feat: fuzzy matching Picnic ↔ Leclerc + page /matches dans le dashboard

Nouvelle table product_matches (status: pending/validated/rejected).
Matching via RapidFuzz token_sort_ratio, seuil configurable (défaut 85%).

Workflow :
  1. python -m tickettracker.cli match [--threshold 85]
     → calcule et stocke les paires candidates
  2. http://localhost:8000/matches
     → l'utilisateur valide ou rejette chaque paire
  3. La comparaison de prix enrichie avec les paires validées

Nouvelles dépendances : rapidfuzz, watchdog (requirements.txt).
10 tests ajoutés (test_matcher.py), tous passent.
Suite complète : 129 passent, 1 xfail, 0 échec.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
2026-02-25 18:02:48 +01:00
parent f360332626
commit be4d4a7076
13 changed files with 804 additions and 23 deletions

View File

@@ -83,13 +83,18 @@ def get_monthly_spending(conn: sqlite3.Connection) -> list[dict]:
def get_compare_prices(conn: sqlite3.Connection) -> list[dict]:
"""Comparaison de prix entre Picnic et Leclerc pour les produits communs.
Utilise la vue price_history. Ne retourne que les produits présents
dans les deux enseignes. Trié par écart décroissant (le plus cher en premier).
Combine deux sources :
- Correspondances exactes (même name_normalized dans les deux enseignes)
- Correspondances fuzzy validées dans product_matches (status='validated')
Les doublons éventuels (un produit déjà en exact ET en fuzzy) sont éliminés
par UNION (qui déduplique) + sélection par nom picnic.
Returns:
Liste de dicts {name, price_picnic, price_leclerc, diff, diff_pct}.
diff = price_leclerc - price_picnic (positif = Leclerc plus cher)
diff_pct = diff / MIN(price_picnic, price_leclerc) * 100
Liste de dicts {name, price_picnic, price_leclerc, diff, diff_pct, match_type}.
diff = price_leclerc - price_picnic (positif = Leclerc plus cher)
diff_pct = diff / MIN(price_picnic, price_leclerc) * 100
match_type = 'exact' ou 'fuzzy'
"""
rows = conn.execute(
"""
@@ -101,32 +106,67 @@ def get_compare_prices(conn: sqlite3.Connection) -> list[dict]:
FROM price_history
WHERE name_normalized IS NOT NULL
GROUP BY name_normalized, store
),
exact_matches AS (
SELECT
a.name_normalized AS name,
a.name_normalized AS name_display,
a.avg_price AS price_picnic,
b.avg_price AS price_leclerc,
ROUND(b.avg_price - a.avg_price, 2) AS diff,
ROUND(
(b.avg_price - a.avg_price)
/ MIN(a.avg_price, b.avg_price) * 100
, 1) AS diff_pct,
'exact' AS match_type
FROM avg_by_store a
JOIN avg_by_store b
ON a.name_normalized = b.name_normalized
AND a.store = 'picnic'
AND b.store = 'leclerc'
),
fuzzy_matches AS (
SELECT
pm.name_picnic AS name,
pm.name_picnic || '' || pm.name_leclerc AS name_display,
ap_p.avg_price AS price_picnic,
ap_l.avg_price AS price_leclerc,
ROUND(ap_l.avg_price - ap_p.avg_price, 2) AS diff,
ROUND(
(ap_l.avg_price - ap_p.avg_price)
/ MIN(ap_p.avg_price, ap_l.avg_price) * 100
, 1) AS diff_pct,
'fuzzy' AS match_type
FROM product_matches pm
JOIN avg_by_store ap_p
ON ap_p.name_normalized = pm.name_picnic AND ap_p.store = 'picnic'
JOIN avg_by_store ap_l
ON ap_l.name_normalized = pm.name_leclerc AND ap_l.store = 'leclerc'
WHERE pm.status = 'validated'
-- Exclure si déjà présent en exact match
AND pm.name_picnic NOT IN (SELECT name FROM exact_matches)
)
SELECT
a.name_normalized AS name,
a.avg_price AS price_picnic,
b.avg_price AS price_leclerc,
