Files
coder/coderd/chatd/chatloop/compaction.go
T
Mathias Fredriksson bdbcd3428b feat(coderd/chatd): unify chat storage on SDK parts and fix file-reference rendering (#22958)
File-reference parts in user messages were flattened to `TextContent` at
write time because fantasy has no file-reference content type. The
frontend never saw them as structured parts.

This moves all write paths (user, assistant, tool) from fantasy envelope
format to `codersdk.ChatMessagePart`. The streaming layer (`chatloop`)
is untouched, the conversion happens at the serialization boundary in
`persistStep`.

Old rows are still readable. `ParseContent` uses a structural heuristic
(`isFantasyEnvelopeFormat`) to distinguish legacy envelopes from SDK
parts. We chose this over try/fallback because fantasy envelopes
partially unmarshal into `ChatMessagePart` (the `type` field matches)
while silently losing content. A guard test enforces that no SDK part
can produce the envelope shape.

This is forward-only: new rows are unreadable by old code. Chat is
behind a feature flag so rollback risk is contained.

Also adds a typed `ChatMessageRole` to replace raw strings and
`fantasy.MessageRole*` casts at the persistence boundary. The type
covers `ChatMessage.Role`, `ChatStreamMessagePart.Role`, the
`PublishMessagePart` callback chain, and all DB write sites.
`fantasy.MessageRole*` remains only where we build `fantasy.Message`
structs for LLM dispatch.

Separately, `ProviderMetadata` was leaking to SSE clients via
`publishMessagePart`. `StripInternal` now runs on both the SSE and REST
paths, covering this.

Other cleanup:

- Old `db2sdk.contentBlockToPart` silently dropped metadata on
text/reasoning/tool-call content. New code preserves it.
- `providerMetadataToOptions` now logs warnings instead of silently
returning nil.
- `db2sdk` shrinks from ~250 lines of parallel conversion to ~15 lines
delegating to `chatprompt.ParseContent()`, removing the `fantasy` import
entirely.

