b, 904, false)
-	b.cancelIntake(904)
-
-	time.Sleep(80 * time.Millisecond)
-	assert.False(t, b.hasPendingIntake(904), "cancelled buffer must not resurrect")
-}
-
-// A stale timer from a cancelled batch must not flush a newer batch early.
-func TestFlushIntake_IgnoresSupersededSequence(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	b.config.DebounceMs = int(bufferOnly / time.Millisecond)
-
-	bufferText(b, 905, false)
-	b.intakeBuffersMu.Lock()
-	staleSeq := b.intakeBuffers[905].seq
-	b.intakeBuffersMu.Unlock()
-
-	b.cancelIntake(905)
-	bufferText(b, 905, false) // new batch, new sequence
-
-	// The old timer firing late must not consume the new batch.
-	b.flushIntake(context.Background(), 905, staleSeq)
-	assert.True(t, b.hasPendingIntake(905), "superseded flush must leave the newer batch armed")
-}
-
-func TestClearChatHistory_CancelsPendingIntake(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, mockTg := setupBotForTest(t, 123)
-	b.config.DebounceMs = int(bufferOnly / time.Millisecond)
-	mockTg.SendMessageFunc = func(_ context.Context, _ *bot.SendMessageParams) (*models.Message, error) {
-		return &models.Message{}, nil
-	}
-
-	const chatID int64 = 906
-	bufferText(b, chatID, false)
-	require.True(t, b.hasPendingIntake(chatID))
-
-	b.clearChatHistory(context.Background(), chatID, 123, 0, 0, "", false)
-
-	assert.False(t, b.hasPendingIntake(chatID),
-		"clearing history must disarm the buffer, or deleted messages get replayed into memory")
-}
-
-func TestDebounceWindow(t *testing.T) {
-	cases := []struct {
-		name string
-		ms   int
-		want time.Duration
-	}{
-		{"unset disables debouncing", 0, 0},
-		{"negative disables debouncing", -1, 0},
-		{"positive converts to duration", 2500, 2500 * time.Millisecond},
-	}
-	for _, tc := range cases {
-		t.Run(tc.name, func(t *testing.T) {
-			c := BotConfig{DebounceMs: tc.ms}
-			assert.Equal(t, tc.want, c.DebounceWindow())
-		})
-	}
-}
-
-func TestCacheHistoryEnabled_DefaultsOn(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	assert.True(t, (&BotConfig{}).CacheHistoryEnabled(), "cache_history defaults to enabled")
-
-	off := false
-	assert.False(t, (&BotConfig{CacheHistory: &off}).CacheHistoryEnabled())
-
-	on := true
-	assert.True(t, (&BotConfig{CacheHistory: &on}).CacheHistoryEnabled())
-}

models.go

@@ -44,10 +44,6 @@ type Message struct {
 	AnsweredOn     *time.Time     `gorm:"index"`
 	ImageFileIDs   []string       `gorm:"type:text;serializer:json"`
 	FilesCleanedAt *time.Time     `gorm:"index"`
-	// TelegramMessageID is Telegram's message_id, kept so an edited_message
-	// update can find and correct the stored copy. Zero on rows stored before
-	// edits were tracked.
-	TelegramMessageID int `gorm:"index"`
 }
 