ROUND(b.avg_price - a.avg_price, 2) AS diff,
ROUND(
(b.avg_price - a.avg_price)
/ MIN(a.avg_price, b.avg_price) * 100
, 1) AS diff_pct
FROM avg_by_store a
JOIN avg_by_store b
ON a.name_normalized = b.name_normalized
AND a.store = 'picnic'
AND b.store = 'leclerc'
ORDER BY ABS(b.avg_price - a.avg_price) DESC
SELECT name, name_display, price_picnic, price_leclerc, diff, diff_pct, match_type
FROM (
SELECT name, name_display, price_picnic, price_leclerc, diff, diff_pct, match_type
FROM exact_matches
UNION ALL
SELECT name, name_display, price_picnic, price_leclerc, diff, diff_pct, match_type
FROM fuzzy_matches
)
ORDER BY ABS(diff) DESC
"""
).fetchall()
return [
{
"name": r["name"],
"name_display": r["name_display"],
"price_picnic": r["price_picnic"],
"price_leclerc": r["price_leclerc"],
"diff": r["diff"],
"diff_pct": r["diff_pct"],
"match_type": r["match_type"],
}
for r in rows
]
@@ -279,6 +319,91 @@ def get_receipt_detail(conn: sqlite3.Connection, receipt_id: int) -> dict | None
}
def get_pending_matches(conn: sqlite3.Connection) -> list[dict]:
"""Paires en attente de validation, avec prix moyens des deux enseignes.
Returns:
Liste de dicts {id, name_picnic, price_picnic, name_leclerc, price_leclerc, score}.
price_picnic / price_leclerc : prix moyen unitaire de ce produit dans la vue
price_history (None si aucune occurrence pour ce nom normalisé).
"""
rows = conn.execute(
"""
SELECT
pm.id,
pm.name_picnic,
pm.name_leclerc,
pm.score,
ROUND(AVG(CASE WHEN ph.store='picnic' THEN ph.unit_price END), 2) AS price_picnic,
ROUND(AVG(CASE WHEN ph.store='leclerc' THEN ph.unit_price END), 2) AS price_leclerc
FROM product_matches pm
LEFT JOIN price_history ph
ON ph.name_normalized IN (pm.name_picnic, pm.name_leclerc)
WHERE pm.status = 'pending'
GROUP BY pm.id
ORDER BY pm.score DESC
"""
).fetchall()
return [
{
"id": r["id"],
"name_picnic": r["name_picnic"],
"name_leclerc": r["name_leclerc"],
"score": r["score"],
"price_picnic": r["price_picnic"],
"price_leclerc": r["price_leclerc"],
}
for r in rows
]
def get_validated_matches(conn: sqlite3.Connection) -> list[dict]:
"""Paires validées pour enrichir get_compare_prices.
Returns:
Liste de dicts {name_picnic, price_picnic, name_leclerc, price_leclerc, diff, diff_pct}.
"""
rows = conn.execute(
"""
WITH avg_prices AS (
SELECT name_normalized, store, ROUND(AVG(unit_price), 2) AS avg_price
FROM price_history
WHERE name_normalized IS NOT NULL
GROUP BY name_normalized, store
)
SELECT
pm.id,
pm.name_picnic,
pm.name_leclerc,
ap_p.avg_price AS price_picnic,
ap_l.avg_price AS price_leclerc,
ROUND(ap_l.avg_price - ap_p.avg_price, 2) AS diff,
ROUND(
(ap_l.avg_price - ap_p.avg_price)
/ MIN(ap_p.avg_price, ap_l.avg_price) * 100
, 1) AS diff_pct
FROM product_matches pm
JOIN avg_prices ap_p ON ap_p.name_normalized = pm.name_picnic AND ap_p.store = 'picnic'
JOIN avg_prices ap_l ON ap_l.name_normalized = pm.name_leclerc AND ap_l.store = 'leclerc'
WHERE pm.status = 'validated'
ORDER BY ABS(ap_l.avg_price - ap_p.avg_price) DESC
"""
).fetchall()
return [
{
"name_picnic": r["name_picnic"],
"name_leclerc": r["name_leclerc"],
"price_picnic": r["price_picnic"],
"price_leclerc": r["price_leclerc"],
"diff": r["diff"],
"diff_pct": r["diff_pct"],
}
for r in rows
]
def get_product_list(conn: sqlite3.Connection) -> list[str]:
"""Liste tous les noms normalisés distincts (non NULL) pour le sélecteur.