Conversation
9854eac to
e8c362b
Compare
|
|
use iter_content to avoid storing transparently compressed data.
ie. we are receiving transparently compressed data and storing the compressed stream instead. i guess we re-introduced this in #1840 (and i was advocating for it 🤦 ) |
|
Ready to merge from my side. |
fneum
left a comment
There was a problem hiding this comment.
LGTM. If I understand the snakemake docs correctly (https://snakemake.readthedocs.io/en/stable/snakefiles/storage.html), the cached-http plugin is not yet used in this PR as it would need:
storage:
provider="cached-http",
if it is to be used by default or calls to storage.cached_http(). Is that right?
No, there are two facilities how stuff is being used.
ie. the plugin is active for zenodo urls. |
|
Gotcha. Thanks for explaining! |
…ures (#1913) * feat: Add snakemake-storage-plugin-cached-http to alleviate zenodo failures * Add release note * fix: don't use progress bar for cost retrieval * Desperate debugging * More desperation * fix: fix progress_retrieve use iter_content to avoid storing transparently compressed data. * chore: clean up desperation * Apply suggestion from @lkstrp --------- Co-authored-by: Lukas Trippe <[email protected]>
Changes proposed in this Pull Request
Add snakemake-storage-plugin-cached-http as planned in #1675 to fix recurring Zenodo retrieval failures.
Fix a re-introduced stream decompression error in
progress_retrieve(follow-up to #1840).Checklist
pixi.toml(usingpixi add <dependency-name>).doc/release_notes.rstis added.