 type ChatMemory struct {

prompt_cache_test.go

@@ -1,157 +0,0 @@
-package main
-
-import (
-	"testing"
-
-	"github.com/anthropics/anthropic-sdk-go/v2"
-	"github.com/anthropics/anthropic-sdk-go/v2/packages/param"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func TestMarkTrailingCacheBreakpoint(t *testing.T) {
-	t.Run("marks the final text block of the final turn", func(t *testing.T) {
-		msgs := []anthropic.BetaMessageParam{
-			anthropic.NewBetaUserMessage(anthropic.NewBetaTextBlock("older")),
-			anthropic.NewBetaUserMessage(
-				anthropic.NewBetaTextBlock("first"),
-				anthropic.NewBetaTextBlock("last"),
-			),
-		}
-
-		markTrailingCacheBreakpoint(msgs)
-
-		last := msgs[1].OfUser.Content[1].OfText
-		require.NotNil(t, last)
-		assert.False(t, param.IsOmitted(last.CacheControl),
-			"the trailing block carries the breakpoint")
-
-		// Everything earlier stays unmarked: one breakpoint, not one per block.
-		assert.True(t, param.IsOmitted(msgs[1].OfUser.Content[0].OfText.CacheControl))
-		assert.True(t, param.IsOmitted(msgs[0].OfUser.Content[0].OfText.CacheControl))
-	})
-
-	t.Run("marks a trailing image block", func(t *testing.T) {
-		msgs := []anthropic.BetaMessageParam{
-			anthropic.NewBetaUserMessage(
-				anthropic.NewBetaImageBlock(anthropic.BetaFileImageSourceParam{FileID: "file_1"}),
-			),
-		}
-
-		markTrailingCacheBreakpoint(msgs)
-
-		require.NotNil(t, msgs[0].OfUser.Content[0].OfImage)
-		assert.False(t, param.IsOmitted(msgs[0].OfUser.Content[0].OfImage.CacheControl))
-	})
-
-	t.Run("tolerates empty input", func(t *testing.T) {
-		assert.NotPanics(t, func() { markTrailingCacheBreakpoint(nil) })
-		assert.NotPanics(t, func() {
-			markTrailingCacheBreakpoint([]anthropic.BetaMessageParam{{}})
-		})
-	})
-}
-
-func TestPrepareContextMessages_CacheHistoryToggle(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	memory := &ChatMemory{
-		Messages: []Message{
-			{IsUser: true, Text: "hello"},
-			{IsUser: false, Text: "hi there"},
-		},
-		Size: 10,
-	}
-
-	t.Run("enabled by default", func(t *testing.T) {
-		msgs := b.prepareContextMessages(memory)
-		require.Len(t, msgs, 2)
-		require.NotNil(t, msgs[0].OfUser, "user memory becomes a user turn")
-		require.NotNil(t, msgs[1].OfAssistant, "bot memory becomes an assistant turn")
-		assert.False(t, param.IsOmitted(msgs[1].OfAssistant.Content[0].OfText.CacheControl))
-	})
-
-	t.Run("opt-out leaves history unmarked", func(t *testing.T) {
-		off := false
-		b.config.CacheHistory = &off
-		defer func() { b.config.CacheHistory = nil }()
-
-		msgs := b.prepareContextMessages(memory)
-		require.Len(t, msgs, 2)
-		assert.True(t, param.IsOmitted(msgs[1].OfAssistant.Content[0].OfText.CacheControl))
-	})
-}
-
-func TestValidateConfig_DebounceMs(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	base := func(ms int) *BotConfig {
-		return &BotConfig{
-			ID:             "b",
-			TelegramToken:  "t",
-			Model:          "claude-sonnet-4-6",
-			MessagePerHour: 1,
-			MessagePerDay:  1,
-			DebounceMs:     ms,
-		}
-	}
-
-	cases := []struct {
-		name    string
-		ms      int
-		wantErr bool
-	}{
-		{"omitted is valid", 0, false},
-		{"typical chat window is valid", 2500, false},
-		{"at the ceiling is valid", maxDebounceMs, false},
-		{"negative is rejected", -1, true},
-		{"above the ceiling is rejected", maxDebounceMs + 1, true},
-	}
-
-	for _, tc := range cases {
-		t.Run(tc.name, func(t *testing.T) {
-			err := validateConfig(base(tc.ms), map[string]bool{}, map[string]bool{})
-			if tc.wantErr {
-				assert.Error(t, err)
-				return
-			}
-			assert.NoError(t, err)
-		})
-	}
-}
-
-func TestValidateConfig_Effort(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	base := func(effort string) *BotConfig {
-		return &BotConfig{
-			ID:             "b",
-			TelegramToken:  "t",
-			Model:          "claude-sonnet-5",
-			MessagePerHour: 1,
-			MessagePerDay:  1,
-			Effort:         effort,
-		}
-	}
-
-	cases := []struct {
-		name    string
-		effort  string
-		wantErr bool
-	}{
-		{"omitted is valid", "", false},
-		{"low is valid", "low", false},
-		{"medium is valid", "medium", false},
-		{"high is valid", "high", false},
-		{"xhigh is valid", "xhigh", false},
-		{"max is valid", "max", false},
-		{"unknown level is rejected", "turbo", true},
-		{"wrong case is rejected", "Low", true},
-	}
-
-	for _, tc := range cases {
-		t.Run(tc.name, func(t *testing.T) {
-			err := validateConfig(base(tc.effort), map[string]bool{}, map[string]bool{})
-			if tc.wantErr {
-				assert.Error(t, err)
-				return
-			}
-			assert.NoError(t, err)
-		})
-	}
-}

telegram_client.go

@@ -9,9 +9,7 @@ import (
 
 type TelegramClient interface {
 	SendMessage(ctx context.Context, params *bot.SendMessageParams) (*models.Message, error)
-	SendMessageDraft(ctx context.Context, params *bot.SendMessageDraftParams) (bool, error)
 	SendAudio(ctx context.Context, params *bot.SendAudioParams) (*models.Message, error)
-	SendChatAction(ctx context.Context, params *bot.SendChatActionParams) (bool, error)
 	SetMyCommands(ctx context.Context, params *bot.SetMyCommandsParams) (bool, error)
 	GetFile(ctx context.Context, params *bot.GetFileParams) (*models.File, error)
 	FileDownloadLink(f *models.File) string

telegram_client_mock.go

@@ -11,9 +11,7 @@ import (
 type MockTelegramClient struct {
 	mock.Mock
 	SendMessageFunc      func(ctx context.Context, params *bot.SendMessageParams) (*models.Message, error)
-	SendMessageDraftFunc func(ctx context.Context, params *bot.SendMessageDraftParams) (bool, error)
 	SendAudioFunc        func(ctx context.Context, params *bot.SendAudioParams) (*models.Message, error)
-	SendChatActionFunc   func(ctx context.Context, params *bot.SendChatActionParams) (bool, error)
 	SetMyCommandsFunc    func(ctx context.Context, params *bot.SetMyCommandsParams) (bool, error)
 	GetFileFunc          func(ctx context.Context, params *bot.GetFileParams) (*models.File, error)
 	FileDownloadLinkFunc func(f *models.File) string
@@ -31,13 +29,6 @@ func (m *MockTelegramClient) SendMessage(ctx context.Context, params *bot.SendMe
 	return nil, args.Error(1)
 }
 
