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
+188
View File
@@ -0,0 +1,188 @@
package main
import (
"context"
"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/logredact"
"github.com/descrybe/descrybe-v2/apps/api/internal/processing"
"github.com/google/uuid"
)
func main() {
log.SetOutput(logredact.Writer(os.Stderr))
eans := []string{
"4548736132597",
"194644207427",
"194644034894",
"195949822384",
"3830061960930",
}
if v := strings.TrimSpace(os.Getenv("PROBE_EANS")); v != "" {
eans = strings.Split(v, ",")
for i := range eans {
eans[i] = strings.TrimSpace(eans[i])
}
}
companyID, err := uuid.Parse("2b3159b0-fc08-415b-b248-35ed02a6baab")
if err != nil {
log.Fatal(err)
}
userID, err := uuid.Parse("0ce3305c-d810-4b56-b1d4-3c1ed510db76")
if err != nil {
log.Fatal(err)
}
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()
pipeline := processing.NewPipeline(pool)
pipeline.BatchSize = cfg.ProcessingBatchSize
// Force mock-llm: platform DB OpenAI binding would otherwise override env.
pipeline.AI = nil
pipeline.Prompts = aiprompts.NewService(pool)
mockKey := strings.TrimSpace(cfg.OpenAIAPIKey)
if mockKey == "" {
mockKey = "local-test"
}
mockBase := strings.TrimSpace(cfg.OpenAIBaseURL)
if mockBase == "" {
mockBase = "http://127.0.0.1:18767/v1"
}
mockModel := strings.TrimSpace(cfg.OpenAIModel)
if mockModel == "" {
mockModel = "mock-llm"
}
log.Printf("forcing completer base=%s model=%s", mockBase, mockModel)
pipeline.Engine = &processing.Engine{
Completer: processing.NewOpenAIClient(mockKey, mockBase, mockModel, cfg.ProcessingRPM, cfg.ProcessingMaxRetries),
Vector: processing.NoopVectorCategorizer{},
EPREL: nil, // skip remote EPREL hangs during local scorecard
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)
}
// Clear competing queue noise.
_, _ = pool.Exec(ctx, `
UPDATE processing_jobs
SET status = 'failed', error = 'yielded to electronics in-process probe', updated_at = now()
WHERE status IN ('pending','running','processing')`)
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 chunks=%d", jobID, len(jobs))
defend := make(chan struct{})
go func() {
t := time.NewTicker(400 * time.Millisecond)
defer t.Stop()
for {
select {
case <-defend:
return
case <-ctx.Done():
return
case <-t.C:
_, _ = pool.Exec(context.Background(), `
UPDATE processing_jobs
SET status = 'failed', error = 'yielded to electronics in-process probe', updated_at = now()
WHERE status IN ('pending','running','processing') AND id <> $1`, jobID)
_, _ = 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 '%abandoned%')`, 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 '%abandoned%')`, 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, 3*time.Minute)
defer runCancel()
if err := pipeline.ProcessJob(runCtx, jobID); err != nil {
close(defend)
log.Fatalf("ProcessJob: %v", err)
}
close(defend)
rows, err := pool.Query(ctx, `
SELECT rp.gtin, pjp.status, COALESCE(pjp.error,''), pjp.processed_product_id IS NOT NULL
FROM processing_job_products pjp
JOIN raw_products rp ON rp.id = pjp.raw_product_id
WHERE pjp.job_id = $1
ORDER BY rp.gtin`, jobID)
if err != nil {
log.Fatalf("query items: %v", err)
}
defer rows.Close()
fmt.Printf("JOB %s\n", jobID)
for rows.Next() {
var gtin, status, errMsg string
var hasPP bool
if err := rows.Scan(&gtin, &status, &errMsg, &hasPP); err != nil {
log.Fatal(err)
}
fmt.Printf("ITEM %s status=%s has_pp=%v err=%s\n", gtin, status, hasPP, errMsg)
}
out := os.Getenv("PROBE_OUT_DIR")
if out == "" {
out = `F:\laragon\www\_MY\descrybe-v2\.codehelper\_process_probe_electronics`
}
_ = os.MkdirAll(out, 0o755)
_ = os.WriteFile(out+string(os.PathSeparator)+"process_id.txt", []byte(jobID.String()+"\n"), 0o644)
}
+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)
}
}
+173
View File
@@ -0,0 +1,173 @@
package main
import (
"context"
"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/logredact"
"github.com/descrybe/descrybe-v2/apps/api/internal/processing"
"github.com/google/uuid"
)
// timeoutCompleter forces enhance deadline so synthesis path is exercised live.
