fix(ai): risposta AI malformata -> 502, non 400 (#36 punto 7)

json.JSONDecodeError e' un ValueError: KeyError/JSONDecodeError sulla
risposta del modello finivano nell'except ValueError del router (pensato
per i nostri errori di configurazione, es. API key mancante) e uscivano
come 400 invece di 502. Nuova AIResponseError dedicata in ai_service.py
per contenuto malformato o JSON non valido; recipes.py la traduce in 502,
toccato solo per questo except.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

src/backend/api/routers/recipes.py

@@ -189,6 +189,11 @@ async def parse_technical_sheet(
         tasks = await ai_service.parse_technical_sheet(pdf_bytes)
     except ValueError as e:
         raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=str(e))
+    except ai_service.AIResponseError as e:
+        # The AI model's own response was malformed (bad JSON, missing
+        # fields): that is the upstream service's fault, not ours (400).
+        logging.getLogger(__name__).error("AI response malformed: %s", e)
+        raise HTTPException(status_code=status.HTTP_502_BAD_GATEWAY, detail="Risposta del servizio AI non valida")
     except Exception as e:
         logging.getLogger(__name__).error("AI parsing error: %s", e)
         raise HTTPException(status_code=status.HTTP_502_BAD_GATEWAY, detail="Errore nel servizio AI")

src/backend/services/ai_service.py

@@ -11,6 +11,17 @@ from src.backend.config import settings
 
 logger = logging.getLogger(__name__)
 
+
+class AIResponseError(Exception):
+    """The AI model's response could not be used: missing fields or invalid JSON.
+
+    Distinct from ValueError (raised for our own misconfiguration, e.g. a
+    missing API key) so the router can tell "we are broken" (400) apart from
+    "the upstream AI service is broken" (502) instead of both landing on the
+    same except clause.
+    """
+
+
 SYSTEM_PROMPT = """Sei un assistente specializzato nell'analisi di schede tecniche industriali.
 Analizza il testo estratto da una scheda tecnica e identifica i blocchi operativi distinti.
 Per ogni blocco, estrai:
@@ -65,7 +76,10 @@ async def parse_technical_sheet(pdf_bytes: bytes) -> list[dict]:
         resp.raise_for_status()
 
     data = resp.json()
-    content = data["choices"][0]["message"]["content"].strip()
+    try:
+        content = data["choices"][0]["message"]["content"].strip()
+    except (KeyError, IndexError, TypeError, AttributeError) as e:
+        raise AIResponseError(f"Risposta del modello AI malformata: {e}") from e
 
     # Strip markdown fences if present
     if content.startswith("```"):
@@ -74,4 +88,10 @@ async def parse_technical_sheet(pdf_bytes: bytes) -> list[dict]:
             content = content[:-3]
         content = content.strip()
 
-    return json.loads(content)
+    try:
+        return json.loads(content)
+    except json.JSONDecodeError as e:
+        # json.JSONDecodeError is itself a ValueError: without this, the
+        # router's `except ValueError` (meant for our own config errors)
+        # would misreport a broken upstream response as our fault (400).
+        raise AIResponseError(f"Risposta del modello AI non e' JSON valido: {e}") from e

src/backend/tests/test_ai_service.py

@@ -0,0 +1,67 @@
+"""Tests for ai_service.py error handling (gitea#36 point 7).
+
+A malformed response from the AI model (missing fields, invalid JSON) must
+raise the dedicated AIResponseError, not a bare KeyError/JSONDecodeError —
+JSONDecodeError in particular is itself a ValueError subclass, so without
+this it was caught by the router's `except ValueError` (meant for our own
+config errors) and misreported as a 400 instead of a 502.
+"""
+import json
+from unittest.mock import AsyncMock, patch
+
+import pytest
+
+from src.backend.services import ai_service
+
+
+class _FakeResponse:
+    def __init__(self, json_data):
+        self._json_data = json_data
+
+    def raise_for_status(self):
+        pass
+
+    def json(self):
+        return self._json_data
+
+
+def _patch_openrouter(monkeypatch, response_json: dict):
+    monkeypatch.setattr(ai_service, "extract_text_from_pdf", lambda _b: "some extracted text")
+    monkeypatch.setattr(ai_service.settings, "openrouter_api_key", "test-key")
+    monkeypatch.setattr(ai_service.settings, "openrouter_model", "test-model")
+
+    fake_client = AsyncMock()
+    fake_client.post = AsyncMock(return_value=_FakeResponse(response_json))
+    fake_client.__aenter__ = AsyncMock(return_value=fake_client)
+    fake_client.__aexit__ = AsyncMock(return_value=False)
+
+    return patch("src.backend.services.ai_service.httpx.AsyncClient", return_value=fake_client)
+
+
+async def test_missing_choices_key_raises_ai_response_error(monkeypatch):
+    with _patch_openrouter(monkeypatch, {"no_choices_here": True}):
+        with pytest.raises(ai_service.AIResponseError):
+            await ai_service.parse_technical_sheet(b"%PDF-fake")
+
+
+async def test_invalid_json_content_raises_ai_response_error(monkeypatch):
+    bad_content = "not valid json {{{"
+    response_json = {"choices": [{"message": {"content": bad_content}}]}
+    with _patch_openrouter(monkeypatch, response_json):
+        with pytest.raises(ai_service.AIResponseError):
+            await ai_service.parse_technical_sheet(b"%PDF-fake")
+
+
+async def test_valid_json_content_returns_tasks(monkeypatch):
+    tasks = [{"title": "MATERIALI", "directive": "Verificare", "description": "..."}]
+    response_json = {"choices": [{"message": {"content": json.dumps(tasks)}}]}
+    with _patch_openrouter(monkeypatch, response_json):
+        result = await ai_service.parse_technical_sheet(b"%PDF-fake")
+    assert result == tasks
+
+
+async def test_empty_extracted_text_returns_empty_list(monkeypatch):
+    """No AI call at all when the PDF has no extractable text."""
+    monkeypatch.setattr(ai_service, "extract_text_from_pdf", lambda _b: "   ")
+    result = await ai_service.parse_technical_sheet(b"%PDF-fake")
+    assert result == []

src/backend/tests/test_recipes_ai_parse_error.py

@@ -0,0 +1,33 @@
+"""Router-level test for gitea#36 point 7: recipes.py translates
+ai_service.AIResponseError to 502, not the config-error 400.
+
+A dedicated file (not test_recipes.py) to avoid colliding with concurrent
+work on the copy-on-write recipe logic in that shared test module.
+"""
+from httpx import AsyncClient
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from src.backend.models.orm.user import User
+from src.backend.services import ai_service
+from src.backend.tests.conftest import auth_headers, create_test_recipe
+
+
+async def test_ai_response_error_is_502(
+    client: AsyncClient,
+    maker_user: User,
+    db_session: AsyncSession,
+    monkeypatch,
+):
+    recipe = await create_test_recipe(db_session, maker_user.id, code="AI-PARSE-502")
+
+    async def _raise_ai_response_error(_pdf_bytes):
+        raise ai_service.AIResponseError("Risposta del modello AI non e' JSON valido: boom")
+
+    monkeypatch.setattr(ai_service, "parse_technical_sheet", _raise_ai_response_error)
+
+    resp = await client.post(
+        f"/api/recipes/{recipe.id}/parse-technical-sheet",
+        headers=auth_headers(maker_user),
+        files={"file": ("sheet.pdf", b"%PDF-1.4 fake", "application/pdf")},
+    )
+    assert resp.status_code == 502