-func (m *MockTelegramClient) SendMessageDraft(ctx context.Context, params *bot.SendMessageDraftParams) (bool, error) {
-	if m.SendMessageDraftFunc != nil {
-		return m.SendMessageDraftFunc(ctx, params)
-	}
-	return true, nil
-}
-
 func (m *MockTelegramClient) SetMyCommands(ctx context.Context, params *bot.SetMyCommandsParams) (bool, error) {
 	if m.SetMyCommandsFunc != nil {
 		return m.SetMyCommandsFunc(ctx, params)
@@ -52,13 +43,6 @@ func (m *MockTelegramClient) SendAudio(ctx context.Context, params *bot.SendAudi
 	return nil, nil
 }
 
-func (m *MockTelegramClient) SendChatAction(ctx context.Context, params *bot.SendChatActionParams) (bool, error) {
-	if m.SendChatActionFunc != nil {
-		return m.SendChatActionFunc(ctx, params)
-	}
-	return true, nil
-}
-
 func (m *MockTelegramClient) GetFile(ctx context.Context, params *bot.GetFileParams) (*models.File, error) {
 	if m.GetFileFunc != nil {
 		return m.GetFileFunc(ctx, params)

turn.go

@@ -1,321 +0,0 @@
-package main
-
-import (
-	"context"
-	"errors"
-	"strconv"
-	"strings"
-
-	"github.com/go-telegram/bot/models"
-)
-
-// cancelReason records why a running turn's context was cancelled, so the
-// turn can tell a deliberate interruption from a real failure.
-type cancelReason int
-
-const (
-	cancelNone cancelReason = iota
-	// cancelRestart: a newer message or an edit arrived before anything was
-	// shown to the user. The turn is abandoned and rerun with it included.
-	cancelRestart
-	// cancelStop: the user pressed Telegram's Stop button under the draft.
-	cancelStop
-)
-
-// errTurnSilent marks a failure runTurn must not report: the turn body has
-// either already told the user, or deliberately has nothing to say.
-var errTurnSilent = errors.New("turn: failure handled by the turn body")
-
-// turnRequest is everything one assistant turn needs to run and reply.
-type turnRequest struct {
-	ctx                                         context.Context
-	chatID, userID                              int64
-	username, firstName, lastName, languageCode string
-	isPremium                                   bool
-	messageTime                                 int
-	businessConnectionID                        string
-	isEmojiOnly                                 bool
-}
-
-// chatTurn is one running assistant turn. At most one exists per chat.
-//
-// This is the second half of the debounce design. The intake buffer decides
-// when a burst is quiet enough to answer; this decides what happens to a
-// message that lands after that, while the model is already working. Without
-// it every late fragment started another concurrent turn, and a slow typist
-// got one reply per fragment, each split into several messages.
-type chatTurn struct {
-	seq    uint64
-	chatID int64
-	cancel context.CancelFunc
-	done   chan struct{}
-
-	// Guarded by Bot.turnsMu.
-	draftID   string       // Telegram draft id shown for this turn, "" if none
-	coveredID uint         // highest user Message.ID this turn's context includes
-	committed bool         // must not be restarted: output shown, or work not cheaply redone
-	reason    cancelReason // why cancel was called, if it was
-	rerun     *turnRequest // a flush that arrived mid-turn and may still need a turn
-	rerunWant uint         // the newest user message that flush wanted answered
-}
-
-func (t *chatTurn) draftIDString() string { return strconv.FormatUint(t.seq, 10) }
-
-func (b *Bot) registerTurnLocked(parent context.Context, chatID int64, coveredID uint, committed bool) (*chatTurn, context.Context) {
-	ctx, cancel := context.WithCancel(parent)
-	b.turnSeq++
-	t := &chatTurn{
-		seq:       b.turnSeq,
-		chatID:    chatID,
-		cancel:    cancel,
-		done:      make(chan struct{}),
-		coveredID: coveredID,
-		committed: committed,
-	}
-	b.turns[chatID] = t
-	return t, ctx
-}
-
-// beginTextTurn applies the late-message policy for a text flush. wantID is the
-// newest user message the flush wants answered. ok=false means the caller must
-// not run a turn, because one of three things happened instead:
-//
-//   - dropped: the running turn's context already includes wantID, so its reply
-//     answers this flush too;
-//   - restart: nothing has been shown yet, so the running turn is cancelled and
-//     recorded to rerun with the new message in context. This is the only case
-//     where tokens are wasted, and it is bounded by the time to first output;
