Skip to content

Conversation

@ajdapretnar
Copy link
Collaborator

@ajdapretnar ajdapretnar commented Jan 6, 2026

Issue

Remove gensim to avoid installation issues.
Fixes #1127.

Description of changes
  • Move gensim's Dictionary into a separate Orange class (copy with some methods removed).
  • Fix bag of words to work with sklearn's methods.
Includes
  • Code changes
  • Tests
  • Documentation

@janezd
Copy link
Contributor

janezd commented Jan 6, 2026

I just noticed that the network add-on also requires gensim to compute graph embeddings. To get rid of it there, we'd need to make the graph embeddings widget "optional"? Or is there an alternative library we can use?

@ajdapretnar
Copy link
Collaborator Author

ajdapretnar commented Jan 6, 2026

@janezd I see two tentative options:

  • replace word2vec with an alternative distribution (https://pypi.org/project/word2vec/) Scrap that, it is completely unmaintained.
  • use a different embedder (FastText or alike)
    But I am no expert here. Essentially, we should find a good sequence model, from how I understand it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Remove Gensim

2 participants