opencode/picoclaw
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fix: resolve merge conflicts with main, adapt transcriber to interface

Browse source
This commit is contained in:
virat-mankali 2026-03-08 21:21:12 +05:30
parent 9118879d59
commit dfe65da016
2 changed files with 28 additions and 34 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

61
pkg/agent/loop.go
View file

@ -624,50 +624,41 @@ func (al *AgentLoop) processMessage(ctx context.Context, msg bus.InboundMessage)
// Transcribe any audio media refs before passing to the agent. // Transcribe any audio media refs before passing to the agent.
userMessage := msg.Content userMessage := msg.Content
if al.transcriber != nil && len(msg.Media) > 0 && al.mediaStore != nil { if al.transcriber != nil && len(msg.Media) > 0 && al.mediaStore != nil {
logger.DebugCF("agent", "Checking media for transcription", map[string]any{
"media_count": len(msg.Media),
})
for _, ref := range msg.Media { for _, ref := range msg.Media {
localPath, meta, err := al.mediaStore.ResolveWithMeta(ref) localPath, meta, err := al.mediaStore.ResolveWithMeta(ref)
if err != nil { if err != nil {
logger.WarnCF("agent", "Failed to resolve media ref for transcription", map[string]any{
"ref": ref, "error": err.Error(),
})
continue continue
} }
if inferMediaType(meta.Filename, meta.ContentType) != "audio" {
logger.DebugCF("agent", "Skipping non-audio media", map[string]any{ mediaType := inferMediaType(meta.Filename, meta.ContentType)
"ref": ref, "filename": meta.Filename,
// Handle audio transcription
if mediaType == "audio" {
logger.InfoCF("agent", "Transcribing audio", map[string]any{
"ref": ref, "path": localPath, "filename": meta.Filename,
}) })
continue result, err := al.transcriber.Transcribe(ctx, localPath)
} if err != nil {
logger.InfoCF("agent", "Transcribing audio", map[string]any{ logger.WarnCF("agent", "Audio transcription failed", map[string]any{
"ref": ref, "path": localPath, "filename": meta.Filename, "ref": ref, "error": err.Error(),
}) })
result, err := al.transcriber.Transcribe(ctx, localPath) continue
if err != nil { }
logger.WarnCF("agent", "Audio transcription failed", map[string]any{ logger.InfoCF("agent", "Transcribed audio", map[string]any{
"ref": ref, "error": err.Error(), "ref": ref, "length": len(result.Text),
}) })
continue // Replace the [voice]/[audio] placeholder with the actual transcript
} userMessage = strings.NewReplacer("[voice]", "", "[audio]", "").Replace(userMessage)
logger.InfoCF("agent", "Transcribed audio", map[string]any{ userMessage = strings.TrimSpace(userMessage)
"ref": ref, "length": len(result.Text), if userMessage != "" {
}) userMessage = userMessage + "\n\n[Voice transcript]: " + result.Text
// Replace the [voice]/[audio] placeholder with the actual transcript } else {
userMessage = strings.NewReplacer("[voice]", "", "[audio]", "").Replace(userMessage) userMessage = result.Text
userMessage = strings.TrimSpace(userMessage) }
if userMessage != "" {
userMessage = userMessage + "\n\n[Voice transcript]: " + result.Text
} else {
userMessage = result.Text
} }
} }
} else if al.transcriber == nil && len(msg.Media) > 0 {
logger.WarnCF("agent", "Transcriber not configured, skipping media", map[string]any{
"media_count": len(msg.Media),
})
} }
return al.runAgentLoop(ctx, agent, processOptions{ return al.runAgentLoop(ctx, agent, processOptions{
@ -1759,3 +1750,5 @@ func extractParentPeer(msg bus.InboundMessage) *routing.RoutePeer {
} }
return &routing.RoutePeer{Kind: parentKind, ID: parentID} return &routing.RoutePeer{Kind: parentKind, ID: parentID}
} }

1
pkg/providers/openai_compat/provider.go
View file

@ -364,6 +364,7 @@ func parseResponse(body io.Reader) (*LLMResponse, error) {
// openaiMessage is the wire-format message for OpenAI-compatible APIs. // openaiMessage is the wire-format message for OpenAI-compatible APIs.
// It mirrors protocoltypes.Message but omits SystemParts, which is an // It mirrors protocoltypes.Message but omits SystemParts, which is an
// internal field that would be unknown to third-party endpoints. // internal field that would be unknown to third-party endpoints.
// Content can be either a string or an array of content blocks for multimodal messages.
type openaiMessage struct { type openaiMessage struct {
Role string `json:"role"` Role string `json:"role"`
Content string `json:"content"` Content string `json:"content"`