-//   - collect: output is already on screen, so the flush is recorded to run as
-//     its own turn once this one finishes, unless the finished turn turns out to
-//     have covered it after all (see finishTurn).
-func (b *Bot) beginTextTurn(req turnRequest, wantID uint) (*chatTurn, context.Context, bool) {
-	b.turnsMu.Lock()
-	defer b.turnsMu.Unlock()
-
-	if running, exists := b.turns[req.chatID]; exists {
-		switch {
-		case wantID <= running.coveredID:
-			InfoLogger.Printf("[%s] turn: chat %d flush dropped, message %d already covered by running turn %d",
-				b.config.ID, req.chatID, wantID, running.seq)
-		case !running.committed:
-			running.reason = cancelRestart
-			running.rerun, running.rerunWant = &req, wantID
-			running.cancel()
-			InfoLogger.Printf("[%s] turn: chat %d restarting turn %d, message %d arrived before any output",
-				b.config.ID, req.chatID, running.seq, wantID)
-		default:
-			running.rerun, running.rerunWant = &req, wantID
-			InfoLogger.Printf("[%s] turn: chat %d collected message %d, runs after turn %d",
-				b.config.ID, req.chatID, wantID, running.seq)
-		}
-		return nil, nil, false
-	}
-
-	turn, ctx := b.registerTurnLocked(req.ctx, req.chatID, wantID, false)
-	return turn, ctx, true
-}
-
-// beginTurn registers a turn for a media handler, waiting for any running turn
-// on the chat to finish first. Media is never dropped or coalesced, and the
-// turn is committed from the start: cancelling half-finished uploads or a voice
-// transcription to save a few tokens is not worth the mess.
-//
-// The wait does not watch parent: the running turn derives from the same
-// long-lived context, so if parent is cancelled it finishes promptly anyway,
-// and registering over it would put two turns on one chat.
-func (b *Bot) beginTurn(parent context.Context, chatID int64) (*chatTurn, context.Context) {
-	for {
-		b.turnsMu.Lock()
-		running, exists := b.turns[chatID]
-		if !exists {
-			turn, ctx := b.registerTurnLocked(parent, chatID, 0, true)
-			b.turnsMu.Unlock()
-			return turn, ctx
-		}
-		b.turnsMu.Unlock()
-		<-running.done
-	}
-}
-
-// markCovered records the newest user message the turn's context includes.
-func (b *Bot) markCovered(turn *chatTurn, coveredID uint) {
-	b.turnsMu.Lock()
-	defer b.turnsMu.Unlock()
-	if coveredID > turn.coveredID {
-		turn.coveredID = coveredID
-	}
-}
-
-// markCommitted flags that the user has seen output, so the turn can no longer
-// be restarted, only collected behind.
-func (b *Bot) markCommitted(turn *chatTurn) {
-	b.turnsMu.Lock()
-	defer b.turnsMu.Unlock()
-	turn.committed = true
-}
-
-func (b *Bot) setTurnDraft(turn *chatTurn, draftID string) {
-	b.turnsMu.Lock()
-	defer b.turnsMu.Unlock()
-	turn.draftID = draftID
-}
-
-func (b *Bot) turnReason(turn *chatTurn) cancelReason {
-	b.turnsMu.Lock()
-	defer b.turnsMu.Unlock()
-	return turn.reason
-}
-
-// stopTurn handles Telegram's stopped_message_generation update. The draft id
-// must match the running turn's, so a stale tap on an old draft cannot kill a
-// newer turn. Reports whether a turn was cancelled.
-func (b *Bot) stopTurn(chatID int64, draftID int) bool {
-	b.turnsMu.Lock()
-	defer b.turnsMu.Unlock()
-
-	running, exists := b.turns[chatID]
-	if !exists || running.draftID == "" || running.draftID != strconv.Itoa(draftID) {
-		return false
-	}
-	running.reason = cancelStop
-	running.cancel()
-	return true
-}
-
-// restartForEdit reruns a turn that is reading the pre-edit text of message
-// msgID and has shown nothing yet. Reports whether it did. A committed turn
-// keeps going: the user has already seen a reply to the old wording.
-func (b *Bot) restartForEdit(req turnRequest, msgID uint) bool {
-	b.turnsMu.Lock()
-	defer b.turnsMu.Unlock()
-
-	running, exists := b.turns[req.chatID]
-	if !exists || running.committed || running.coveredID < msgID {
-		return false
-	}
-	running.reason = cancelRestart
-	running.rerun, running.rerunWant = &req, msgID
-	running.cancel()
-	return true
-}
-
-// discardRerun drops a flush collected behind the chat's running turn. Used
-// by /clear, where rerunning against an emptied chat would call the model
-// with no user message at all.
-func (b *Bot) discardRerun(chatID int64) {
-	b.turnsMu.Lock()
-	defer b.turnsMu.Unlock()
-	if running, exists := b.turns[chatID]; exists {
-		running.rerun, running.rerunWant = nil, 0
-	}
-}
-
-// finishTurn unregisters the turn and returns the flush that still needs a turn
