Update app.py

app.py

@@ -1,268 +1,280 @@
-# app.py — Formulaire EAN/Libellé + logs (UI & Space) — style "IPC / Insee"
-import io, sys, json
-import gradio as gr
-from contextlib import redirect_stdout, redirect_stderr
-
-# 👉 adapte le module si ton fichier agent a un autre nom
-from quick_deploy_agent import build_agent, parse_result
-
-
-# --- en haut de app.py, ajoute cet util ---
-import re
-
-def _extract_json(text: str):
-    """
-    Essaie de parser un JSON valable depuis une chaîne.
-    - d'abord json.loads direct
-    - sinon, extrait le contenu d'un bloc ```json ... ``` puis parse
-    - renvoie None si rien de valable
-    """
-    if isinstance(text, dict):
-        return text
-    if not isinstance(text, str):
-        return None
-    # 1) direct
-    try:
-        return json.loads(text)
-    except Exception:
-        pass
-    # 2) dans des backticks
-    try:
-        m = re.search(r"```(?:json)?\s*(\{[\s\S]*\})\s*```", text)
-        if m:
-            return json.loads(m.group(1))
-    except Exception:
-        pass
-    return None
-
-
-# ---------- util "tee" pour dupliquer les logs ----------
-class _Tee(io.TextIOBase):
-    def __init__(self, *streams): self.streams = streams
-    def write(self, s):
-        for st in self.streams:
-            try: st.write(s)
-            except Exception: pass
-        return len(s)
-    def flush(self):
-        for st in self.streams:
-            try: st.flush()
-            except Exception: pass
-
-# ---------- instancie l'agent + logs verbeux ----------
-agent = build_agent()
-if hasattr(agent, "verbosity_level"):
-    agent.verbosity_level = 3  # 0..3 (3 = très verbeux)
-
-# ---------- prompt interne (l'utilisateur ne tape rien) ----------
-TASK_TMPL = """\
-Classe ce produit en COICOP:
-EAN: {ean}
-Libellé: {label}
-
-Outils autorisés UNIQUEMENT : validate_ean, openfoodfacts_product_by_ean,
-map_off_to_coicop, coicop_regex_rules, coicop_semantic_similarity, resolve_coicop_candidates.
-N'UTILISE PAS python_interpreter. N'ÉCRIS PAS DE CODE. N'INDEXE JAMAIS la sortie d'un tool.
-
-Pipeline :
-1) validate_ean(ean)
-2) openfoodfacts_product_by_ean(ean)
-3) map_off_to_coicop(off_payload=<sortie brute de (2)>) ou, si nécessaire,
-   map_off_to_coicop(product_name, categories_tags, ingredients_text)
-4) coicop_regex_rules(text=LIBELLÉ UTILISATEUR)
-5) coicop_semantic_similarity(text=LIBELLÉ UTILISATEUR, topk=5)
-6) resolve_coicop_candidates([...], topn=3)
-
-Règles strictes :
-- Les outils renvoient des objets Python (dict/list). Tu peux indexer directement.
-- Pour (4) et (5), utilise le libellé utilisateur (pas besoin de lire la réponse d’OFF).
-- Retourne uniquement un JSON valide (objet), sans backticks.
