From abaa0c53c8f2aaff3c533747c2f310d8f60839c9 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Thu, 30 Apr 2020 16:22:59 -0700
Subject: ingest: don't 'want' non-PDF ingest

---
 python/sandcrawler/ingest.py | 5 +++++
 1 file changed, 5 insertions(+)

(limited to 'python')

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
-- 
cgit v1.2.3