Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fixing data_directory not actually changing download directory and enforced spacy and thinc version requirements #229

Merged
merged 6 commits into from
Nov 8, 2024

Conversation

Torchee
Copy link
Contributor

@Torchee Torchee commented Oct 5, 2024

Description:
The fix/update is two-fold:

  1. Solving the problem with ConvoKit download.
  2. Ensuring ConvoKit would work with numpy 2.0.0+

Solutions to each update:

  1. Updated util.py so that convokit downloads data to the directory specified in config.yml instead of to "downloads" always.
  2. Updated setup.py to enforce spacy>=3.8.2 and thinc>=8.3.0, <8.4.0. Pre Spacy 3.8.2 was not compatible with numpy 2.0.0+ due to thinc issues. Spacy 3.8.2 is compatible with numpy 2.0.0+ but currently requires thinc to be between these versions, so as a temporary solution we enforce these requirements. We will continue to keep an eye on spacy releases and update the requirements if there are new releases targeting this issue.
    For additional insight into the issue:
    Numpy v2.0.0 breaks the ability to download models using spaCy explosion/spaCy#13528
    [BUG] -- numpy 2.0.0 raises ValueError in thinc.backends.numpy_ops explosion/thinc#939

@Torchee Torchee changed the title fixing data_directory not actually changing download directory fixing data_directory not actually changing download directory and enforced spacy and thinc version requirements Nov 7, 2024
Updated troubleshooting documentation with description of package incompatibility issue and the updated requirements to fix it
@seanzhangkx8 seanzhangkx8 merged commit e1a1b19 into CornellNLP:master Nov 8, 2024
5 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
2 participants