-"""
-
-
-def classify(label: str, ean: str):
-    import re
-    label = (label or "").strip()
-    ean = (ean or "").strip()
-
-    if not label and not ean:
-        return json.dumps({"error": "Veuillez saisir un libellé ou un EAN."}, ensure_ascii=False, indent=2), "—"
-
-    # Construire le message pour l’agent
-    task = TASK_TMPL.format(ean=ean or "N/A", label=label or "N/A")
-
-    # Buffers UI + duplication vers logs Space (stdout/stderr)
-    buf_out, buf_err = io.StringIO(), io.StringIO()
-    tee_out, tee_err = _Tee(sys.stdout, buf_out), _Tee(sys.stderr, buf_err)
-
-    # Utilitaire local : extraire un vrai JSON d'une chaîne (gère les ```json ... ```)
-    def _extract_json(text: str):
-        if isinstance(text, dict):
-            return text
-        if not isinstance(text, str):
-            return None
-        # 1) direct
-        try:
-            return json.loads(text)
-        except Exception:
-            pass
-        # 2) JSON dans des backticks
-        try:
-            m = re.search(r"```(?:json)?\s*(\{[\s\S]*\})\s*```", text)
-            if m:
-                return json.loads(m.group(1))
-        except Exception:
-            pass
-        return None
-
-    print("\n===== Agent run start =====")
-    print(task)
-
-    try:
-        # Exécuter l’agent en dupliquant les logs vers l’UI et le Space
-        with redirect_stdout(tee_out), redirect_stderr(tee_err):
-            res = agent.run(task)
-
-        # Parsing robuste du résultat
-        obj = None
-        if isinstance(res, dict):
-            obj = res
-        if obj is None:
-            obj = _extract_json(res)
-        if obj is None:
-            try:
-                obj = parse_result(res)  # ton parseur existant
-            except Exception:
-                obj = None
-        if obj is None:
-            obj = {"raw": str(res)}  # dernier recours
-
-        logs_ui = "\n".join(
-            s for s in [buf_out.getvalue().strip(), buf_err.getvalue().strip()] if s
-        ) or "(aucun log)"
-
-        print("===== Agent run end =====\n")
-        return json.dumps(obj, ensure_ascii=False, indent=2), logs_ui
-
-    except Exception as e:
-        logs_ui = "\n".join(
-            s for s in [buf_out.getvalue().strip(), buf_err.getvalue().strip()] if s
-        ) or "(aucun log)"
-        print(f"✖ Agent error: {e}")
-        return json.dumps({"error": str(e)}, ensure_ascii=False, indent=2), logs_ui
-
-
-def fill_example():
-    # EAN réel OFF (Les p'tits crémeux – Aldi – 216 g)
-    return "Camembert au lait cru AOP 250g - ALDI", "2006050033638"
-
-# ---------- thème & CSS (style IPC / Insee) ----------
-theme = gr.themes.Soft(
-    primary_hue="blue",
-    secondary_hue="indigo",
-    neutral_hue="slate",
-)
-
-custom_css = """
-:root{
-  --insee-primary: #89c2d9;
-  --insee-primary-700:#89c2d9;
-  --insee-accent: #ff5c35;
-  --insee-neutral-50:#f8fafc;
-  --insee-neutral-100:#f1f5f9;
-  --insee-neutral-900:#0f172a;
-  --radius-lg: 16px;
-}
-.gradio-container{
-  font-family: Inter, system-ui, -apple-system, Segoe UI, Roboto, Ubuntu, Cantarell, "Helvetica Neue", Arial, "Noto Sans", sans-serif;
-  background: linear-gradient(180deg, var(--insee-neutral-50), #ffffff 30%);
-}
-#app-header{
-  background: linear-gradient(90deg, var(--insee-primary), var(--insee-primary-700));
-  color: white; padding: 18px 22px; border-radius: 18px;
-  box-shadow: 0 10px 30px rgba(11,61,145,0.25); margin-bottom: 16px;
-}
-#app-header h1{ margin:0; font-size: 22px; letter-spacing: .3px; }
-#app-header p{ margin:6px 0 0; opacity:.92 }
-
-.card{
-  background: #fff; border: 1px solid var(--insee-neutral-100); border-radius: var(--radius-lg);
-  box-shadow: 0 8px 24px rgba(15, 23, 42, 0.06);
-  padding: 18px;
-}
-
-.gr-textbox textarea{
-  font-size: 16px; line-height: 1.35; padding: 14px 16px;
-  border: 2px solid #e5e7eb; border-radius: 14px;
-}
-.gr-textbox textarea:focus{
-  outline: 3px solid color-mix(in oklab, var(--insee-primary) 22%, white);
-  border-color: var(--insee-primary);
-  box-shadow: 0 0 0 4px rgba(11,61,145,0.15);
-}
-
-.gr-button{
-  font-weight: 700; letter-spacing: .2px; border-radius: 14px;
-  transition: transform .06s ease, box-shadow .06s ease;
-}
-.gr-button:hover{ transform: translateY(-1px); }
-.gr-button:focus{
-  outline: 3px solid color-mix(in oklab, var(--insee-primary) 30%, white);
