fix

2026-08-16 21:35:48 +02:00
parent 106f602232
commit 389608ea56
28 changed files with 2491 additions and 115 deletions
@@ -0,0 +1,99 @@
+package main
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"os"
+	"strings"
+	"time"
+
+	"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/processing"
+	"github.com/google/uuid"
+)
+
+func main() {
+	out := `F:\laragon\www\_MY\descrybe-v2\.codehelper\_live_llm_formula_20260816`
+	jobID := uuid.MustParse("3d926dfa-de8a-4afc-b1dd-8707633c673e")
+	cfg, err := config.Load()
+	if err != nil {
+		panic(err)
+	}
+	ctx, cancel := context.WithTimeout(context.Background(), 60*time.Second)
+	defer cancel()
+	pool, err := db.NewPool(ctx, cfg.DatabaseURL, db.PoolOptions{
+		MaxConns: 4, MinConns: 0,
+		MaxConnLifetime: time.Hour, HealthCheckPeriod: 30 * time.Second,
+	})
+	if err != nil {
+		panic(err)
+	}
+	defer pool.Close()
+
+	rows, err := pool.Query(ctx, `
+		SELECT rp.gtin,
+		       COALESCE(pp.processed_name, pp.name, '') AS title,
+		       COALESCE(NULLIF(pp.processed_description, ''), pp.description, '') AS description,
+		       COALESCE(pp.category, '') AS category,
+		       COALESCE(c.name, '') AS category_name,
+		       c.description_template,
+		       pp.id
+		FROM processing_job_products pjp
+		JOIN raw_products rp ON rp.id = pjp.raw_product_id
+		LEFT JOIN processed_products pp ON pp.id = pjp.processed_product_id
+		LEFT JOIN categories c ON c.company_id = pp.company_id AND c.unique_id = pp.category
+		WHERE pjp.job_id = $1
+		ORDER BY rp.gtin`, jobID)
+	if err != nil {
+		panic(err)
+	}
+	defer rows.Close()
+	products := []map[string]any{}
+	for rows.Next() {
+		var gtin, title, desc, cat, catName string
+		var tmpl any
+		var ppID *uuid.UUID
+		if err := rows.Scan(&gtin, &title, &desc, &cat, &catName, &tmpl, &ppID); err != nil {
+			panic(err)
+		}
+		idStr := ""
+		if ppID != nil {
+			idStr = ppID.String()
+		}
+		lower := strings.ToLower(desc)
+		products = append(products, map[string]any{
+			"ean":                  gtin,
+			"product_id":           idStr,
+			"title":                title,
+			"description":          desc,
+			"category":             cat,
+			"category_name":        catName,
+			"description_template": tmpl,
+			"formula_constraint":   processing.FormatDescriptionFormulaConstraint(tmpl),
+			"has_h1":               strings.Contains(lower, "<h1"),
+			"has_h2":               strings.Contains(lower, "<h2"),
+			"has_p":                strings.Contains(lower, "<p"),
+			"has_ul":               strings.Contains(lower, "<ul"),
+			"desc_len":             len(desc),
+			"desc_ne_title":        strings.TrimSpace(desc) != "" && strings.TrimSpace(desc) != strings.TrimSpace(title),
+		})
+	}
+	raw, _ := json.MarshalIndent(products, "", "  ")
+	_ = os.WriteFile(out+`\db_products.json`, raw, 0o644)
+	fmt.Printf("wrote %d products\n", len(products))
+	for _, p := range products {
+		t := p["title"].(string)
+		if len(t) > 40 {
+			t = t[:40]
+		}
+		fmt.Printf("%s cat=%s/%s h1=%v h2=%v p=%v ul=%v desc_ne_title=%v len=%v title=%q\n",
+			p["ean"], p["category"], p["category_name"], p["has_h1"], p["has_h2"], p["has_p"], p["has_ul"], p["desc_ne_title"], p["desc_len"], t)
+		d := p["description"].(string)
+		if len(d) > 220 {
+			d = d[:220]
+		}
+		fmt.Printf("  desc_preview=%q\n", d)
+	}
+}