fix

2026-08-17 01:30:28 +02:00
parent c2429873b3
commit 0dceb3a404
18 changed files with 1021 additions and 219 deletions
@@ -0,0 +1,236 @@
+// Temporary local probe for A1/Demo V1 projection cleanup. Do not commit.
+package main
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"os"
+	"path/filepath"
+	"time"
+
+	"github.com/descrybe/descrybe-v2/apps/api/internal/billing"
+	"github.com/descrybe/descrybe-v2/apps/api/internal/processing"
+	"github.com/google/uuid"
+	"github.com/jackc/pgx/v5/pgxpool"
+)
+
+func main() {
+	ctx, cancel := context.WithTimeout(context.Background(), 2*time.Minute)
+	defer cancel()
+	dsn := os.Getenv("DATABASE_URL")
+	if dsn == "" {
+		dsn = "postgres://descrybe:descrybe@localhost:5433/descrybe?sslmode=disable"
+	}
+	pool, err := pgxpool.New(ctx, dsn)
+	if err != nil {
+		fatal(err)
+	}
+	defer pool.Close()
+
+	outDir := filepath.Join("..", "..", ".codehelper", "_a1_v1_response_20260817")
+	if abs, err := filepath.Abs(outDir); err == nil {
+		outDir = abs
+	}
+	_ = os.MkdirAll(outDir, 0o755)
+
+	type companyRow struct {
+		ID     uuid.UUID
+		Name   string
+		Legacy string
+	}
+	companies := []companyRow{}
+	rows, err := pool.Query(ctx, `
+		SELECT id, name, COALESCE(legacy_company_id, '')
+		FROM companies
+		WHERE legacy_company_id = $1
+		   OR lower(name) LIKE '%demo%'
+		ORDER BY CASE WHEN legacy_company_id = $1 THEN 0 ELSE 1 END, name
+		LIMIT 6`, billing.A1LegacyCompanyID)
+	if err != nil {
+		fatal(err)
+	}
+	for rows.Next() {
+		var c companyRow
+		if err := rows.Scan(&c.ID, &c.Name, &c.Legacy); err != nil {
+			fatal(err)
+		}
+		companies = append(companies, c)
+	}
+	rows.Close()
+	if len(companies) == 0 {
+		fatal(fmt.Errorf("no A1/Demo companies found"))
+	}
+
+	p := processing.NewPipeline(pool)
+	eans := []string{"8806091734365", "8422160053580"}
+	summary := map[string]any{"generated_at": time.Now().UTC().Format(time.RFC3339), "companies": []any{}}
+
+	for _, co := range companies {
+		omitMeta := billing.IsA1CohortCompany(co.Legacy, co.Name)
+		coSummary := map[string]any{
+			"company_id":   co.ID.String(),
+			"name":         co.Name,
+			"legacy_id":    co.Legacy,
+			"omit_seo_meta": omitMeta,
+			"items":        []any{},
+		}
+
+		// Prefer existing processed rows for the sample EANs; fall back to any 2 recent.
+		type prod struct {
+			EAN, Title, Desc, Cat, CatName string
+			Attrs                          []byte
+		}
+		prods := []prod{}
+		q := `
+			SELECT COALESCE(p.product_id,''), COALESCE(p.processed_name,p.name,''),
+				COALESCE(p.processed_description, p.description, ''),
+				COALESCE(p.category,''), COALESCE(c.name,''),
+				COALESCE(p.processed_attributes, p.attributes, '{}'::jsonb)
+			FROM processed_products p
+			LEFT JOIN categories c ON c.unique_id = p.category AND c.company_id = p.company_id
+			WHERE p.company_id = $1 AND p.product_id = ANY($2)
+			ORDER BY p.updated_at DESC NULLS LAST
+			LIMIT 4`
+		pr, err := pool.Query(ctx, q, co.ID, eans)
+		if err != nil {
+			fatal(err)
+		}
+		for pr.Next() {
+			var x prod
+			if err := pr.Scan(&x.EAN, &x.Title, &x.Desc, &x.Cat, &x.CatName, &x.Attrs); err != nil {
+				fatal(err)
+			}
+			prods = append(prods, x)
+		}
+		pr.Close()
+		if len(prods) < 2 {
+			pr2, err := pool.Query(ctx, `
+				SELECT COALESCE(p.product_id,''), COALESCE(p.processed_name,p.name,''),
+					COALESCE(p.processed_description, p.description, ''),
+					COALESCE(p.category,''), COALESCE(c.name,''),
+					COALESCE(p.processed_attributes, p.attributes, '{}'::jsonb)
+				FROM processed_products p
+				LEFT JOIN categories c ON c.unique_id = p.category AND c.company_id = p.company_id
+				WHERE p.company_id = $1 AND COALESCE(p.processed_name,p.name,'') <> ''
+				ORDER BY p.updated_at DESC NULLS LAST
+				LIMIT 2`, co.ID)
+			if err != nil {
+				fatal(err)
+			}
+			for pr2.Next() {
+				var x prod
+				if err := pr2.Scan(&x.EAN, &x.Title, &x.Desc, &x.Cat, &x.CatName, &x.Attrs); err != nil {
+					fatal(err)
+				}
