From 3b0bb20ad0b9ce58992a45e1fcf863069119f560 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Fri, 3 Apr 2020 18:33:17 -0700
Subject: include ia_pdf_url when available

---
 fatcat_covid19/transform.py | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fatcat_covid19/transform.py b/fatcat_covid19/transform.py
index 16774ab..3f942ba 100644
--- a/fatcat_covid19/transform.py
+++ b/fatcat_covid19/transform.py
@@ -152,6 +152,10 @@ def fulltext_to_elasticsearch(row, force_bool=True):
             t['fulltext']['thumbnail_url'] = "/" + full['thumbnail_path']
         if full.get('grobid_xml_path'):
             t['fulltext']['grobid_xml_url'] = "/" + full['grobid_xml_path']
+        for url in full.get('urls', []):
+            if url.get('rel') in ('webarchive', 'archive') and 'archive.org/' in url['url']:
+                t['fulltext']['ia_pdf_url'] = url['url']
+                break
 
     if 'fulltext_grobid' in row:
         grobid = row['fulltext_grobid']
-- 
cgit v1.2.3