Caching and large file download

[dgunning]

@paultiq you mentioned that caching might be causing a performance issue when downloading bulk files. Do you know why when download_filings is called there is a 2 minute wait before the download progress bar displays?

Downloading feed file(s): 0it [00:00, ?it/s]async call made without an async bucket.

[paultiq]

There's two different things here, but updating the cache rules for #378 to exclude these files is the best first step, I'll take a look...

It's somewhat tricky to get streaming progress and caching to play nice together, but I have that on another branch right now.

[paultiq]

I updated the cache rules & added an option to bypass the cache. This option is now used when streaming bulk files, which ensures they are rate limited but not cached.

[dgunning]

Can we also turn down the logging - to DEBUG? This happens in some settings not all, but I think users might be bothered seeing this on every request

2025-08-14 10:57:09 | httpxthrottlecache.httpxclientmanager | INFO | Creating new HTTPX Client
2025-08-14 10:57:09 | httpxthrottlecache.controller | INFO | matched .*\.sec\.gov, using value data.sec.gov: {'/submissions.*': 600, '/include/ticker\\.txt.*': 600, '/files/company_tickers\\.json.*

[paultiq]

Yes, I think we'd just change the logging inside edgartools to set httpxthrottlecache to warning, rather than changing it in the lib itself.

What do you think?

[dgunning]

That's fine