Run scheduler task immediately

README.md

@@ -8,6 +8,7 @@ MR Reviewer automatise la revue de code sur GitLab en s'appuyant sur un moteur d
 ## Fonctionnalités principales
 
 - Interrogation périodique de GitLab pour récupérer les MRs ouvertes
+- Possibilité de ne traiter qu'un projet spécifique via `PROJECT_ID`
 - Analyse du diff via Mistral et publication automatisée des commentaires
 - Gestion du nombre maximal de tokens et filtrage des MRs trop volumineuses
 - Système de cache pour éviter les doublons
@@ -33,6 +34,7 @@ GITLAB_URL=https://gitlab.com/api/v4
 MISTRAL_API_KEY=sk-xxxx
 CACHE_FILE=cache.json
 POLLING_INTERVAL=300
+PROJECT_ID=123  # Laisser vide pour traiter toutes vos MR
 MAX_TOKENS=8000
 LOG_LEVEL=INFO
 # Optionnel : fichier où enregistrer les logs
@@ -47,7 +49,7 @@ python -m app.main
 
 ## Comportement général
 
-- Le scheduler déclenche `review_merge_requests()` à l'intervalle défini
+- Le scheduler déclenche immédiatement `review_merge_requests()` puis à l'intervalle défini
 - Pour chaque MR détectée :
   - si elle a déjà été traitée, elle est ignorée
   - si le diff dépasse la limite autorisée, un message explicatif est envoyé

app/adapters/gitlab_adapter.py

@@ -42,14 +42,15 @@ def _request(self, method: str, url: str, **kwargs) -> tuple[any, dict]:
             raise
         return resp.json(), resp.headers
 
-    def get_open_merge_requests(self) -> Iterable[MergeRequest]:
-        """
-        Retrieve all open merge requests from GitLab.
+    def get_open_merge_requests(self, project_id: int | None = None) -> Iterable[MergeRequest]:
+        """Retrieve open merge requests.
 
-        Returns:
-            Iterable[MergeRequest]: An iterable of MergeRequest objects representing open merge requests.
+        When ``project_id`` is provided, only MRs of that project are returned.
         """
-        url = f"{self.base_url}/merge_requests"
+        if project_id is not None:
+            url = f"{self.base_url}/projects/{project_id}/merge_requests"
+        else:
+            url = f"{self.base_url}/merge_requests"
         params = {"state": "opened", "per_page": 100, "page": 1}
         results: List[MergeRequest] = []
         while True:

app/main.py

@@ -1,4 +1,5 @@
 import time
+
 from app.adapters.gitlab_adapter import GitLabAdapter
 from app.adapters.mistral_adapter import MistralAdapter
 from app.adapters.cache_adapter import JSONCacheAdapter
@@ -9,25 +10,23 @@
 
 
 def main():
-    """
-    Main entry point for the application.
-
-    Initializes the logging, adapters, and starts the scheduler to review merge requests.
-    This function is responsible for setting up the environment and starting the review process.
-    """
+    """Main entry point for the application."""
     configure_logging()
     gitlab = GitLabAdapter(token=settings.gitlab_token, base_url=settings.gitlab_url)
     ai = MistralAdapter(api_key=settings.mistral_api_key, model=settings.mistral_model)
     cache = JSONCacheAdapter(cache_file=settings.cache_file)
 
     start_scheduler(
-        lambda: run_merge_request_review(gitlab, ai, cache), settings.polling_interval
+        lambda: run_merge_request_review(gitlab, ai, cache, project_id=settings.project_id),
+        settings.polling_interval,
     )
 
     try:
         while True:
             time.sleep(1)
     except KeyboardInterrupt:
+        pass
+    finally:
         stop_scheduler()
 
 

app/ports/output/gitlab_port.py

@@ -10,8 +10,12 @@ class GitLabPort(ABC):
     """Interface used to interact with GitLab."""
 
     @abstractmethod
-    def get_open_merge_requests(self) -> Iterable[MergeRequest]:
-        """Return currently opened merge requests."""
+    def get_open_merge_requests(self, project_id: int | None = None) -> Iterable[MergeRequest]:
+        """Return currently opened merge requests.
+
+        Args:
+            project_id: restrict search to this project when provided.
+        """
 
         raise NotImplementedError
 

app/scheduler/polling.py

@@ -21,6 +21,7 @@ def start_scheduler(task: Callable[[], None], interval: int) -> None:
     global _scheduler
     scheduler = BackgroundScheduler()
     scheduler.add_job(task, "interval", seconds=interval)
+    task()
     scheduler.start()
     _scheduler = scheduler
     logger.info("Scheduler démarré (intervalle=%ss)", interval)

app/usecases/review_merge_requests.py

@@ -10,12 +10,14 @@
 from config.settings import settings
 
 
-def run_merge_request_review(gitlab: GitLabPort, ai: AIPort, cache: CachePort) -> None:
+def run_merge_request_review(
+    gitlab: GitLabPort, ai: AIPort, cache: CachePort, project_id: int | None = None
+) -> None:
     """Fetch opened MRs and post an automated review when possible."""
     logger = logging.getLogger("Review")
     engine = ReviewEngine(ai, settings.max_tokens)
 
-    for mr in gitlab.get_open_merge_requests():
+    for mr in gitlab.get_open_merge_requests(project_id=project_id):
         if cache.is_up_to_date(mr.id, mr.sha):
             logger.debug("MR %s déjà analysée", mr.id)
             continue

config/settings.py

@@ -12,6 +12,7 @@ class Settings(BaseSettings):
 
     cache_file: str = Field(default="cache.json", env="CACHE_FILE")
     polling_interval: int = Field(default=300, env="POLLING_INTERVAL")
+    project_id: int | None = Field(default=None, env="PROJECT_ID")
     max_tokens: int = Field(default=8000, env="MAX_TOKENS")
     log_level: str = Field(default="INFO", env="LOG_LEVEL")
     log_file: str | None = Field(default=None, env="LOG_FILE")

docs/ARCHITECTURE.md

@@ -77,7 +77,7 @@ Définit les interfaces abstraites que les adaptateurs techniques doivent implé
   Interface : `review_diff(diff: str) -> str`
 
 - **gitlab_port.py**  
-  Interface : `get_open_merge_requests()`, `get_diff(mr_id)`, `post_comment(mr_id, text)`
+  Interface : `get_open_merge_requests(project_id=None)`, `get_diff(mr_id)`, `post_comment(mr_id, text)`
 
 - **cache_port.py**  
   Interface : `is_up_to_date(mr_id, sha)`, `update_reviewed(mr_id, sha)`

tests/integration/test_usecase_run_review.py

@@ -5,7 +5,8 @@
 class DummyGitLab:
     def __init__(self):
         self.comments = []
-    def get_open_merge_requests(self):
+
+    def get_open_merge_requests(self, project_id=None):
         return [MergeRequest(id=1, project_id=1, sha="s1")]
     def get_diff(self, project_id, mr_id):
         return "good diff"
@@ -33,7 +34,7 @@ def test_run_merge_request_review(monkeypatch):
     ai = DummyAI()
     cache = DummyCache()
     monkeypatch.setattr(usecase, "settings", SimpleNamespace(max_tokens=100))
-    usecase.run_merge_request_review(gitlab, ai, cache)
+    usecase.run_merge_request_review(gitlab, ai, cache, project_id=None)
     assert ai.called == "good diff"
     assert gitlab.comments == ["reviewed"]
     assert cache.data[1] == "s1"
@@ -44,5 +45,5 @@ def test_skip_when_diff_too_large(monkeypatch):
     ai = DummyAI()
     cache = DummyCache()
     monkeypatch.setattr(usecase, "settings", SimpleNamespace(max_tokens=1))
-    usecase.run_merge_request_review(gitlab, ai, cache)
+    usecase.run_merge_request_review(gitlab, ai, cache, project_id=None)
     assert gitlab.comments == ["Diff trop volumineux pour analyse automatique."]

tests/unit/test_gitlab_adapter.py

@@ -37,6 +37,22 @@ def fake_request(method, url, **kwargs):
     assert calls[-1][2]["params"]["page"] == "2"
 
 
+def test_get_open_merge_requests_for_project(monkeypatch):
+    resp = DummyResponse([], headers={})
+    captured = {}
+
+    def fake_request(method, url, **kwargs):
+        captured["url"] = url
+        captured["params"] = kwargs["params"]
+        return resp
+
+    monkeypatch.setattr(requests, "request", fake_request)
+    adapter = GitLabAdapter(token="t", base_url="http://x")
+    list(adapter.get_open_merge_requests(project_id=5))
+    assert captured["url"].endswith("/projects/5/merge_requests")
+    assert captured["params"]["state"] == "opened"
+
+
 def test_get_diff_and_post_comment(monkeypatch):
     def fake_request(method, url, **kwargs):
         if method == "GET":

tests/unit/test_main.py

@@ -24,6 +24,7 @@ def test_main_calls_dependencies(
     mock_settings.mistral_model = "model"
     mock_settings.cache_file = "cache.json"
     mock_settings.polling_interval = 42
+    mock_settings.project_id = None
 
     with (
         patch("app.main.time.sleep", side_effect=KeyboardInterrupt),
@@ -66,10 +67,13 @@ def test_main_runs_lambda(mock_start_scheduler):
         mock_settings.mistral_model = "model"
         mock_settings.cache_file = "cache.json"
         mock_settings.polling_interval = 42
+        mock_settings.project_id = 123
 
         app.main.main()
 
         # Extract the lambda and call it
         lambda_func = mock_start_scheduler.call_args[0][0]
         lambda_func()
-        mock_review.assert_called_once_with(fake_gitlab, fake_ai, fake_cache)
+        mock_review.assert_called_once_with(
+            fake_gitlab, fake_ai, fake_cache, project_id=123
+        )

tests/unit/test_polling_scheduler.py

@@ -30,6 +30,7 @@ def task():
         called.append(True)
 
     polling.start_scheduler(task, interval=5)
+    assert called == [True]
     assert dummy.started
     assert dummy.jobs[0][1] == "interval" and dummy.jobs[0][2] == 5
     assert polling._scheduler is dummy