-// of its own, if any.
-//
-// A restart always reruns: the cancelled turn delivered nothing. A collected
-// flush reruns only if this turn's final context did not include its message.
-// Media turns snapshot after their uploads or transcription, so a text that
-// arrived during that work was collected and then answered anyway; rerunning
-// it would produce a second, unprompted reply. Stop does not discard the
-// flush: the message predates the tap and still deserves an answer.
-func (b *Bot) finishTurn(turn *chatTurn) *turnRequest {
-	b.turnsMu.Lock()
-	if b.turns[turn.chatID] == turn {
-		delete(b.turns, turn.chatID)
-	}
-	rerun, want, reason, covered := turn.rerun, turn.rerunWant, turn.reason, turn.coveredID
-	b.turnsMu.Unlock()
-
-	turn.cancel()
-	close(turn.done)
-
-	if rerun == nil {
-		return nil
-	}
-	if reason != cancelRestart && want <= covered {
-		InfoLogger.Printf("[%s] turn %d for chat %d covered collected message %d; no rerun",
-			b.config.ID, turn.seq, turn.chatID, want)
-		return nil
-	}
-	return rerun
-}
-
-// endTurn finishes the turn and runs whatever was collected behind it on its
-// own goroutine, so a chain of collected turns does not nest on the stack of
-// whichever update handler happened to run the first one.
-func (b *Bot) endTurn(turn *chatTurn) {
-	if next := b.finishTurn(turn); next != nil {
-		go b.respondToChat(*next)
-	}
-}
-
-// runTurn drives a registered turn to completion: keeps the typing indicator
-// alive, picks the reply sink (draft or segments), runs call, delivers and
-// records the reply, and afterwards runs anything collected while it was busy.
-//
-// call receives the turn's cancellable context and the sink, and returns the
-// joined reply text. Both the debounced text path and the photo path go
-// through here, so the two behave identically once the model is involved.
-//
-// Delivery of a finished reply uses the request's context, not the turn's: a
-// Stop or restart that lands after the stream ended must not turn a complete
-// reply into one that is stored but never sent.
-func (b *Bot) runTurn(turn *chatTurn, turnCtx context.Context, req turnRequest,
-	call func(ctx context.Context, out replyOutput) (string, error)) {
-	defer b.endTurn(turn)
-
-	stopTyping := b.startChatAction(turnCtx, req.chatID, req.businessConnectionID, models.ChatActionTyping)
-	defer stopTyping()
-
-	out := b.newReplyOutput(turnCtx, turn, req.chatID, req.businessConnectionID)
-	defer out.close()
-
-	joined, err := call(turnCtx, out)
-	if err != nil {
-		if errors.Is(err, errTurnSilent) {
-			return
-		}
-		switch b.turnReason(turn) {
-		case cancelRestart:
-			InfoLogger.Printf("[%s] turn %d for chat %d abandoned for restart", b.config.ID, turn.seq, req.chatID)
-			return
-		case cancelStop:
-			b.deliverStoppedTurn(turn, req, out)
-			return
-		}
-		ErrorLogger.Printf("Error getting Anthropic response: %v", err)
-		if sendErr := b.sendResponse(req.ctx, req.chatID, b.anthropicErrorResponse(err, req.userID), req.businessConnectionID); sendErr != nil {
-			ErrorLogger.Printf("Error sending response: %v", sendErr)
-		}
-		return
-	}
-
-	if finishErr := out.finish(req.ctx, joined); finishErr != nil {
-		ErrorLogger.Printf("[%s] delivering reply to chat %d: %v", b.config.ID, req.chatID, finishErr)
-		return
-	}
-	if _, storeErr := b.screenOutgoingMessage(req.chatID, joined); storeErr != nil {
-		ErrorLogger.Printf("Error recording assistant turn: %v", storeErr)
-	}
-}
-
-// deliverStoppedTurn keeps what the user watched being written. Telegram drops
-// the draft shortly after a stop unless the bot re-sends it as a real message,
-// and memory must hold the same text the user saw, or the next turn continues
-// from a reply that was never delivered.
-func (b *Bot) deliverStoppedTurn(turn *chatTurn, req turnRequest, out replyOutput) {
-	partial := strings.TrimSpace(out.partial())
-	InfoLogger.Printf("[%s] turn %d for chat %d stopped by user with %d chars shown",
-		b.config.ID, turn.seq, req.chatID, len(partial))
-	if partial == "" {
-		return
-	}
-	if err := b.sendLongMessage(req.ctx, req.chatID, partial, req.businessConnectionID); err != nil {
-		ErrorLogger.Printf("[%s] sending stopped partial to chat %d: %v", b.config.ID, req.chatID, err)
-		return
-	}
-	if _, err := b.screenOutgoingMessage(req.chatID, partial); err != nil {
-		ErrorLogger.Printf("Error recording stopped assistant turn: %v", err)
-	}
-}

