Add timeouts to model-download HTTP requests - #53
Merged
dralgroup merged 1 commit intoAug 3, 2026
Conversation
Two requests.get() calls that download model files had no timeout, so a stalled or unresponsive server hangs the process indefinitely: - model_cls.py: downloading a model from a link - torchani_interface.py: fetching the ANI-1xnr parameters zip Add timeout=30, consistent with the rest of the codebase, which already sets timeouts on its network calls (MLatom.py, _update_check.py, _version.py). requests' timeout is a per-read inactivity limit, not a total cap, so it does not abort slow-but-progressing streamed downloads.
Owner
|
thank you, released in 3.24.0 with your contribution credited in the commit history and release notes; we built on it so that a download which does not succeed is also recognized instead of being kept as if it were the model. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Two
requests.get()calls that download model files run without atimeout, so a stalled or unresponsive server hangs the MLatom process indefinitely (no way to recover except killing it):mlatom/model_cls.py:287— downloading a model from a link (streamed).mlatom/interfaces/torchani_interface.py:1937— fetching the ANI-1xnr parameters zip from GitHub.Fix
Add
timeout=30to both, consistent with the rest of the codebase, which already sets timeouts on its network calls (MLatom.pyusestimeout=3,_update_check.pyand_version.pyuse timeouts too).Note: requests'
timeoutis a per-read inactivity limit, not a total-duration cap, so this does not abort slow-but-progressing streamed downloads — it only unblocks the process when the connection goes dead.