type timeoutCompleter struct{}
func (timeoutCompleter) Complete(context.Context, string, string) (processing.Completion, error) {
return processing.Completion{}, context.DeadlineExceeded
}
func main() {
log.SetOutput(logredact.Writer(os.Stderr))
ean := strings.TrimSpace(os.Getenv("PROBE_EAN"))
if ean == "" {
ean = "4719331854218"
}
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()
pipeline := processing.NewPipeline(pool)
pipeline.BatchSize = cfg.ProcessingBatchSize
pipeline.AI = nil
pipeline.Prompts = aiprompts.NewService(pool)
pipeline.Engine = &processing.Engine{
Completer: timeoutCompleter{},
Vector: processing.NoopVectorCategorizer{},
EPREL: nil,
ProviderMode: processing.AIProviderInternal,
}
var rawID uuid.UUID
if err := pool.QueryRow(ctx, `
SELECT id FROM raw_products
WHERE company_id = $1 AND gtin = $2`, companyID, ean).Scan(&rawID); err != nil {
log.Fatalf("raw product %s: %v", ean, err)
}
log.Printf("ean=%s raw_id=%s", ean, rawID)
_, _ = pool.Exec(ctx, `
UPDATE processed_products
SET processed_description = '', description = '',
field_sources = COALESCE(field_sources, '{}'::jsonb) - 'enhance_input_hash',
updated_at = now()
WHERE company_id = $1 AND product_id = $2`, companyID, ean)
_, _ = pool.Exec(ctx, `
UPDATE processing_jobs
SET status = 'failed', error = 'yielded to thin-synth probe', updated_at = now()
WHERE status IN ('pending','running','processing')`)
jobs, err := pipeline.StartJob(ctx, companyID, userID, []uuid.UUID{rawID}, "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)
defend := make(chan struct{})
go func() {
t := time.NewTicker(300 * time.Millisecond)
defer t.Stop()
for {
select {
case <-defend:
return
case <-ctx.Done():
return
case <-t.C:
_, _ = pool.Exec(context.Background(), `
UPDATE processing_jobs
SET status = 'failed', error = 'yielded to thin-synth probe', updated_at = now()
WHERE status IN ('pending','running','processing') AND id <> $1`, jobID)
_, _ = 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 '%parked%' OR error ILIKE '%yield%' OR error ILIKE '%reclaim%'
OR error ILIKE '%exclusive%' OR error ILIKE '%cleared%' OR error ILIKE '%P0%')`, 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`, 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, 2*time.Minute)
defer runCancel()
if err := pipeline.ProcessJob(runCtx, jobID); err != nil {
close(defend)
log.Fatalf("ProcessJob: %v", err)
}
close(defend)
var name, desc, mode string
var descLen int
if err := pool.QueryRow(ctx, `
SELECT COALESCE(name,''), COALESCE(processed_description,''), length(COALESCE(processed_description,'')),
COALESCE(ai_provider_mode,'')
FROM processed_products
WHERE company_id = $1 AND product_id = $2`, companyID, ean).
