Skip to content

Release v. 0.1.8 - #372

Merged
sundy1994 merged 44 commits into
mainfrom
dev
Jul 20, 2026
Merged

Release v. 0.1.8#372
sundy1994 merged 44 commits into
mainfrom
dev

Conversation

@xehu

@xehu xehu commented Jul 15, 2026

Copy link
Copy Markdown
Collaborator

Merge dev into main, releasing a new version of the package.

Key Changes

  • Robustness for politeness features such as WH question detection
  • Improved logging capability
  • Added functionality to drop redundant (highly correlated) columns automatically
  • Removed functionality for processing partial conversations (e.g., first %)
  • Fixed bug with information diversity, in which the metric was reversed
  • Update documentation

xehu and others added 30 commits June 3, 2025 11:46
sundy1994 and others added 14 commits May 7, 2026 09:43
…ver the FB-generated cols (not all cols/metadata); (2) edit the log headers to specify what file the log is for
The test_drop_redundant_columns.py tests (committed in 93b074e) read this
CSV, but it was silently skipped by the *.csv .gitignore rule, leaving the
test without its fixture on the remote. Force-add it so the test is runnable.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The blanket *.csv ignore rule silently dropped new test fixtures (e.g.
test_redundant_columns.csv), requiring git add -f. Negate the rule for the
fixture directory so fixtures stage normally; CSVs elsewhere stay ignored.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The test workflow invokes pytest file-by-file, so test_drop_redundant_columns.py
was never collected. Add it to the Run tests step.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add structured logging and feature output diagnostics to FeatureBuilder
The pre_install step installed spacy==3.7.2 in isolation, before the
numpy<2.0.0 constraint in requirements.txt applied. pip then resolved
numpy 2.x, which is ABI-incompatible with spacy 3.7.2's numpy-1.x-built
wheels, causing `python -m spacy download en_core_web_sm` to crash on
import and failing the Read the Docs build.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
`python -m spacy download en_core_web_sm` resolves the model version via
a runtime compatibility lookup that came back empty, producing a
malformed URL (.../download/-en_core_web_sm/-en_core_web_sm.tar.gz) and
a 404 that failed the RTD build. Pin and pip-install the matching model
wheel (en_core_web_sm-3.7.1, compatible with spacy 3.7.2) directly,
which is more reproducible and avoids the lookup entirely.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Fix inverted information diversity score (cosine distance vs. similarity)
@xehu
xehu requested a review from sundy1994 July 15, 2026 18:34
@sundy1994
sundy1994 merged commit 2b20678 into main Jul 20, 2026
1 check passed
@xehu
xehu deleted the dev branch July 20, 2026 16:33
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.

2 participants