-  box-shadow: 0 0 0 4px rgba(11,61,145,.18);
-}
-
-/* primaire */
-button.primary{
-  background: linear-gradient(180deg, var(--insee-primary), var(--insee-primary-700));
-  color: white; border: none;
-}
-/* bouton secondaire "fantôme" */
-button.ghost{
-  background: white; color: var(--insee-primary); border: 2px solid var(--insee-primary);
-}
-button.ghost:hover{
-  background: color-mix(in oklab, var(--insee-primary) 6%, white);
-}
-
-.gr-code pre, .gr-code code{
-  font-size: 13.5px; line-height: 1.45; border-radius: 14px;
-}
-"""
-
-
-# ---------- UI ----------
-with gr.Blocks(title="OpenFoodFactsAgent (COICOP)", theme=theme, css=custom_css) as demo:
-    gr.HTML("""
-    <div id="app-header">
-      <h1>🧮 Agent COICOP — Classification IPC</h1>
-      <p>Analyse de libellé ou d’EAN, interrogation Open Food Facts, règles & sémantique, fusion des candidats.</p>
-    </div>
-    """)
-
-    with gr.Row():
-        with gr.Column(scale=1):
-            with gr.Group(elem_classes=["card"]):
-                gr.Markdown("### Saisie")
-                label_in = gr.Textbox(label="Libellé de produit",
-                                      placeholder="Ex. Camembert au lait cru AOP 250g - ALDI",
-                                      lines=2)
-                ean_in   = gr.Textbox(label="EAN (optionnel)",
-                                      placeholder="Ex. 2006050033638")
-                with gr.Row():
-                    btn_go = gr.Button("Classifier en COICOP", elem_classes=["primary"])
-                    btn_ex = gr.Button("Exemple", elem_classes=["ghost"])
-
-        with gr.Column(scale=1):
-            with gr.Group(elem_classes=["card"]):
-                gr.Markdown("### Résultat JSON")
-                out_json = gr.Code(language="json")
-
-            with gr.Group(elem_classes=["card"]):
-                gr.Markdown("### Logs de l’agent")
-                out_logs = gr.Textbox(lines=16)
-
-    btn_go.click(fn=classify, inputs=[label_in, ean_in], outputs=[out_json, out_logs])
-    btn_ex.click(fn=fill_example, outputs=[label_in, ean_in])
-
-if __name__ == "__main__":
-    demo.launch()
+# app.py — Formulaire EAN/Libellé + logs (UI & Space) — style "IPC / Insee"
+import io, sys, json
+import gradio as gr
+from contextlib import redirect_stdout, redirect_stderr
+
+# 👉 adapte le module si ton fichier agent a un autre nom
+from quick_deploy_agent import build_agent, parse_result
+
+
+# --- en haut de app.py, ajoute cet util ---
+import re
+
+def _extract_json(text: str):
+    """
+    Essaie de parser un JSON valable depuis une chaîne.
+    - d'abord json.loads direct
+    - sinon, extrait le contenu d'un bloc ```json ... ``` puis parse
+    - renvoie None si rien de valable
+    """
+    if isinstance(text, dict):
+        return text
+    if not isinstance(text, str):
+        return None
+    # 1) direct
+    try:
+        return json.loads(text)
+    except Exception:
+        pass
+    # 2) dans des backticks
+    try:
+        m = re.search(r"```(?:json)?\s*(\{[\s\S]*\})\s*```", text)
+        if m:
+            return json.loads(m.group(1))
+    except Exception:
+        pass
+    return None
+
+
+# ---------- util "tee" pour dupliquer les logs ----------
+class _Tee(io.TextIOBase):
+    def __init__(self, *streams): self.streams = streams
+    def write(self, s):
+        for st in self.streams:
+            try: st.write(s)
+            except Exception: pass
+        return len(s)
+    def flush(self):
+        for st in self.streams:
+            try: st.flush()
+            except Exception: pass
+
+# ---------- instancie l'agent + logs verbeux ----------
+agent = build_agent()
+if hasattr(agent, "verbosity_level"):
+    agent.verbosity_level = 3  # 0..3 (3 = très verbeux)
+
+# ---------- prompt interne (l'utilisateur ne tape rien) ----------
+TASK_TMPL = """\
+Classe ce produit en COICOP:
+EAN: {ean}
+Libellé: {label}
+
+Outils autorisés :
+- validate_ean
+- openfoodfacts_product_by_ean
+- map_off_to_coicop
+- coicop_regex_rules
+- coicop_semantic_similarity
+- merge_candidates
+- resolve_coicop_candidates
+- web_search
+- web_get
+- python_interpreter  # autorisé si besoin pour manipuler des listes/dicts
+
+Règles strictes :
+- Utilise python_interpreter uniquement pour manipuler des résultats (listes/dicts, filtrage, fusion).
+- N’écris pas de code inutile (pas de fichiers, pas de réseaux hors outils).
+- Pour (4) et (5), utilise le libellé utilisateur tel quel.
+- Retourne uniquement un JSON valide (objet), sans backticks.
+
+Pipeline :
+1) v = validate_ean(ean)
+2) off = openfoodfacts_product_by_ean(ean)         # si EAN vide ou invalide, passe à l’étape 4
+3) offmap = map_off_to_coicop(off_payload=off)      # si off.ok != True, ignorer
+4) rx = coicop_regex_rules(text=LIBELLÉ UTILISATEUR)
+5) sem = coicop_semantic_similarity(text=LIBELLÉ UTILISATEUR, topk=5)
+6) merged = merge_candidates(candidates_lists=[offmap, rx, sem], min_k=3, fallback_bias="cheese")
+7) res = resolve_coicop_candidates(json_lists=[merged], topn=3)
+
+→ Retourne res tel quel (objet contenant final, alternatives, candidates_top).
+"""
+
+
+
+def classify(label: str, ean: str):
+    import re
+    label = (label or "").strip()
+    ean = (ean or "").strip()
+
+    if not label and not ean:
+        return json.dumps({"error": "Veuillez saisir un libellé ou un EAN."}, ensure_ascii=False, indent=2), "—"
+
+    # Construire le message pour l’agent
+    task = TASK_TMPL.format(ean=ean or "N/A", label=label or "N/A")
+
+    # Buffers UI + duplication vers logs Space (stdout/stderr)
+    buf_out, buf_err = io.StringIO(), io.StringIO()
+    tee_out, tee_err = _Tee(sys.stdout, buf_out), _Tee(sys.stderr, buf_err)
+
+    # Utilitaire local : extraire un vrai JSON d'une chaîne (gère les ```json ... ```)
+    def _extract_json(text: str):
+        if isinstance(text, dict):
+            return text
+        if not isinstance(text, str):
+            return None
+        # 1) direct
+        try:
+            return json.loads(text)
+        except Exception:
+            pass
+        # 2) JSON dans des backticks
+        try:
+            m = re.search(r"```(?:json)?\s*(\{[\s\S]*\})\s*```", text)
+            if m:
+                return json.loads(m.group(1))
+        except Exception:
+            pass
+        return None
+
+    print("\n===== Agent run start =====")
+    print(task)
+
+    try:
+        # Exécuter l’agent en dupliquant les logs vers l’UI et le Space
+        with redirect_stdout(tee_out), redirect_stderr(tee_err):
+            res = agent.run(task)
+
+        # Parsing robuste du résultat
+        obj = None
+        if isinstance(res, dict):
+            obj = res
+        if obj is None:
+            obj = _extract_json(res)
+        if obj is None:
+            try:
+                obj = parse_result(res)  # ton parseur existant
+            except Exception:
+                obj = None
+        if obj is None:
+            obj = {"raw": str(res)}  # dernier recours
+
+        logs_ui = "\n".join(
+            s for s in [buf_out.getvalue().strip(), buf_err.getvalue().strip()] if s
+        ) or "(aucun log)"
+
+        print("===== Agent run end =====\n")
+        return json.dumps(obj, ensure_ascii=False, indent=2), logs_ui
+
+    except Exception as e:
+        logs_ui = "\n".join(
+            s for s in [buf_out.getvalue().strip(), buf_err.getvalue().strip()] if s
+        ) or "(aucun log)"
+        print(f"✖ Agent error: {e}")
+        return json.dumps({"error": str(e)}, ensure_ascii=False, indent=2), logs_ui
+
+
+def fill_example():
+    # EAN réel OFF (Les p'tits crémeux – Aldi – 216 g)
+    return "Camembert au lait cru AOP 250g - ALDI", "2006050033638"
+
+# ---------- thème & CSS (style IPC / Insee) ----------
+theme = gr.themes.Soft(
+    primary_hue="blue",
+    secondary_hue="indigo",
+    neutral_hue="slate",
+)
+
+custom_css = """
+:root{
+  --insee-primary: #89c2d9;
+  --insee-primary-700:#89c2d9;