Scan(&name, &desc, &descLen, &mode); err != nil {
log.Fatalf("load processed: %v", err)
}
fmt.Printf("EAN %s\n", ean)
fmt.Printf("name=%q\n", name)
fmt.Printf("processed_description_len=%d\n", descLen)
fmt.Printf("processed_description=%q\n", desc)
fmt.Printf("ai_provider_mode=%q\n", mode)
echo := strings.EqualFold(strings.TrimSpace(desc), strings.TrimSpace(name))
pass := descLen > 0 && !echo && !strings.Contains(strings.ToLower(desc), "ready for retail listing")
fmt.Printf("PASS_DESC_NE_TITLE=%v\n", pass)
if !pass {
os.Exit(1)
}
out := os.Getenv("PROBE_OUT_DIR")
if out == "" {
out = `F:\laragon\www\_MY\descrybe-v2\.codehelper\_process_probe_thin_synth`
}
_ = os.MkdirAll(out, 0o755)
_ = os.WriteFile(out+string(os.PathSeparator)+"process_id.txt", []byte(jobID.String()+"\n"), 0o644)
_ = os.WriteFile(out+string(os.PathSeparator)+"score.txt", []byte(fmt.Sprintf(
"ean=%s\nname=%s\npd_len=%d\npd=%s\nmode=%s\nPASS=%v\n", ean, name, descLen, desc, mode, pass,
)), 0o644)
}
+36
View File
@@ -9,6 +9,7 @@ import (
"strconv"
"strings"
"github.com/descrybe/descrybe-v2/apps/api/internal/processing"
"github.com/google/uuid"
"github.com/jackc/pgx/v5"
"github.com/jackc/pgx/v5/pgxpool"
@@ -48,6 +49,7 @@ func migrateCatalogAndFeeds(
migrateProcessedProducts(ctx, mysqlDB, pg, companyMap, feedMap, rawMap, allow, report, dryRun)
backfillProcessedDescriptionsFromMapped(ctx, pg, companyMap, allow, report, dryRun)
backfillProcessedNamesFromMapped(ctx, pg, companyMap, allow, report, dryRun)
backfillProcessedCategoriesFromMapped(ctx, pg, companyMap, allow, report, dryRun)
}
if domains.has("feeds") {
migrateExportFeeds(ctx, mysqlDB, pg, companyMap, feedMap, allow, report, dryRun)
@@ -956,6 +958,40 @@ func backfillProcessedNamesFromMapped(
}
}
// backfillProcessedCategoriesFromMapped fills empty processed.category from mapped unique_ids.
func backfillProcessedCategoriesFromMapped(
ctx context.Context,
pg *pgxpool.Pool,
companyMap map[string]string,
allow map[string]bool,
report map[string]int,
dryRun bool,
) {
if dryRun || pg == nil {
return
}
total := 0
for legacy, cid := range companyMap {
if len(allow) > 0 && !allow[legacy] {
continue
}
u, err := uuid.Parse(cid)
if err != nil {
continue
}
res, err := processing.BackfillProcessedCategoriesFromMapped(ctx, pg, u)
if err != nil {
log.Printf("processed category backfill company=%s: %v", cid, err)
continue
}
total += int(res.Updated)
}
if total > 0 {
report["processed_categories_backfilled"] = total
log.Printf("backfilled %d processed_products.category from mapped_data", total)
}
}
func migrateExportFeeds(
ctx context.Context,
mysqlDB *sql.DB,
@@ -0,0 +1,77 @@
// Command repair-category-prompts replaces legacy A1 / Platform Demo categories.prompt
// HTML marketing formulas with aiprompts.CategoryEnhanceUserTemplate (JSON-compatible
// user overlay: {{name}} {{description}} {{attrs}} {{category}} {{language}}; stores
// both prompt->>'sl' and "*"). Does not touch title_template / description_template
// (formulas are appended as plain-text instructions at enhance render time).
//
// Usage (from apps/api):
//
// go run ./cmd/repair-category-prompts -dry-run
// go run ./cmd/repair-category-prompts -apply
//
// DATABASE_URL / -postgres required. Default is dry-run (count only).
package main
import (
"context"
"encoding/json"
"flag"
"fmt"
"log"
"os"
"strings"
"time"
"github.com/descrybe/descrybe-v2/apps/api/internal/catalog"
"github.com/jackc/pgx/v5/pgxpool"
)
func main() {
postgresURL := flag.String("postgres", os.Getenv("DATABASE_URL"), "Postgres URL (DATABASE_URL)")
dryRun := flag.Bool("dry-run", true, "count rows that would update (default true)")
apply := flag.Bool("apply", false, "write updates (implies not dry-run)")
flag.Parse()
if strings.TrimSpace(*postgresURL) == "" {
log.Fatal("-postgres / DATABASE_URL is required")
}
doDryRun := *dryRun
if *apply {
doDryRun = false
}
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Minute)
defer cancel()
pg, err := pgxpool.New(ctx, *postgresURL)
if err != nil {
log.Fatalf("postgres: %v", err)
}
defer pg.Close()
// Always print dry-run counts first when applying, so operators see the delta.
