From e99d9f2fddcb8b52ba52128b290ec5e0f367392f Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Tue, 3 Nov 2020 17:17:02 -0800
Subject: gen_file_metadata: detect JATS XML and use application/jats+xml

---
 python/sandcrawler/misc.py | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/python/sandcrawler/misc.py b/python/sandcrawler/misc.py
index 8c91246..67e5c0b 100644
--- a/python/sandcrawler/misc.py
+++ b/python/sandcrawler/misc.py
@@ -26,6 +26,10 @@ def gen_file_metadata(blob: bytes) -> dict:
     """
     assert blob
     mimetype = magic.Magic(mime=True).from_buffer(blob)
+    if mimetype in ("application/xml", "text/xml"):
+        # crude check for JATS XML, using only first 1 kB of file
+        if b"<article " in blob[:1024] and not b"<html" in blob[:1024]:
+            mimetype = "application/jats+xml"
     hashes = [
         hashlib.sha1(),
         hashlib.sha256(),
-- 
cgit v1.2.3