turn_test.go

@@ -1,215 +0,0 @@
-package main
-
-import (
-	"context"
-	"testing"
-	"time"
-
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func textReq(chatID int64) turnRequest {
-	return turnRequest{ctx: context.Background(), chatID: chatID, userID: 555, username: "tester"}
-}
-
-func TestBeginTextTurn_RunsWhenIdle(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-
-	turn, ctx, ok := b.beginTextTurn(textReq(1), 5)
-	require.True(t, ok)
-	require.NotNil(t, turn)
-	assert.NoError(t, ctx.Err())
-
-	assert.Nil(t, b.finishTurn(turn))
-	_, _, ok = b.beginTextTurn(textReq(1), 6)
-	assert.True(t, ok, "chat is free again after finish")
-}
-
-func TestBeginTextTurn_DropsFlushAlreadyCovered(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	turn, ctx, _ := b.beginTextTurn(textReq(2), 5)
-
-	_, _, ok := b.beginTextTurn(textReq(2), 5)
-
-	assert.False(t, ok)
-	assert.NoError(t, ctx.Err(), "a covered flush must not cancel the running turn")
-	assert.Nil(t, b.finishTurn(turn), "nothing to rerun")
-}
-
-func TestBeginTextTurn_RestartsUncommittedTurn(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	turn, ctx, _ := b.beginTextTurn(textReq(3), 5)
-
-	_, _, ok := b.beginTextTurn(textReq(3), 7)
-
-	assert.False(t, ok)
-	assert.ErrorIs(t, ctx.Err(), context.Canceled, "a newer message before any output cancels the turn")
-	assert.Equal(t, cancelRestart, b.turnReason(turn))
-	next := b.finishTurn(turn)
-	require.NotNil(t, next, "the cancelled turn must be rerun")
-	assert.Equal(t, int64(3), next.chatID)
-}
-
-func TestFinishTurn_RestartRerunsEvenIfSnapshotRacedIn(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	turn, _, _ := b.beginTextTurn(textReq(30), 5)
-	_, _, ok := b.beginTextTurn(textReq(30), 7)
-	require.False(t, ok)
-
-	// The cancelled turn's snapshot happened to include message 7, but it
-	// delivered nothing, so the rerun must still happen.
-	b.markCovered(turn, 7)
-
-	assert.NotNil(t, b.finishTurn(turn), "a restart always reruns")
-}
-
-func TestBeginTextTurn_CollectsBehindCommittedTurn(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	turn, ctx, _ := b.beginTextTurn(textReq(4), 5)
-	b.markCommitted(turn)
-
-	_, _, ok := b.beginTextTurn(textReq(4), 7)
-
-	assert.False(t, ok)
-	assert.NoError(t, ctx.Err(), "output already shown: never cancel")
-	assert.Equal(t, cancelNone, b.turnReason(turn))
-	require.NotNil(t, b.finishTurn(turn), "the collected flush runs afterwards")
-}
-
-func TestFinishTurn_DropsCollectedFlushTheTurnEndedUpCovering(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	// A media turn: committed from the start, coverage known only later.
-	turn, _ := b.beginTurn(context.Background(), 40)
-	_, _, ok := b.beginTextTurn(textReq(40), 7)
-	require.False(t, ok, "collected behind the media turn")
-
-	// The media turn's late snapshot includes message 7 and answers it.
-	b.markCovered(turn, 7)
-
-	assert.Nil(t, b.finishTurn(turn), "rerunning would answer message 7 twice")
-}
-
-func TestStopTurn_RequiresMatchingDraft(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	turn, ctx, _ := b.beginTextTurn(textReq(5), 5)
-
-	assert.False(t, b.stopTurn(5, 1), "no draft shown yet: nothing to stop")
-	b.setTurnDraft(turn, "42")
-	assert.False(t, b.stopTurn(5, 41), "a stale draft id is ignored")
-	assert.NoError(t, ctx.Err())
-
-	assert.True(t, b.stopTurn(5, 42))
-	assert.ErrorIs(t, ctx.Err(), context.Canceled)
-	assert.Equal(t, cancelStop, b.turnReason(turn))
-	b.finishTurn(turn)
-}
-
-func TestFinishTurn_StopKeepsCollectedFlush(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	turn, _, _ := b.beginTextTurn(textReq(6), 5)
-	b.markCommitted(turn)
-	b.setTurnDraft(turn, turn.draftIDString())
-	_, _, ok := b.beginTextTurn(textReq(6), 7)
-	require.False(t, ok)
-
-	require.True(t, b.stopTurn(6, int(turn.seq)))
-
-	assert.NotNil(t, b.finishTurn(turn), "a message sent before the Stop tap still deserves an answer")
-}
-
-func TestDiscardRerun_DropsCollectedFlush(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	turn, _, _ := b.beginTextTurn(textReq(8), 5)
-	b.markCommitted(turn)
-	_, _, ok := b.beginTextTurn(textReq(8), 7)
-	require.False(t, ok)
-
-	b.discardRerun(8)
-
-	assert.Nil(t, b.finishTurn(turn), "/clear must not let the collected flush rerun on an empty chat")
-}
-
-func TestRestartForEdit(t *testing.T) {
-	b, _ := setupBotForTest(t, 123)
-
-	t.Run("uncommitted turn reading the edited message restarts", func(t *testing.T) {
-		turn, ctx, _ := b.beginTextTurn(textReq(9), 5)
-		assert.True(t, b.restartForEdit(textReq(9), 5))
-		assert.ErrorIs(t, ctx.Err(), context.Canceled)
-		assert.Equal(t, cancelRestart, b.turnReason(turn))
-		assert.NotNil(t, b.finishTurn(turn))
-	})
-
-	t.Run("committed turn keeps going", func(t *testing.T) {
-		turn, ctx, _ := b.beginTextTurn(textReq(10), 5)
-		b.markCommitted(turn)
-		assert.False(t, b.restartForEdit(textReq(10), 5))
-		assert.NoError(t, ctx.Err())
-		b.finishTurn(turn)
-	})
-
-	t.Run("turn that never read the message is left alone", func(t *testing.T) {
-		turn, ctx, _ := b.beginTextTurn(textReq(11), 3)
-		assert.False(t, b.restartForEdit(textReq(11), 5))
-		assert.NoError(t, ctx.Err())
-		b.finishTurn(turn)
-	})
-
-	t.Run("no running turn", func(t *testing.T) {
-		assert.False(t, b.restartForEdit(textReq(12), 5))
-	})
-}
-
-func TestBeginTurn_WaitsForRunningTurn(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	running, _, _ := b.beginTextTurn(textReq(7), 5)
-
-	acquired := make(chan *chatTurn, 1)
-	go func() {
-		turn, _ := b.beginTurn(context.Background(), 7)
-		acquired <- turn
-	}()
-
-	select {
-	case <-acquired:
-		t.Fatal("a media turn must wait for the running turn")
-	case <-time.After(50 * time.Millisecond):
-	}
-
-	b.finishTurn(running)
-
-	select {
-	case turn := <-acquired:
-		assert.True(t, turn.committed, "media turns are committed from the start")
-		b.finishTurn(turn)
-	case <-time.After(time.Second):
-		t.Fatal("the media turn never acquired the chat")
-	}
-}
-
-func TestBeginTurn_CancelledParentStillWaits(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, _ := setupBotForTest(t, 123)
-	running, _, _ := b.beginTextTurn(textReq(13), 5)
-	parent, cancel := context.WithCancel(context.Background())
-	cancel()
-
-	acquired := make(chan struct{})
-	go func() {
-		turn, _ := b.beginTurn(parent, 13)
-		b.finishTurn(turn)
-		close(acquired)
-	}()
-
-	select {
-	case <-acquired:
-		t.Fatal("a cancelled parent must not register over the running turn")
-	case <-time.After(50 * time.Millisecond):
-	}
-	b.finishTurn(running)
-	select {
-	case <-acquired:
-	case <-time.After(time.Second):
-		t.Fatal("never acquired after the running turn finished")
-	}
-}

