From 97aa00038fbce39097d5f78b8891bbb88b71af75 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Mon, 24 Oct 2022 10:19:17 -0700
Subject: html: pubpub platform detection

---
 python/sandcrawler/ingest_html.py | 2 ++
 1 file changed, 2 insertions(+)

(limited to 'python')

diff --git a/python/sandcrawler/ingest_html.py b/python/sandcrawler/ingest_html.py
index 1c2c3fd..fb42e71 100644
--- a/python/sandcrawler/ingest_html.py
+++ b/python/sandcrawler/ingest_html.py
@@ -260,6 +260,8 @@ def html_guess_platform(
                 in doc.html
             ):
                 return "ojs"
+            if '<a href="https://www.pubpub.org">Published with' in doc.html:
+                return "pubpub"
             if 'Powered by <a target="_blank" href="http://arphahub.com">' in doc.html:
                 return "arpha"
             if "<meta property='og:image' content='http://cms.galenos.com.tr' />" in doc.html:
-- 
cgit v1.2.3