Stabilize search pipeline and improve preview diagnostics
build-push / docker (push) Successful in 4m14s

This commit is contained in:
AI Assistant
2026-03-13 18:32:54 +09:00
parent 6f3149a443
commit 7dfb1ad2de
8 changed files with 463 additions and 45 deletions
+148 -13
View File
@@ -76,6 +76,27 @@ type PreviewResponse struct {
Qualities []map[string]any `json:"qualities"`
}
type searchDebugSummary struct {
Total int `json:"total"`
BySource map[string]int `json:"bySource"`
WithPreview int `json:"withPreview"`
WithThumbnail int `json:"withThumbnail"`
Top []map[string]any `json:"top"`
Warning string `json:"warning,omitempty"`
DurationMS int64 `json:"durationMs,omitempty"`
GeminiCandidateCap int `json:"geminiCandidateCap,omitempty"`
}
type geminiBatchStats struct {
CandidateCap int `json:"candidateCap"`
Requested int `json:"requested"`
Batches int `json:"batches"`
Succeeded int `json:"succeeded"`
Failed int `json:"failed"`
RecommendedCount int `json:"recommendedCount"`
Errors []string `json:"errors,omitempty"`
}
func RegisterRoutes(router *gin.Engine, app *App) {
router.GET("/healthz", func(c *gin.Context) {
c.JSON(http.StatusOK, gin.H{"status": "ok"})
@@ -259,6 +280,7 @@ func (a *App) runDownload(recordID int64, url, start, end, quality, outputPath s
}
func (a *App) searchMedia(c *gin.Context) {
started := time.Now()
var req struct {
Query string `json:"query"`
Platforms []string `json:"platforms"`
@@ -277,18 +299,24 @@ func (a *App) searchMedia(c *gin.Context) {
if len(queryVariants) == 0 {
queryVariants = []string{req.Query}
}
a.debug("search query variants", gin.H{"query": req.Query, "variants": queryVariants, "platforms": req.Platforms})
a.debug("search query variants", gin.H{
"query": req.Query,
"platforms": req.Platforms,
"variants": queryVariants,
"variantCount": len(queryVariants),
"requestIdHint": time.Now().UnixNano(),
})
enabledPlatforms := normalizePlatforms(req.Platforms)
a.Hub.Broadcast("progress", gin.H{"type": "search", "status": "searching " + selectedPlatformLabel(enabledPlatforms), "progress": 35})
results, err := a.SearchService.SearchMedia(queryVariants, enabledPlatforms)
if err != nil {
a.debug("search backend failed", gin.H{"error": err.Error(), "variants": queryVariants})
a.debug("search backend failed", gin.H{"error": err.Error(), "variants": queryVariants, "durationMs": time.Since(started).Milliseconds()})
a.Hub.Broadcast("progress", gin.H{"type": "search", "status": "search failed", "progress": 100, "message": err.Error()})
c.JSON(http.StatusBadGateway, gin.H{"error": err.Error()})
return
}
a.debug("search backend results", gin.H{"count": len(results), "results": results})
a.debug("search backend summary", summarizeSearchResults(results, time.Since(started), 0, ""))
if len(results) == 0 {
warning := "SearXNG returned no renderable results."
a.Hub.Broadcast("progress", gin.H{"type": "search", "status": "no renderable search results", "progress": 100, "message": warning})
@@ -302,10 +330,10 @@ func (a *App) searchMedia(c *gin.Context) {
rankQuery = strings.Join(queryVariants[:min(len(queryVariants), 3)], " ")
}
scored := rankSearchResults(rankQuery, results)
a.debug("search ranked results", gin.H{"count": len(scored), "results": scored})
a.Hub.Broadcast("progress", gin.H{"type": "search", "status": "analyzing all candidate visuals with Gemini Vision", "progress": 75})
recommended := evaluateAllCandidatesWithGemini(a.GeminiService, req.Query, scored)
a.debug("search gemini recommendations", gin.H{"count": len(recommended), "results": recommended})
a.debug("search ranked summary", summarizeSearchResults(scored, time.Since(started), geminiCandidateLimit(len(scored)), ""))
a.Hub.Broadcast("progress", gin.H{"type": "search", "status": "analyzing top candidate visuals with Gemini Vision", "progress": 75})
recommended, geminiStats := evaluateAllCandidatesWithGemini(a.GeminiService, req.Query, scored)
a.debug("search gemini evaluation", geminiStats)
err = nil
if len(recommended) == 0 {
err = fmt.Errorf("gemini vision returned no recommended items across all candidate batches")
@@ -316,6 +344,7 @@ func (a *App) searchMedia(c *gin.Context) {
fallback = append(fallback, services.AIRecommendation{
Title: result.Title,
Link: result.Link,
Snippet: result.Snippet,
ThumbnailURL: result.ThumbnailURL,
PreviewVideoURL: result.PreviewVideoURL,
Source: result.Source,
@@ -324,12 +353,15 @@ func (a *App) searchMedia(c *gin.Context) {
})
}
warning := err.Error()
a.debug("search fallback summary", summarizeRecommendationResults(fallback, time.Since(started), warning))
a.Hub.Broadcast("progress", gin.H{"type": "search", "status": "Gemini Vision fallback to ranked results", "progress": 90, "message": warning})