typing.go

@@ -1,65 +0,0 @@
-package main
-
-import (
-	"context"
-	"time"
-
-	"github.com/go-telegram/bot"
-	"github.com/go-telegram/bot/models"
-)
-
-// typingRefreshInterval re-arms the chat action before Telegram expires it.
-// The Bot API sets the status "for 5 seconds or less", so anything at or above
-// 5s leaves visible gaps. Telegram also clears the status as soon as the bot
-// sends a message, so streamed segments naturally interrupt it until the next
-// tick; there is no API call to clear it early.
-const typingRefreshInterval = 4 * time.Second
-
-// startChatAction shows a chat action (typing, uploading a photo, recording a
-// voice note) and keeps it alive until the returned stop function runs.
-//
-// The returned function is idempotent and MUST be deferred by the caller. A
-// keepalive loop that can outlive its turn is the failure mode behind
-// openclaw/openclaw#27177, where the indicator stuck on until the process was
-// restarted, so the loop here owns a derived context and exits on the first of:
-// stop being called, or the parent context ending.
-func (b *Bot) startChatAction(
-	ctx context.Context,
-	chatID int64,
-	businessConnectionID string,
-	action models.ChatAction,
-) (stop func()) {
-	actionCtx, cancel := context.WithCancel(ctx)
-
-	send := func() {
-		params := &bot.SendChatActionParams{
-			ChatID: chatID,
-			Action: action,
-		}
-		if businessConnectionID != "" {
-			params.BusinessConnectionID = businessConnectionID
-		}
-		if _, err := b.tgBot.SendChatAction(actionCtx, params); err != nil {
-			// Cosmetic only: a failed indicator must never abort the turn.
-			InfoLogger.Printf("[%s] chat action %q failed for chat %d: %v",
-				b.config.ID, action, chatID, err)
-		}
-	}
-
-	send()
-
-	go func() {
-		ticker := time.NewTicker(typingRefreshInterval)
-		defer ticker.Stop()
-		for {
-			select {
-			case <-actionCtx.Done():
-				return
-			case <-ticker.C:
-				send()
-			}
-		}
-	}()
-
-	return cancel
-}

