Minimal Archive for Testing / Different Archive Formats
Opened this issue · 5 comments
Bringing part of the discussion from #7 over here:
Some users (@krishna015) seem to be getting a different archive which can't be processed using the script atm. I've tried using different languages but haven't been able to reproduce how they're getting that archive.
#9 also mentions a need for a minimal archive to help with testing.
I think we were all under the assumption that the archive looks like #7 (comment)
If anyone has examples of how to get an archive that looks different, that'd be helpful!
@MetricMike I think if we can try to create some pull request for that, to detect which sort of archive is that? I don't have access to such archive so someone else have to do it. Also probably it has to use different methods for parsing as that's one big file for messages
@MetricMike thanks for help on other PR. So what about test suite?
Thank you @marzann for helping out with this:
For now I'm going to focus my efforts with my facebook archive:
- Does the archive look like we expect it to?
- Are we actually able to analyze messages/rank friends/etc
- How long does each step take?
We'll probably have some overlap and significant diffs so I'm excited to see where we end up landing and working with you to make future development easier 😄
In order to analyse the different archives, an option might be, to make a script that will replace the text with a random string of the same length. That way we can ask user to send their archives in for further analyzation without compromising their privacy. That will also help in parsing (maybe different structures, who knows atm?)
What ya thinking?