Deploy OpenClaw PR API

src/slop_farmer/app/pr_search_api.py

@@ -10,7 +10,7 @@ from fastapi import FastAPI, HTTPException, Request
 from fastapi.responses import JSONResponse
 
 from slop_farmer.config import PrSearchRefreshOptions
-from slop_farmer.data.ghreplica_api import GhrProbeClient
+from slop_farmer.data.ghreplica_api import GhReplicaProbeUnavailableError, GhrProbeClient
 from slop_farmer.reports.pr_search_service import (
     get_pr_search_candidate_clusters,
     get_pr_search_cluster,
@@ -98,6 +98,12 @@ def create_app(settings: PrSearchApiSettings | None = None) -> FastAPI:
         status_code = 404 if _looks_not_found(exc) else 400
         return JSONResponse({"detail": str(exc)}, status_code=status_code)
 
+    @app.exception_handler(GhReplicaProbeUnavailableError)
+    async def handle_probe_unavailable(
+        _request: Request, exc: GhReplicaProbeUnavailableError
+    ) -> JSONResponse:
+        return JSONResponse({"detail": str(exc)}, status_code=exc.status_code)
+
     @app.get("/healthz")
     async def healthz() -> dict[str, bool]:
         return {"ok": True}

src/slop_farmer/data/ghreplica_api.py

@@ -19,6 +19,14 @@ class GhReplicaApiRequestError(RuntimeError):
         super().__init__(f"ghreplica API request failed: {status_code} {path} {detail}")
 
 
+class GhReplicaProbeUnavailableError(RuntimeError):
+    """Raised when ghreplica cannot yet serve a live probe payload."""
+
+    def __init__(self, detail: str, *, status_code: int = 503):
+        self.status_code = status_code
+        super().__init__(detail)
+
+
 class GhrProbeClient:
     provider = "ghreplica"
 
@@ -52,14 +60,54 @@ class GhrProbeClient:
             raise GhReplicaApiRequestError(exc.code, path, detail) from exc
         return json.loads(payload)
 
+    def _request_json_or_none(self, path: str) -> Any | None:
+        try:
+            return self._request_json(path)
+        except GhReplicaApiRequestError as exc:
+            if exc.status_code == 404:
+                return None
+            raise
+
     def get_pull_request(self, owner: str, repo: str, number: int) -> dict[str, Any]:
-        payload = self._request_json(f"/v1/github/repos/{owner}/{repo}/pulls/{number}")
+        try:
+            payload = self._request_json(f"/v1/github/repos/{owner}/{repo}/pulls/{number}")
+        except GhReplicaApiRequestError as exc:
+            if exc.status_code == 404:
+                raise GhReplicaProbeUnavailableError(
+                    f"PR #{number} was not found in ghreplica.",
+                    status_code=404,
+                ) from exc
+            raise
         if not isinstance(payload, dict):
             raise RuntimeError(f"Expected dict payload for pull request, got {type(payload)!r}")
         return payload
 
     def iter_pull_files(self, owner: str, repo: str, number: int) -> Iterable[dict[str, Any]]:
-        payload = self._request_json(f"/v1/changes/repos/{owner}/{repo}/pulls/{number}/files")
+        try:
+            payload = self._request_json(f"/v1/changes/repos/{owner}/{repo}/pulls/{number}/files")
+        except GhReplicaApiRequestError as exc:
+            if exc.status_code != 404:
+                raise
+            status = self.get_pull_request_status(owner, repo, number)
+            if isinstance(status, dict):
+                detail_bits = []
+                for key in (
+                    "indexed",
+                    "backfill_in_progress",
+                    "changed_files",
+                    "indexed_file_count",
+                ):
+                    if key in status:
+                        detail_bits.append(f"{key}={status[key]}")
+                suffix = f" ({', '.join(detail_bits)})" if detail_bits else ""
+                raise GhReplicaProbeUnavailableError(
+                    f"PR #{number} is not available in ghreplica yet{suffix}.",
+                    status_code=503,
+                ) from exc
+            raise GhReplicaProbeUnavailableError(
+                f"PR #{number} was not found in ghreplica changed-file replica.",
+                status_code=404,
+            ) from exc
         rows = payload if isinstance(payload, list) else payload.get("files")
         if not isinstance(rows, list):
             raise RuntimeError(
@@ -85,7 +133,9 @@ class GhrProbeClient:
             }
 
     def get_pull_request_status(self, owner: str, repo: str, number: int) -> dict[str, Any] | None:
-        payload = self._request_json(f"/v1/changes/repos/{owner}/{repo}/pulls/{number}/status")
+        payload = self._request_json_or_none(
+            f"/v1/changes/repos/{owner}/{repo}/pulls/{number}/status"
+        )
         if payload is None:
             return None
         if not isinstance(payload, dict):