typing_test.go

@@ -1,133 +0,0 @@
-package main
-
-import (
-	"context"
-	"sync"
-	"sync/atomic"
-	"testing"
-	"time"
-
-	"github.com/go-telegram/bot"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func TestStartChatAction_SendsImmediately(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, mockTg := setupBotForTest(t, 123)
-
-	var mu sync.Mutex
-	var got []*bot.SendChatActionParams
-	mockTg.SendChatActionFunc = func(_ context.Context, p *bot.SendChatActionParams) (bool, error) {
-		mu.Lock()
-		defer mu.Unlock()
-		got = append(got, p)
-		return true, nil
-	}
-
-	stop := b.startChatAction(context.Background(), 42, "biz-7", "typing")
-	stop()
-
-	mu.Lock()
-	defer mu.Unlock()
-	require.Len(t, got, 1, "the indicator must show before the slow work starts, not after")
-	assert.Equal(t, int64(42), got[0].ChatID)
-	assert.EqualValues(t, "typing", got[0].Action)
-	assert.Equal(t, "biz-7", got[0].BusinessConnectionID,
-		"business chats need the connection id or the indicator never renders")
-}
-
-func TestStartChatAction_OmitsEmptyBusinessConnectionID(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, mockTg := setupBotForTest(t, 123)
-
-	var mu sync.Mutex
-	var captured *bot.SendChatActionParams
-	mockTg.SendChatActionFunc = func(_ context.Context, p *bot.SendChatActionParams) (bool, error) {
-		mu.Lock()
-		defer mu.Unlock()
-		captured = p
-		return true, nil
-	}
-
-	stop := b.startChatAction(context.Background(), 42, "", "typing")
-	stop()
-
-	mu.Lock()
-	defer mu.Unlock()
-	require.NotNil(t, captured)
-	assert.Empty(t, captured.BusinessConnectionID)
-}
-
-// The openclaw#27177 guard: once the turn ends, the keepalive must stop. A loop
-// that can re-arm after completion left the indicator stuck on until restart.
-func TestStartChatAction_StopHaltsKeepalive(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, mockTg := setupBotForTest(t, 123)
-
-	var calls atomic.Int32
-	mockTg.SendChatActionFunc = func(_ context.Context, _ *bot.SendChatActionParams) (bool, error) {
-		calls.Add(1)
-		return true, nil
-	}
-
-	stop := b.startChatAction(context.Background(), 42, "", "typing")
-	stop()
-	after := calls.Load()
-
-	// Well past a refresh tick had the loop survived.
-	time.Sleep(50 * time.Millisecond)
-	assert.Equal(t, after, calls.Load(), "no chat action may be sent after stop")
-}
-
-// Cancelling the parent context must also tear the loop down, so a turn aborted
-// upstream cannot leak a goroutine that keeps calling Telegram.
-func TestStartChatAction_ParentCancelHaltsKeepalive(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, mockTg := setupBotForTest(t, 123)
-
-	var calls atomic.Int32
-	mockTg.SendChatActionFunc = func(_ context.Context, _ *bot.SendChatActionParams) (bool, error) {
-		calls.Add(1)
-		return true, nil
-	}
-
-	ctx, cancel := context.WithCancel(context.Background())
-	stop := b.startChatAction(ctx, 42, "", "typing")
-	defer stop()
-
-	cancel()
-	after := calls.Load()
-	time.Sleep(50 * time.Millisecond)
-	assert.Equal(t, after, calls.Load(), "parent cancellation must stop the keepalive")
-}
-
-func TestStartChatAction_StopIsIdempotent(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, mockTg := setupBotForTest(t, 123)
-	mockTg.SendChatActionFunc = func(_ context.Context, _ *bot.SendChatActionParams) (bool, error) {
-		return true, nil
-	}
-
-	stop := b.startChatAction(context.Background(), 42, "", "typing")
-	// The voice path calls stop early and again via defer.
-	assert.NotPanics(t, func() {
-		stop()
-		stop()
-	})
-}
-
-// A failed indicator is cosmetic and must never surface as a turn failure.
-func TestStartChatAction_SendErrorIsNonFatal(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	b, mockTg := setupBotForTest(t, 123)
-	mockTg.SendChatActionFunc = func(_ context.Context, _ *bot.SendChatActionParams) (bool, error) {
-		return false, assert.AnError
-	}
-
-	assert.NotPanics(t, func() {
-		stop := b.startChatAction(context.Background(), 42, "", "typing")
-		stop()
-	})
-}
-
-// Telegram clears the status after "5 seconds or less", so the refresh must be
-// strictly under that or the indicator visibly drops out mid-turn.
-func TestTypingRefreshInterval_UnderTelegramExpiry(t *testing.T) { //NOSONAR go:S100 -- underscore separation is idiomatic in Go test names
-	assert.Less(t, typingRefreshInterval, 5*time.Second,
-		"Telegram expires a chat action after at most 5s")
-}