How to get actual snapshot files? #25
Description
Playwright keeps changing their (undocumented) JSONL output, so I figured I would give CTRF a try. I have it integrated into my pipeline, but the outputs are not what I expect. For example, when playwright completes a run and I upload the output folder as an artifact in a GitHub Actions pipeline, I get a .jsonl file that points to files in a resources folder. Those files are the expected and actual screenshots. However, this package seems to output only a .json file that points to files on the ephemeral Actions file system -- the files are no where to be found in the output.
How can I gather all of these files listed in the test.attachments key? I'm of course running multiple shards, so I need to gather up all of the output, combine it, and then feed ti to the next process.