+				prods = append(prods, x)
+			}
+			pr2.Close()
+		}
+		if len(prods) == 0 {
+			coSummary["note"] = "no processed products"
+			summary["companies"] = append(summary["companies"].([]any), coSummary)
+			continue
+		}
+
+		items := []processing.V1ProcessJobItem{}
+		for _, x := range prods {
+			attrs := map[string]any{}
+			_ = json.Unmarshal(x.Attrs, &attrs)
+			item := processing.V1ProcessJobItem{
+				"ean":           x.EAN,
+				"status":        "processed",
+				"title":         x.Title,
+				"name":          x.Title,
+				"description":   x.Desc,
+				"category":      x.Cat,
+				"category_name": x.CatName,
+				"attributes":    attrs,
+				"eprel":         nil,
+				"id":            uuid.New().String(),
+			}
+			item["processed_product_id"] = item["id"]
+			item["raw_product_id"] = uuid.New().String()
+			if !omitMeta {
+				item["meta_title"] = nil
+				item["meta_description"] = nil
+			}
+			item = processing.EnforceV1ProcessCompletedItemOpts(item, processing.EnforceV1Opts{
+				Language:    "sl",
+				Allowed:     map[string]struct{}{},
+				OmitSEOMeta: omitMeta,
+			})
+			items = append(items, item)
+		}
+
+		payload := map[string]any{
+			"data": map[string]any{
+				"status":          "COMPLETED",
+				"processing_type": "full",
+				"total_items":     len(items),
+				"items":           items,
+				"company":         co.Name,
+				"omit_seo_meta":   omitMeta,
+			},
+		}
+		b, err := json.MarshalIndent(payload, "", "  ")
+		if err != nil {
+			fatal(err)
+		}
+		label := "demo"
+		if omitMeta {
+			label = "a1"
+		}
+		fname := fmt.Sprintf("%s_%s.json", label, co.ID.String()[:8])
+		path := filepath.Join(outDir, fname)
+		if err := os.WriteFile(path, b, 0o644); err != nil {
+			fatal(err)
+		}
+		coSummary["sample_path"] = path
+		coSummary["item_count"] = len(items)
+		coSummary["items"] = items
+		summary["companies"] = append(summary["companies"].([]any), coSummary)
+		fmt.Println("wrote", path, "items=", len(items), "omit_meta=", omitMeta)
+
+		// Also exercise LoadV1 against the latest completed job for this company when present.
+		var jobID uuid.UUID
+		err = pool.QueryRow(ctx, `
+			SELECT id FROM processing_jobs
+			WHERE company_id = $1 AND status IN ('completed','COMPLETED')
+			ORDER BY updated_at DESC NULLS LAST
+			LIMIT 1`, co.ID).Scan(&jobID)
+		if err == nil {
+			loaded, loadErr := p.LoadV1ProcessJobItems(ctx, co.ID, jobID, "full")
+			if loadErr == nil && len(loaded) > 0 {
+				if len(loaded) > 2 {
+					loaded = loaded[:2]
+				}
+				live := map[string]any{
+					"data": map[string]any{
+						"process_id":      jobID.String(),
+						"status":          "COMPLETED",
+						"processing_type": "full",
+						"total_items":     len(loaded),
+						"items":           loaded,
+						"company":         co.Name,
+						"omit_seo_meta":   omitMeta,
+						"source":          "LoadV1ProcessJobItems",
+					},
+				}
+				lb, _ := json.MarshalIndent(live, "", "  ")
+				lpath := filepath.Join(outDir, fmt.Sprintf("%s_live_job_%s.json", label, jobID.String()[:8]))
+				_ = os.WriteFile(lpath, lb, 0o644)
+				fmt.Println("wrote", lpath, "live_items=", len(loaded))
+			}
+		}
+	}
+
+	sb, _ := json.MarshalIndent(summary, "", "  ")
+	_ = os.WriteFile(filepath.Join(outDir, "summary.json"), sb, 0o644)
+	fmt.Println("done", outDir)
+}
+
+func fatal(err error) {
+	fmt.Fprintln(os.Stderr, err)
+	os.Exit(1)
+}
@@ -0,0 +1,84 @@
+package main
+
+import (
+	"encoding/json"
+	"fmt"
+	"os"
+	"path/filepath"
+
+	"github.com/descrybe/descrybe-v2/apps/api/internal/processing"
+)
+
+func main() {
+	badItems := []processing.V1ProcessJobItem{
+		{
+			"ean":                  "8806091734365",
+			"status":               "processed",
+			"title":                "LG Ameriški hladilnikGSXV80PZLE",
+			"name":                 "LG Ameriški hladilnikGSXV80PZLE",
+			"category":             "11",
+			"category_name":        "Hladilniki",