+  --insee-accent: #ff5c35;
+  --insee-neutral-50:#f8fafc;
+  --insee-neutral-100:#f1f5f9;
+  --insee-neutral-900:#0f172a;
+  --radius-lg: 16px;
+}
+.gradio-container{
+  font-family: Inter, system-ui, -apple-system, Segoe UI, Roboto, Ubuntu, Cantarell, "Helvetica Neue", Arial, "Noto Sans", sans-serif;
+  background: linear-gradient(180deg, var(--insee-neutral-50), #ffffff 30%);
+}
+#app-header{
+  background: linear-gradient(90deg, var(--insee-primary), var(--insee-primary-700));
+  color: white; padding: 18px 22px; border-radius: 18px;
+  box-shadow: 0 10px 30px rgba(11,61,145,0.25); margin-bottom: 16px;
+}
+#app-header h1{ margin:0; font-size: 22px; letter-spacing: .3px; }
+#app-header p{ margin:6px 0 0; opacity:.92 }
+
+.card{
+  background: #fff; border: 1px solid var(--insee-neutral-100); border-radius: var(--radius-lg);
+  box-shadow: 0 8px 24px rgba(15, 23, 42, 0.06);
+  padding: 18px;
+}
+
+.gr-textbox textarea{
+  font-size: 16px; line-height: 1.35; padding: 14px 16px;
+  border: 2px solid #e5e7eb; border-radius: 14px;
+}
+.gr-textbox textarea:focus{
+  outline: 3px solid color-mix(in oklab, var(--insee-primary) 22%, white);
+  border-color: var(--insee-primary);
+  box-shadow: 0 0 0 4px rgba(11,61,145,0.15);
+}
+
+.gr-button{
+  font-weight: 700; letter-spacing: .2px; border-radius: 14px;
+  transition: transform .06s ease, box-shadow .06s ease;
+}
+.gr-button:hover{ transform: translateY(-1px); }
+.gr-button:focus{
+  outline: 3px solid color-mix(in oklab, var(--insee-primary) 30%, white);
+  box-shadow: 0 0 0 4px rgba(11,61,145,.18);
+}
+
+/* primaire */
+button.primary{
+  background: linear-gradient(180deg, var(--insee-primary), var(--insee-primary-700));
+  color: white; border: none;
+}
+/* bouton secondaire "fantôme" */
+button.ghost{
+  background: white; color: var(--insee-primary); border: 2px solid var(--insee-primary);
+}
+button.ghost:hover{
+  background: color-mix(in oklab, var(--insee-primary) 6%, white);
+}
+
+.gr-code pre, .gr-code code{
+  font-size: 13.5px; line-height: 1.45; border-radius: 14px;
+}
+"""
+
+
+# ---------- UI ----------
+with gr.Blocks(title="OpenFoodFactsAgent (COICOP)", theme=theme, css=custom_css) as demo:
+    gr.HTML("""
+    <div id="app-header">
+      <h1>🧮 Agent COICOP — Classification IPC</h1>
+      <p>Analyse de libellé ou d’EAN, interrogation Open Food Facts, règles & sémantique, fusion des candidats.</p>
+    </div>
+    """)
+
+    with gr.Row():
+        with gr.Column(scale=1):
+            with gr.Group(elem_classes=["card"]):
+                gr.Markdown("### Saisie")
+                label_in = gr.Textbox(label="Libellé de produit",
+                                      placeholder="Ex. Camembert au lait cru AOP 250g - ALDI",
+                                      lines=2)
+                ean_in   = gr.Textbox(label="EAN (optionnel)",
+                                      placeholder="Ex. 2006050033638")
+                with gr.Row():
+                    btn_go = gr.Button("Classifier en COICOP", elem_classes=["primary"])
+                    btn_ex = gr.Button("Exemple", elem_classes=["ghost"])
+
+        with gr.Column(scale=1):
+            with gr.Group(elem_classes=["card"]):
+                gr.Markdown("### Résultat JSON")
+                out_json = gr.Code(language="json")
+
+            with gr.Group(elem_classes=["card"]):
+                gr.Markdown("### Logs de l’agent")
+                out_logs = gr.Textbox(lines=16)
+
+    btn_go.click(fn=classify, inputs=[label_in, ean_in], outputs=[out_json, out_logs])
+    btn_ex.click(fn=fill_example, outputs=[label_in, ean_in])
+
+if __name__ == "__main__":
+    demo.launch()