preview, err := catalog.RepairA1DemoCategoryEnhancePrompts(ctx, pg, true)
if err != nil {
log.Fatalf("dry-run: %v", err)
}
printResult("dry-run", preview)
if doDryRun {
return
}
res, err := catalog.RepairA1DemoCategoryEnhancePrompts(ctx, pg, false)
if err != nil {
log.Fatalf("apply: %v", err)
}
printResult("apply", res)
}
func printResult(phase string, res catalog.RepairCategoryEnhancePromptsResult) {
b, err := json.MarshalIndent(res, "", " ")
if err != nil {
log.Printf("%s: %+v", phase, res)
return
}
fmt.Printf("%s:\n%s\n", phase, string(b))
}
@@ -0,0 +1,27 @@
package main
import (
"context"
"fmt"
"log"
"github.com/descrybe/descrybe-v2/apps/api/internal/processing"
"github.com/google/uuid"
"github.com/jackc/pgx/v5/pgxpool"
)
// backfillA1ProcessedAttributes rewrites A1 processed_products.attributes and
// processed_attributes with AttrsForPersist + category_attributes allowlists
// (same rules as processOne). Use -mode backfill-attributes for polluted local DBs
// where poll already looked clean but the catalog row still stored feed junk.
func backfillA1ProcessedAttributes(ctx context.Context, pg *pgxpool.Pool, companyID uuid.UUID) (int64, error) {
n, err := processing.BackfillCompanyProcessedAttributes(ctx, pg, companyID)
if err != nil {
return 0, fmt.Errorf("A1 attributes backfill: %w", err)
}
return n, nil
}
func logAttributesBackfillResult(updated int64) {
log.Printf("attributes backfill (A1): processed_updated=%d", updated)
}
+68 -31
View File
@@ -12,6 +12,7 @@ import (
"strings"
"unicode"
"github.com/descrybe/descrybe-v2/apps/api/internal/aiprompts"
"github.com/descrybe/descrybe-v2/apps/api/internal/security"
"github.com/google/uuid"
"github.com/jackc/pgx/v5/pgxpool"
@@ -28,8 +29,9 @@ type categoryPromptsFile struct {
}
type categoryPromptEntry struct {
Name string `json:"name"`
Prompt string `json:"prompt"`
Name string `json:"name"`
UniqueID string `json:"unique_id,omitempty"`
Prompt string `json:"prompt"`
}
var (
@@ -145,26 +147,43 @@ func applyCategoryPrompts(ctx context.Context, pg *pgxpool.Pool, companyID uuid.
}
byNorm := make(map[string]string, len(file.Entries))
for _, e := range file.Entries {
name := strings.TrimSpace(e.Name)
prompt := prepareCategoryPrompt(e.Prompt)
if name == "" || prompt == "" {
out.Skipped++
continue
}
key := normalizeCategoryName(name)
if key == "" {
out.Skipped++
continue
}
byNorm[key] = prompt
byUnique := make(map[string]string, len(file.Entries))
// Overlay sets the shared JSON-compatible enhance user template (not legacy HTML).
// Seed JSON selects which categories get a prompt (prefer unique_id, else name).
// Unique title/description formulas stay in title_template / description_template.
// Stored under "sl" + "*" so prompt->>'sl' and LangPromptAny both resolve;
// language stays via {{language}}.
canonical := prepareCategoryPrompt(aiprompts.CategoryEnhanceUserTemplate)
if canonical == "" {
return out, fmt.Errorf("CategoryEnhanceUserTemplate sanitized to empty")
}
if len(byNorm) == 0 {
for _, e := range file.Entries {
uid := strings.TrimSpace(e.UniqueID)
name := strings.TrimSpace(e.Name)
if uid == "" && name == "" {
out.Skipped++
continue
}
if uid != "" {
byUnique[strings.ToLower(uid)] = canonical
}
if name != "" {
key := normalizeCategoryName(name)
if key == "" {
if uid == "" {
out.Skipped++
}
continue
}
byNorm[key] = canonical
}
}
if len(byNorm) == 0 && len(byUnique) == 0 {
return out, fmt.Errorf("no usable category prompts in %s", promptsPath)
}
rows, err := pg.Query(ctx, `
SELECT id, name
SELECT id, COALESCE(unique_id, ''), name
FROM categories
WHERE company_id = $1`, companyID)
if err != nil {
@@ -172,22 +191,34 @@ func applyCategoryPrompts(ctx context.Context, pg *pgxpool.Pool, companyID uuid.