c.JSON(http.StatusOK, gin.H{"results": fallback, "warning": warning, "queries": queryVariants})
return
}
response := gin.H{"results": mergeRecommendations(recommended, scored, 20), "queries": queryVariants}
merged := mergeRecommendations(recommended, scored, 20)
a.debug("search complete summary", summarizeRecommendationResults(merged, time.Since(started), ""))
response := gin.H{"results": merged, "queries": queryVariants}
a.Hub.Broadcast("progress", gin.H{"type": "search", "status": "search complete", "progress": 100})
c.JSON(http.StatusOK, response)
}
@@ -406,20 +438,31 @@ func selectedPlatformLabel(platforms map[string]bool) string {
return strings.Join(labels, ", ")
}
func evaluateAllCandidatesWithGemini(service *services.GeminiService, query string, ranked []services.SearchResult) []services.AIRecommendation {
func evaluateAllCandidatesWithGemini(service *services.GeminiService, query string, ranked []services.SearchResult) ([]services.AIRecommendation, geminiBatchStats) {
const chunkSize = 8
limit := geminiCandidateLimit(len(ranked))
stats := geminiBatchStats{
CandidateCap: limit,
Requested: min(limit, len(ranked)),
}
merged := make([]services.AIRecommendation, 0, len(ranked))
seen := map[string]bool{}
for start := 0; start < len(ranked); start += chunkSize {
for start := 0; start < limit; start += chunkSize {
end := start + chunkSize
if end > len(ranked) {
end = len(ranked)
if end > limit {
end = limit
}
batch := ranked[start:end]
stats.Batches++
recommended, err := service.Recommend(query, batch)
if err != nil {
stats.Failed++
if len(stats.Errors) < 5 {
stats.Errors = append(stats.Errors, err.Error())
}
continue
}
stats.Succeeded++
for _, item := range recommended {
if item.Link == "" || seen[item.Link] {
continue
@@ -428,7 +471,8 @@ func evaluateAllCandidatesWithGemini(service *services.GeminiService, query stri
merged = append(merged, item)
}
}
return merged
stats.RecommendedCount = len(merged)
return merged, stats
}
func rankSearchResults(query string, results []services.SearchResult) []services.SearchResult {
@@ -515,6 +559,7 @@ func mergeRecommendations(recommended []services.AIRecommendation, ranked []serv
merged = append(merged, services.AIRecommendation{
Title: item.Title,
Link: item.Link,
Snippet: item.Snippet,
ThumbnailURL: item.ThumbnailURL,
PreviewVideoURL: item.PreviewVideoURL,
Source: item.Source,
@@ -525,6 +570,96 @@ func mergeRecommendations(recommended []services.AIRecommendation, ranked []serv
return merged
}
func geminiCandidateLimit(total int) int {
switch {
case total <= 8:
return total
case total <= 16:
return 12
default:
return 16
}
}
func summarizeSearchResults(results []services.SearchResult, duration time.Duration, geminiCap int, warning string) searchDebugSummary {
bySource := map[string]int{}
withPreview := 0
withThumbnail := 0
top := make([]map[string]any, 0, min(6, len(results)))
for idx, item := range results {
bySource[item.Source]++
if strings.TrimSpace(item.PreviewVideoURL) != "" {
withPreview++
}
if strings.TrimSpace(item.ThumbnailURL) != "" {
withThumbnail++
}
if idx < 6 {
top = append(top, map[string]any{
"title": truncateText(item.Title, 120),
"source": item.Source,
"hasPreview": item.PreviewVideoURL != "",
"hasThumbnail": item.ThumbnailURL != "",
"displayLink": item.DisplayLink,
"snippetSample": truncateText(item.Snippet, 160),
})
}
}
return searchDebugSummary{
Total: len(results),
BySource: bySource,
WithPreview: withPreview,
WithThumbnail: withThumbnail,
Top: top,
Warning: warning,
DurationMS: duration.Milliseconds(),
GeminiCandidateCap: geminiCap,
}
}
func summarizeRecommendationResults(results []services.AIRecommendation, duration time.Duration, warning string) searchDebugSummary {
bySource := map[string]int{}
withPreview := 0
withThumbnail := 0
top := make([]map[string]any, 0, min(6, len(results)))
for idx, item := range results {
bySource[item.Source]++
if strings.TrimSpace(item.PreviewVideoURL) != "" {
withPreview++
}
if strings.TrimSpace(item.ThumbnailURL) != "" {
withThumbnail++
}
if idx < 6 {
top = append(top, map[string]any{
"title": truncateText(item.Title, 120),
"source": item.Source,
"hasPreview": item.PreviewVideoURL != "",
"hasThumbnail": item.ThumbnailURL != "",
"reasonSample": truncateText(item.Reason, 120),
"snippetSample": truncateText(item.Snippet, 160),
})
}
}
return searchDebugSummary{
Total: len(results),
BySource: bySource,
WithPreview: withPreview,
WithThumbnail: withThumbnail,
Top: top,
Warning: warning,
DurationMS: duration.Milliseconds(),
}
}
func truncateText(text string, limit int) string {
trimmed := strings.TrimSpace(text)
if len(trimmed) <= limit {
return trimmed
}
return trimmed[:limit] + "..."
}
func EnsurePaths(downloadsDir, workerScript string) error {
if err := os.MkdirAll(downloadsDir, 0o755); err != nil {
return err