aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
-rw-r--r--python/sandcrawler/ingest.py5
1 files changed, 5 insertions, 0 deletions
diff --git a/python/sandcrawler/ingest.py b/python/sandcrawler/ingest.py
index 0be7653..82b43fe 100644
--- a/python/sandcrawler/ingest.py
+++ b/python/sandcrawler/ingest.py
@@ -238,6 +238,11 @@ class IngestFileWorker(SandcrawlerWorker):
error_message="ingest worker internal timeout",
)
+ def want(self, request):
+ if not request.get('ingest_type') in ('file', 'pdf'):
+ return False
+ return True
+
def process(self, request):
# backwards compatibility