Add Stage 2.8 recall, quality gate, retries, and publish idempotency

ai_daily_report/collect.py

@@ -5,6 +5,7 @@ from datetime import datetime, timezone
 from time import perf_counter
 from typing import Callable, Iterable, Any
 
+from .clients import FetchTextError
 from .models import SourceConfig, SourceResult
 
 
@@ -12,11 +13,19 @@ Fetcher = Callable[[SourceConfig, str], list[dict[str, Any]]]
 
 
 def _status_from_exception(exc: Exception) -> str:
+    if isinstance(exc, FetchTextError):
+        return exc.error_type
     if isinstance(exc, TimeoutError):
         return "timeout"
     return "error"
 
 
+def _retry_count_from_exception(exc: Exception) -> int:
+    if isinstance(exc, FetchTextError):
+        return max(0, exc.attempts - 1)
+    return 0
+
+
 def _collect_one(config: SourceConfig, run_date: str, fetcher: Fetcher) -> SourceResult:
     fetched_at = datetime.now(timezone.utc).isoformat()
     if not config.enabled:
@@ -51,6 +60,7 @@ def _collect_one(config: SourceConfig, run_date: str, fetcher: Fetcher) -> Sourc
             status=_status_from_exception(exc),
             error=f"{type(exc).__name__}: {exc}",
             elapsed_ms=elapsed_ms,
+            retry_count=_retry_count_from_exception(exc),
             fetched_at=fetched_at,
         )
 
@@ -91,5 +101,10 @@ def collect_sources(
         "raw_item_count": sum(len(result.items) for result in results),
         "source_counts": {result.source: len(result.items) for result in results},
         "statuses": {result.source: result.status for result in results},
+        "error_types": {
+            result.source: result.status
+            for result in results
+            if not result.ok and result.status != "disabled"
+        },
     }
     return results, report