Refs #22821
2026-03-13 17:53:26 +02:00

318 lines
9.5 KiB
Go

package chatloop
import (
"context"
"encoding/json"
"strings"
"time"
"charm.land/fantasy"
"golang.org/x/xerrors"
"github.com/coder/coder/v2/codersdk"
)
const (
defaultCompactionThresholdPercent = int32(70)
minCompactionThresholdPercent = int32(0)
maxCompactionThresholdPercent = int32(100)
defaultCompactionSummaryPrompt = "You are performing a context compaction. " +
"Summarize the conversation so a new assistant can seamlessly " +
"continue the work in progress.\n\n" +
"Include:\n" +
"- The user's overall goal and current task\n" +
"- Key decisions made and their rationale\n" +
"- Concrete technical details: file paths, function names, " +
"commands, APIs, and configurations\n" +
"- Errors encountered and how they were resolved\n" +
"- Current state of the work: what is DONE, what is IN PROGRESS, " +
"and what REMAINS to be done\n" +
"- The specific action the assistant was performing or about to " +
"perform when this summary was triggered\n\n" +
"Be dense and factual. Every sentence should convey essential " +
"context for continuation. Do not include pleasantries or " +
"conversational filler."
defaultCompactionSystemSummaryPrefix = "The following is a summary of " +
"the earlier conversation. The assistant was actively working when " +
"the context was compacted. Continue the work described below:"
defaultCompactionTimeout = 90 * time.Second
)
type CompactionOptions struct {
ThresholdPercent int32
ContextLimit int64
SummaryPrompt string
SystemSummaryPrefix string
Timeout time.Duration
Persist func(context.Context, CompactionResult) error
// ToolCallID and ToolName identify the synthetic tool call
// used to represent compaction in the message stream.
ToolCallID string
ToolName string
// PublishMessagePart publishes streaming parts to connected
// clients so they see "Summarizing..." / "Summarized" UI
// transitions during compaction.
PublishMessagePart func(codersdk.ChatMessageRole, codersdk.ChatMessagePart)
OnError func(error)
}
type CompactionResult struct {
SystemSummary string
SummaryReport string
ThresholdPercent int32
UsagePercent float64
ContextTokens int64
ContextLimit int64
}
// tryCompact checks whether context usage exceeds the compaction
// threshold and, if so, generates and persists a summary. Returns
// (true, nil) when compaction was performed, (false, nil) when not
// needed, and (false, err) on failure.
func tryCompact(
ctx context.Context,
model fantasy.LanguageModel,
compaction *CompactionOptions,
contextLimitFallback int64,
stepUsage fantasy.Usage,
stepMetadata fantasy.ProviderMetadata,
allMessages []fantasy.Message,
) (bool, error) {
config, ok := normalizedCompactionConfig(compaction)
if !ok {
return false, nil
}
contextTokens := contextTokensFromUsage(stepUsage)
if contextTokens <= 0 {
return false, nil
}
metadataLimit := extractContextLimit(stepMetadata)
contextLimit := resolveContextLimit(
metadataLimit.Int64,
config.ContextLimit,
contextLimitFallback,
)
usagePercent, compact := shouldCompact(
contextTokens, contextLimit, config.ThresholdPercent,
)
if !compact {
return false, nil
}
// Publish the "Summarizing..." tool-call indicator so
// connected clients see activity during summary generation.
if config.PublishMessagePart != nil && config.ToolCallID != "" {
config.PublishMessagePart(
codersdk.ChatMessageRoleAssistant,
codersdk.ChatMessageToolCall(config.ToolCallID, config.ToolName, nil),
)
}
summary, err := generateCompactionSummary(
ctx, model, allMessages, config,
)
if err != nil {
return false, err
}
if summary == "" {
// Publish a tool-result error so connected clients
// see the compaction failure.
publishCompactionError(config, "compaction produced an empty summary")
return false, xerrors.New("compaction produced an empty summary")
}
systemSummary := strings.TrimSpace(
config.SystemSummaryPrefix + "\n\n" + summary,
)
persistCtx := context.WithoutCancel(ctx)
err = config.Persist(persistCtx, CompactionResult{
SystemSummary: systemSummary,
SummaryReport: summary,
ThresholdPercent: config.ThresholdPercent,
UsagePercent: usagePercent,
ContextTokens: contextTokens,
ContextLimit: contextLimit,
})
if err != nil {
publishCompactionError(config, "failed to persist compaction result")
return false, xerrors.Errorf("persist compaction: %w", err)
}
// Publish the "Summarized" tool-result part so the client
// transitions from the in-progress indicator to the final
// state.
if config.PublishMessagePart != nil && config.ToolCallID != "" {
resultJSON, _ := json.Marshal(map[string]any{
"summary": summary,
"source": "automatic",
"threshold_percent": config.ThresholdPercent,
