This commit is contained in:
2026-08-16 16:57:36 +02:00
parent 96f8c1115c
commit 532d439c41
106 changed files with 10147 additions and 494 deletions
+192
View File
@@ -0,0 +1,192 @@
package main
import (
"context"
"encoding/json"
"fmt"
"log"
"os"
"os/signal"
"strings"
"syscall"
"time"
"github.com/descrybe/descrybe-v2/apps/api/internal/aiprompts"
"github.com/descrybe/descrybe-v2/apps/api/internal/config"
"github.com/descrybe/descrybe-v2/apps/api/internal/db"
"github.com/descrybe/descrybe-v2/apps/api/internal/eprel"
"github.com/descrybe/descrybe-v2/apps/api/internal/logredact"
"github.com/descrybe/descrybe-v2/apps/api/internal/processing"
"github.com/google/uuid"
)
func main() {
log.SetOutput(logredact.Writer(os.Stderr))
eans := []string{
"4242005342488",
"4242005479948",
"4242002996967",
"4242005327065",
"4242005382330",
}
companyID := uuid.MustParse("2b3159b0-fc08-415b-b248-35ed02a6baab")
userID := uuid.MustParse("0ce3305c-d810-4b56-b1d4-3c1ed510db76")
cfg, err := config.Load()
if err != nil {
log.Fatalf("config: %v", err)
}
ctx, cancel := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
defer cancel()
pool, err := db.NewPool(ctx, cfg.DatabaseURL, db.PoolOptions{
MaxConns: int32(cfg.DBMaxConns),
MinConns: int32(cfg.DBMinConns),
MaxConnLifetime: cfg.DBMaxConnLifetime,
MaxConnLifetimeJitter: cfg.DBMaxConnLifetimeJitter,
MaxConnIdleTime: cfg.DBMaxConnIdleTime,
HealthCheckPeriod: cfg.DBHealthCheckPeriod,
StatementTimeout: cfg.DBStatementTimeout,
})
if err != nil {
log.Fatalf("db: %v", err)
}
defer pool.Close()
out := os.Getenv("PROBE_OUT_DIR")
if out == "" {
out = `F:\laragon\www\_MY\descrybe-v2\.codehelper\_probe_batch_a`
}
_ = os.MkdirAll(out, 0o755)
pipeline := processing.NewPipeline(pool)
pipeline.BatchSize = cfg.ProcessingBatchSize
pipeline.AI = nil
pipeline.Prompts = aiprompts.NewService(pool)
mockKey := strings.TrimSpace(cfg.OpenAIAPIKey)
if mockKey == "" {
mockKey = "local-test"
}
mockBase := "http://127.0.0.1:18767/v1"
mockModel := "mock-llm"
log.Printf("forcing completer base=%s model=%s (ModeLabel=internal)", mockBase, mockModel)
client := processing.NewOpenAIClient(mockKey, mockBase, mockModel, cfg.ProcessingRPM, cfg.ProcessingMaxRetries)
client.ModeLabel = processing.AIProviderInternal
eprelClient := eprel.NewClient(eprel.Options{Enabled: true, Timeout: 20 * time.Second})
pipeline.Engine = &processing.Engine{
Completer: client,
Vector: processing.NoopVectorCategorizer{},
EPREL: eprelClient,
ProviderMode: processing.AIProviderInternal,
}
rawIDs := make([]uuid.UUID, 0, len(eans))
for _, ean := range eans {
var id uuid.UUID
err := pool.QueryRow(ctx, `
SELECT id FROM raw_products
WHERE company_id = $1 AND gtin = $2`, companyID, ean).Scan(&id)
if err != nil {
log.Fatalf("raw product %s: %v", ean, err)
}
rawIDs = append(rawIDs, id)
log.Printf("ean=%s raw_id=%s", ean, id)
}
// Self-heal only: do NOT fail/reclaim other jobs.
if n, err := processing.BackfillMissingMeta(ctx, pool, companyID); err != nil {
log.Printf("BackfillMissingMeta: %v", err)
} else {
log.Printf("BackfillMissingMeta updated=%d", n)
}
jobs, err := pipeline.StartJob(ctx, companyID, userID, rawIDs, "full")
if err != nil {
log.Fatalf("StartJob: %v", err)
}
if len(jobs) == 0 {
log.Fatal("StartJob returned no jobs")
}
jobID := jobs[0].ID
log.Printf("job_id=%s", jobID)
_ = os.WriteFile(out+string(os.PathSeparator)+"process_id.txt", []byte(jobID.String()+"\n"), 0o644)
defend := make(chan struct{})
go func() {
t := time.NewTicker(500 * time.Millisecond)
defer t.Stop()
for {
select {
case <-defend:
return
case <-ctx.Done():
return
case <-t.C:
// Restore ONLY our job if a sibling parked/yielded it. Never touch other jobs.
_, _ = pool.Exec(context.Background(), `
UPDATE processing_jobs
SET status = 'running', error = NULL, updated_at = now()
WHERE id = $1 AND status IN ('failed','cancelled')
AND (error ILIKE '%P0%' OR error ILIKE '%parked%' OR error ILIKE '%yield%'
OR error ILIKE '%reclaim%' OR error ILIKE '%cleared%' OR error ILIKE '%superseded%'
OR error ILIKE '%exclusive%' OR error ILIKE '%batch_%' OR error ILIKE '%inproc%')`, jobID)
_, _ = pool.Exec(context.Background(), `
UPDATE processing_job_products
SET status = 'pending', error = NULL, updated_at = now()
WHERE job_id = $1 AND status IN ('failed','cancelled')
AND processed_product_id IS NULL
AND (error ILIKE '%P0%' OR error ILIKE '%parked%' OR error ILIKE '%yield%'
OR error ILIKE '%reclaim%' OR error ILIKE '%cleared%' OR error ILIKE '%superseded%'
OR error ILIKE '%exclusive%' OR error ILIKE '%batch_%' OR error ILIKE '%inproc%')`, jobID)
}
}
}()
_, _ = pool.Exec(ctx, `
UPDATE processing_jobs
SET status = 'running', started_at = COALESCE(started_at, now()), updated_at = now()
WHERE id = $1`, jobID)
runCtx, runCancel := context.WithTimeout(ctx, 12*time.Minute)
defer runCancel()
if err := pipeline.ProcessJob(runCtx, jobID); err != nil {
close(defend)
log.Fatalf("ProcessJob: %v", err)
}
close(defend)
items, err := pipeline.LoadV1ProcessJobItems(ctx, companyID, jobID, "full")
if err != nil {
log.Fatalf("LoadV1ProcessJobItems: %v", err)
}
payload := map[string]any{
"data": map[string]any{
"process_id": jobID.String(),
"status": "COMPLETED",
"processing_type": "full",
"total_items": len(items),
"items": items,
},
}
raw, err := json.MarshalIndent(payload, "", " ")
if err != nil {
log.Fatalf("marshal: %v", err)
}
if err := os.WriteFile(out+string(os.PathSeparator)+"final.json", raw, 0o644); err != nil {
log.Fatalf("write final: %v", err)
}
fmt.Printf("JOB %s items=%d\n", jobID, len(items))
for _, it := range items {
ean, _ := it["ean"].(string)
st, _ := it["status"].(string)
title, _ := it["title"].(string)
if len(title) > 50 {
title = title[:50]
}
_, hasEprel := it["eprel"].(map[string]any)
ai, _ := it["ai_provider_mode"].(string)
fmt.Printf("ITEM %s status=%s eprel=%v ai=%s title=%s\n", ean, st, hasEprel, ai, title)
}
}