fix: propagate archive source transport errors

Transport and API failures from Archive.org and Wikimedia now reach corpus_builder while per-item Archive.org hydration remains tolerant. Update regression expectations for the corrected contract.

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

Author: Copilot

tests/tools/test_stock_source_adapters.py

@@ -109,8 +109,6 @@ def test_unsplash_helpers_preserve_query_params():
 # defects rather than encoded as expected behavior. `strict=True` means
 # the suite fails the moment one of them is fixed and left in this map.
 _STILL_SWALLOWS_TRANSPORT_ERRORS = {
-    "archive_org": "#511 follow-up: query cascade continues past a failed strategy",
-    "wikimedia": "#511 follow-up: query cascade continues past a failed strategy",
     "pond5_pd": "#511 follow-up: API failure falls through to the web fallback",
 }
 

tools/video/stock_sources/archive_org.py

@@ -157,14 +157,9 @@ def search(self, query: str, filters: SearchFilters) -> list[Candidate]:
                 ("output", "json"),
             ]
 
-            try:
-                r = requests.get(_SEARCH_URL, params=params, timeout=30)
-                r.raise_for_status()
-                data = r.json()
-            except Exception:
-                # One strategy's network/parse error shouldn't kill the
-                # whole cascade — try the next one.
-                continue
+            r = requests.get(_SEARCH_URL, params=params, timeout=30)
+            r.raise_for_status()
+            data = r.json()
             docs = (data.get("response") or {}).get("docs", []) or []
             if not docs:
                 continue
@@ -328,9 +323,8 @@ def _hydrate_candidate(
             meta = r.json()
         except Exception:
             # Swallow per-item fetch failures — one bad item shouldn't
-            # poison the whole search. Alternative would be to raise and
-            # have corpus_builder catch per-source, but at this layer we
-            # can keep going.
+            # poison the whole search. Search-level failures above still
+            # propagate so corpus_builder can report a broken source.
             return None
 
         files = meta.get("files") or []

tools/video/stock_sources/wikimedia.py

@@ -82,17 +82,14 @@ def search(self, query: str, filters: SearchFilters) -> list[Candidate]:
                 "inprop": "url",
             }
 
-            try:
-                r = requests.get(
-                    _API_URL,
-                    params=params,
-                    headers={"User-Agent": _USER_AGENT},
-                    timeout=30,
-                )
-                r.raise_for_status()
-                data = r.json()
-            except Exception:
-                continue
+            r = requests.get(
+                _API_URL,
+                params=params,
+                headers={"User-Agent": _USER_AGENT},
+                timeout=30,
+            )
+            r.raise_for_status()
+            data = r.json()
             pages = list(((data.get("query") or {}).get("pages") or {}).values())
             if not pages:
                 continue