}
defer rows.Close()
ids := make([]uuid.UUID, 0, len(byNorm))
prompts := make([]string, 0, len(byNorm))
matchedKeys := make(map[string]struct{}, len(byNorm))
ids := make([]uuid.UUID, 0, len(byNorm)+len(byUnique))
prompts := make([]string, 0, len(byNorm)+len(byUnique))
matchedNorm := make(map[string]struct{}, len(byNorm))
matchedUID := make(map[string]struct{}, len(byUnique))
for rows.Next() {
var id uuid.UUID
var name string
if err := rows.Scan(&id, &name); err != nil {
var uniqueID, name string
if err := rows.Scan(&id, &uniqueID, &name); err != nil {
return out, err
}
key := normalizeCategoryName(name)
prompt, ok := byNorm[key]
if !ok {
prompt := ""
if uidKey := strings.ToLower(strings.TrimSpace(uniqueID)); uidKey != "" {
if p, ok := byUnique[uidKey]; ok {
prompt = p
matchedUID[uidKey] = struct{}{}
}
}
if prompt == "" {
key := normalizeCategoryName(name)
if p, ok := byNorm[key]; ok {
prompt = p
matchedNorm[key] = struct{}{}
}
}
if prompt == "" {
continue
}
matchedKeys[key] = struct{}{}
ids = append(ids, id)
prompts = append(prompts, prompt)
}
@@ -195,22 +226,28 @@ func applyCategoryPrompts(ctx context.Context, pg *pgxpool.Pool, companyID uuid.
return out, err
}
for key := range byUnique {
if _, ok := matchedUID[key]; !ok {
out.Unmatched = append(out.Unmatched, "uid:"+key)
}
}
for key := range byNorm {
if _, ok := matchedKeys[key]; !ok {
out.Unmatched = append(out.Unmatched, key)
if _, ok := matchedNorm[key]; !ok {
// Name unmatched is noise when unique_id matched the same row; still report for seed hygiene.
out.Unmatched = append(out.Unmatched, "name:"+key)
}
}
sort.Strings(out.Unmatched)
if len(ids) == 0 {
return out, fmt.Errorf("no A1 categories matched any of %d seed prompts (check names)", len(byNorm))
return out, fmt.Errorf("no A1 categories matched any of %d seed prompts (check unique_id/names)", len(byNorm)+len(byUnique))
}
// Single parameterized batch update — company_id gate prevents cross-tenant writes.
// ASSUMPTION: A1 seed company content language is Slovenian ("sl").
// sl + * = prompt->>'sl' and company.LangPromptAny ({{language}} at render).
tag, err := pg.Exec(ctx, `
UPDATE categories AS c
SET prompt = jsonb_build_object('sl', v.prompt), updated_at = now()
SET prompt = jsonb_build_object('sl', v.prompt, '*', v.prompt), updated_at = now()
FROM (
SELECT * FROM unnest($1::uuid[], $2::text[]) AS t(id, prompt)
) AS v
+21 -8
View File
@@ -7,11 +7,13 @@
// part of npm run seed:a1 (skipped on import + cleared after). Use
// -mode recover-jobs only when restoring legacy job history from a MySQL dump.
//
// After reimport, legacy per-category GPT prompts are overlaid from
// scripts/seed/a1-category-prompts.json (Name→Prompt export), matched by
// normalized category name, with v1 placeholders rewritten to {{name}} /
// {{description}}. Use -skip-category-prompts to skip, or
// -mode apply-category-prompts to overlay without a full wipe/reimport.
// After reimport, per-category enhance prompts are overlaid from
// scripts/seed/a1-category-prompts.json (Name list), matched by normalized
// category name, writing aiprompts.CategoryEnhanceUserTemplate (JSON-compatible;
// includes {{attrs}}/{{language}}). Legacy HTML bodies in that JSON are ignored.
// Use -skip-category-prompts to skip, or -mode apply-category-prompts to overlay
// without a full wipe/reimport. For polluted A1+Platform Demo prompts without
// reimport, run: go run ./cmd/repair-category-prompts -apply
//
// Categories: dump/archive store assignment on processed_products.category
// (category unique_id). Feed mappings do not map category. After reimport,
@@ -26,6 +28,9 @@
// exported/imported as-is (skip-processed must not strip them). For a
// polluted local DB without a full wipe, use -mode backfill-categories.
// Fixture EANs are purged from non-A1 tenants automatically.
// Processed attribute junk (poll clean, DB still has zavora/vzmetenje, …): use
// -mode backfill-attributes — same SanitizeProductAttributes + category allowlist
// as processOne/enhance.
//
// Usage:
//
@@ -34,6 +39,7 @@
// go run ./cmd/seed-a1 -mode apply-category-prompts -file ../../scripts/seed/a1-demo-data.sql.gz
// go run ./cmd/seed-a1 -mode recover-jobs -mysql-dump path/to/descrybe_new.sql
// go run ./cmd/seed-a1 -mode backfill-categories -mysql-dump path/to/descrybe_new.sql
// go run ./cmd/seed-a1 -mode backfill-attributes
//
// DATABASE_URL / -postgres required.
// Day-to-day: `npm run seed:a1` (reimport + category prompts + mapped category backfill). Maintainer snapshot: `npm run seed:a1:export`.
@@ -77,7 +83,7 @@ type tableSpec struct {
func main() {
postgresURL := flag.String("postgres", os.Getenv("DATABASE_URL"), "Postgres URL (DATABASE_URL)")
mode := flag.String("mode", "reimport", "export | reimport | recover-jobs | apply-category-prompts | backfill-categories")
mode := flag.String("mode", "reimport", "export | reimport | recover-jobs | apply-category-prompts | backfill-categories | backfill-attributes")
file := flag.String("file", "", "gzipped seed archive path (required for export/reimport)")
mysqlDump := flag.String("mysql-dump", os.Getenv("SEED_A1_MYSQL_DUMP"), "mysqldump path for recover-jobs / backfill-categories (or SEED_A1_MYSQL_DUMP)")
company := flag.String("company", defaultA1CompanyID, "Postgres companies.id for A1")
@@ -95,7 +101,8 @@ func main() {
}
modeVal := strings.ToLower(strings.TrimSpace(*mode))
if modeVal != "recover-jobs" && modeVal != "apply-category-prompts" && modeVal != "backfill-categories" && strings.TrimSpace(*file) == "" {
needsFile := modeVal != "recover-jobs" && modeVal != "apply-category-prompts" && modeVal != "backfill-categories" && modeVal != "backfill-attributes"
if needsFile && strings.TrimSpace(*file) == "" {
log.Fatal("-file is required (e.g. ../../scripts/seed/a1-demo-data.sql.gz)")
}
dumpPath := resolveMySQLDumpPath(*mysqlDump)
@@ -184,13 +191,19 @@ func main() {
}
res.PurgedOtherRaw, res.PurgedOtherPP = rawN, ppN
logCategoryBackfillResult(res, dumpPath)
case "backfill-attributes":
n, err := backfillA1ProcessedAttributes(ctx, pg, companyID)
if err != nil {
log.Fatalf("backfill-attributes: %v", err)
}
logAttributesBackfillResult(n)
case "recover-jobs":
if err := recoverJobsFromMySQLDump(ctx, pg, companyID, dumpPath); err != nil {
log.Fatalf("recover-jobs: %v", err)
}
log.Printf("recovered A1 processing jobs into company %s from %s", companyID, dumpPath)
default:
log.Fatalf("unknown -mode %q (want export|reimport|recover-jobs|apply-category-prompts|backfill-categories)", *mode)
log.Fatalf("unknown -mode %q (want export|reimport|recover-jobs|apply-category-prompts|backfill-categories|backfill-attributes)", *mode)
}
}