"usage_percent": usagePercent,
"context_tokens": contextTokens,
"context_limit_tokens": contextLimit,
})
config.PublishMessagePart(
codersdk.ChatMessageRoleTool,
codersdk.ChatMessageToolResult(config.ToolCallID, config.ToolName, resultJSON, false),
)
}
return true, nil
}
// publishCompactionError sends a tool-result error part so
// connected clients see that compaction failed.
func publishCompactionError(config CompactionOptions, msg string) {
if config.PublishMessagePart == nil || config.ToolCallID == "" {
return
}
errJSON, _ := json.Marshal(map[string]any{
"error": msg,
})
config.PublishMessagePart(
codersdk.ChatMessageRoleTool,
codersdk.ChatMessageToolResult(config.ToolCallID, config.ToolName, errJSON, true),
)
}
// normalizedCompactionConfig returns a copy of the compaction options
// with defaults applied. The bool is false when compaction is
// disabled (nil options, missing Persist callback, or threshold at
// 100%).
func normalizedCompactionConfig(opts *CompactionOptions) (CompactionOptions, bool) {
if opts == nil {
return CompactionOptions{}, false
}
config := *opts
if config.Persist == nil {
return CompactionOptions{}, false
}
if strings.TrimSpace(config.SummaryPrompt) == "" {
config.SummaryPrompt = defaultCompactionSummaryPrompt
}
if strings.TrimSpace(config.SystemSummaryPrefix) == "" {
config.SystemSummaryPrefix = defaultCompactionSystemSummaryPrefix
}
if config.Timeout <= 0 {
config.Timeout = defaultCompactionTimeout
}
if config.ThresholdPercent < minCompactionThresholdPercent ||
config.ThresholdPercent > maxCompactionThresholdPercent {
config.ThresholdPercent = defaultCompactionThresholdPercent
}
if config.ThresholdPercent == maxCompactionThresholdPercent {
return CompactionOptions{}, false
}
return config, true
}
// contextTokensFromUsage returns the total context token count from
// a step's usage report. It sums input, cache-read, and
// cache-creation tokens when available, falling back to TotalTokens
// if none of the granular fields are set.
func contextTokensFromUsage(usage fantasy.Usage) int64 {
total := int64(0)
hasContextTokens := false
if usage.InputTokens > 0 {
total += usage.InputTokens
hasContextTokens = true
}
if usage.CacheReadTokens > 0 {
total += usage.CacheReadTokens
hasContextTokens = true
}
if usage.CacheCreationTokens > 0 {
total += usage.CacheCreationTokens
hasContextTokens = true
}
if !hasContextTokens && usage.TotalTokens > 0 {
total = usage.TotalTokens
}
return total
}
// resolveContextLimit picks the first positive value from metadata,
// configured limit, and fallback — in that priority order. Returns
// 0 when none are positive.
func resolveContextLimit(metadataLimit, configLimit, fallback int64) int64 {
if metadataLimit > 0 {
return metadataLimit
}
if configLimit > 0 {
return configLimit
}
if fallback > 0 {
return fallback
}
return 0
}
// shouldCompact returns the usage percentage and whether it exceeds
// the threshold. Returns (0, false) when contextLimit is
// non-positive.
func shouldCompact(contextTokens, contextLimit int64, thresholdPercent int32) (float64, bool) {
if contextLimit <= 0 {
return 0, false
}
usagePercent := (float64(contextTokens) / float64(contextLimit)) * 100
return usagePercent, usagePercent >= float64(thresholdPercent)
}
// generateCompactionSummary asks the model to summarize the
// conversation so far. The provided messages should contain the
// complete history (system prompt, user/assistant turns, tool
// results). A final user message with the summary prompt is appended
// before calling the model.
func generateCompactionSummary(
ctx context.Context,
model fantasy.LanguageModel,
messages []fantasy.Message,
options CompactionOptions,
) (string, error) {
summaryPrompt := make([]fantasy.Message, 0, len(messages)+1)
summaryPrompt = append(summaryPrompt, messages...)
summaryPrompt = append(summaryPrompt, fantasy.Message{
Role: fantasy.MessageRoleUser,
Content: []fantasy.MessagePart{
fantasy.TextPart{Text: options.SummaryPrompt},
},
})
toolChoice := fantasy.ToolChoiceNone
summaryCtx, cancel := context.WithTimeout(ctx, options.Timeout)
defer cancel()
response, err := model.Generate(summaryCtx, fantasy.Call{
Prompt: summaryPrompt,
ToolChoice: &toolChoice,
})
if err != nil {
return "", xerrors.Errorf("generate summary text: %w", err)
}
parts := make([]string, 0, len(response.Content))
for _, block := range response.Content {
textBlock, ok := fantasy.AsContentType[fantasy.TextContent](block)
if !ok {
continue
}
text := strings.TrimSpace(textBlock.Text)
if text == "" {
continue
}
parts = append(parts, text)
}
return strings.TrimSpace(strings.Join(parts, " ")), nil
}