+			"description":          "<h1>LG Ameriški hladilnikGSXV80PZLE</h1><p>LG Ameriški hladilnikGSXV80PZLE je izdelek v kategoriji Hladilniki znamke LG. Ključne specifikacije: width 179, height 91.3, depth 73.5.</p><h2>Ključne lastnosti</h2><ul><li>width 179</li><li>height 91.3</li><li>depth 73.5</li></ul><p>LG Ameriški hladilnikGSXV80PZLE je izdelek v kategoriji Hladilniki znamke LG. Ključne specifikacije: width 179, height 91.3, depth 73.5.</p><p>LG Ameriški hladilnikGSXV80PZLE je izdelek v kategoriji Hladilniki znamke LG. Ključne specifikacije: width 179, height 91.3, depth 73.5.</p><p>LG Ameriški hladilnikGSXV80PZLE je izdelek v kategoriji Hladilniki znamke LG. Ključne specifikacije: width 179, height 91.3, depth 73.5.</p>",
+			"attributes":           map[string]any{"brand": "LG", "depth": "73.5", "energy_class": "E", "height": "91.3", "warranty": "24 mesecev", "width": "179"},
+			"meta_title":           "LG Ameriški hladilnikGSXV80PZLE | Hladilniki",
+			"meta_description":     "x",
+			"eprel":                map[string]any{"id": "1158009", "energy_class": "E"},
+			"id":                   "1bc20e89-3779-4c66-8617-8e72555cd363",
+			"processed_product_id": "1bc20e89-3779-4c66-8617-8e72555cd363",
+			"raw_product_id":       "fff13ee3-e78c-4185-b524-b64128bbc38b",
+			"main_image":           "https://example/lg.jpg",
+			"more_images":          nil,
+		},
+		{
+			"ean":                  "8422160053580",
+			"status":               "processed",
+			"title":                "Ufesa digitalni zračni cvrtnik 23L Magnum",
+			"name":                 "Ufesa digitalni zračni cvrtnik 23L Magnum",
+			"category":             "112",
+			"category_name":        "Ostali aparati",
+			"description":          "<h1>Ufesa digitalni zračni cvrtnik 23L Magnum</h1><p>Ufesa digitalni zračni cvrtnik 23L Magnum je izdelek v kategoriji Ostali aparati znamke UFESA. Ključne specifikacije: width 0,00m, height 0,00m, depth 0,00m.</p><h2>Ključne lastnosti</h2><ul><li>width 0,00m</li><li>height 0,00m</li><li>depth 0,00m</li><li>weight 11,0000kg</li></ul><p>Ufesa digitalni zračni cvrtnik 23L Magnum je izdelek v kategoriji Ostali aparati znamke UFESA. Ključne specifikacije: width 0,00m, height 0,00m, depth 0,00m.</p>",
+			"attributes":           map[string]any{"brand": "UFESA", "depth": "0,00m", "height": "0,00m", "warranty": "24 mesecev", "weight": "11,0000kg", "width": "0,00m"},
+			"meta_title":           "x",
+			"meta_description":     "y",
+			"eprel":                nil,
+			"id":                   "c708b505-dbf0-40c5-ba09-022d15daff97",
+			"processed_product_id": "c708b505-dbf0-40c5-ba09-022d15daff97",
+			"raw_product_id":       "fff6c685-931b-4fa5-846b-89b2503a5a77",
+			"main_image":           "https://example/ufesa.jpg",
+			"more_images":          []string{"https://example/2.jpg"},
+		},
+	}
+	cleaned := make([]processing.V1ProcessJobItem, 0, len(badItems))
+	for _, it := range badItems {
+		cleaned = append(cleaned, processing.EnforceV1ProcessCompletedItemOpts(it, processing.EnforceV1Opts{
+			Language:    "sl",
+			Allowed:     map[string]struct{}{},
+			OmitSEOMeta: true,
+		}))
+	}
+	out := map[string]any{
+		"data": map[string]any{
+			"process_id":      "13b97eab-2329-4735-82fa-3afe263849fa",
+			"status":          "COMPLETED",
+			"processing_type": "full",
+			"total_items":     len(cleaned),
+			"items":           cleaned,
+			"note":            "projection cleanup of production sample (A1 OmitSEOMeta)",
+		},
+	}
+	b, err := json.MarshalIndent(out, "", "  ")
+	if err != nil {
+		panic(err)
+	}
+	path := filepath.Join("..", "..", ".codehelper", "_a1_v1_response_20260817", "production_sample_cleaned.json")
+	if err := os.WriteFile(path, b, 0o644); err != nil {
+		panic(err)
+	}
+	fmt.Println("wrote", path)
+	for _, it := range cleaned {
+		fmt.Printf("ean=%v\ntitle=%v\ncategory=%v category_id=%v\nmeta_present=%v/%v\nattrs=%v\ndesc=%v\n\n",
+			it["ean"], it["title"], it["category"], it["category_id"],
+			it["meta_title"] != nil, it["meta_description"] != nil,
+			it["attributes"], it["description"])
+	}
+}