Tim Allison created TIKA-2349:
---------------------------------
Summary: Try to match digests when finding equivalent embedded
files in tika-eval Compare
Key: TIKA-2349
URL: https://issues.apache.org/jira/browse/TIKA-2349
Project: Tika
Issue Type: Improvement
Components: tika-eval
Reporter: Tim Allison
Priority: Trivial
In tika-eval {{Compare}}, there's some code that tries to find equivalent
embedded files. The challenge is that embedded resource paths may contain
different names in different versions of Tika, or they may be at different
offsets within the {{List<Metadata>}}. In the most recent regression run, we
had several cases where the {{List<Metadata>}}s were the same size, but the
attachments were ordered differently.
Instead of relying solely on position, let's try to match digest keys first,
then back off to path names, and finally, back off equal locations in the list,
if the lists have the same size.
More areas for improvement will exist for this bit of code...
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
