fix(kb-reingest): two regressions found by 2026-04-28 incident on source 141

(1) discard_job was deleting the S3 file even on re-ingest jobs. For a
fresh upload the s3_key points at inbox/<topic>/<kind>/<uuid>-<file>
which the user just dropped — deletable. For a re-ingest, s3_key points
at the EXISTING source's file (kb_source.original_path); deleting it
orphans the source and bricks every future re-process attempt with
"original file unavailable". User hit this on source 141: started a
re-process, hit discard, then the next click on re-process returned
HTTP 400 because the underlying PDF was gone. Now skip the S3 delete
when source_id is set.

(2) start_reingest was setting original_filename = src.title or
filename_from_key(key). When the source's title had no extension
(literally "החלטה" in this case), parse_first_pages couldn't dispatch
by suffix and returned empty text — the classifier then logged
"empty text — skipping" and returned {} for ai_suggestions. Now use
_filename_from_key(key) directly so the parser always sees a real
.pdf/.docx/.txt extension.

Source 141's PDF is permanently lost (MinIO bucket is unversioned).
Recovery options for the user: (a) upload the same PDF to MinIO at
the original key path manually, (b) delete source 141 and re-upload
as a fresh source.

api/services/kb/admin_sources.py

@@ -307,6 +307,12 @@ async def start_reingest(
         "reingest": True,
     }
     batch_id = _uuid.uuid4().hex
+    # Use the actual filename (with extension) — NOT the source title.
+    # parse_first_pages dispatches by extension (.pdf/.docx/.txt); a
+    # title like "החלטה" without an extension makes the parser bail
+    # and return empty text, which then makes the classifier return {}.
+    # See incident 2026-04-28: source 141 (title="החלטה") had every
+    # re-ingest classify as empty until this was fixed.
     pool = await get_pool()
     async with pool.acquire() as conn:
         row = await conn.fetchrow(
@@ -318,7 +324,7 @@ async def start_reingest(
             RETURNING id
             """,
             source_id,
-            src.get("title") or _filename_from_key(key),
+            _filename_from_key(key),
             key,
             src["kind"],
             src["topic_id"],

api/services/kb/ingest_jobs.py

@@ -537,11 +537,22 @@ async def discard_job(job_id: int, requested_by: str | None) -> dict:
     if job["status"] in ("done",):
         raise ValueError(f"job {job_id} already done — cannot discard")
 
-    # Best-effort S3 cleanup; the row stays as audit trail.
+    # Best-effort S3 cleanup — but ONLY for fresh uploads (source_id IS
-    try:
+    # NULL). For re-ingest jobs (source_id set), the s3_key points to
-        kb_s3._client().delete_object(Bucket=kb_s3._bucket(), Key=job["s3_key"])
+    # the existing source's file in kb_source.original_path; deleting
-    except Exception as e:
+    # it would orphan the source and break any future re-process.
-        logger.warning("[discard] S3 delete failed for %s: %s", job["s3_key"], e)
+    # See incident 2026-04-28: source 141 lost its PDF when a discarded
+    # re-ingest deleted the underlying file.
+    if job.get("source_id") is None:
+        try:
+            kb_s3._client().delete_object(Bucket=kb_s3._bucket(), Key=job["s3_key"])
+        except Exception as e:
+            logger.warning("[discard] S3 delete failed for %s: %s", job["s3_key"], e)
+    else:
+        logger.info(
+            "[discard] keeping S3 file for re-ingest job %s (source_id=%s)",
+            job_id, job["source_id"],
+        )
 
     reason = f"discarded by {requested_by or 'user'}"
     async with pool.acquire() as conn: