Skip to content

[pull] master from stanfordmlgroup:master#1

Open
pull[bot] wants to merge 313 commits intoFoundryAI:masterfrom
stanfordmlgroup:master
Open

[pull] master from stanfordmlgroup:master#1
pull[bot] wants to merge 313 commits intoFoundryAI:masterfrom
stanfordmlgroup:master

Conversation

@pull
Copy link

@pull pull bot commented Jul 17, 2020

See Commits and Changes for more details.


Created by pull[bot] (v2.0.0-alpha.4)

Can you help keep this open source service alive? 💖 Please sponsor : )

ryan-wolbeck and others added 30 commits January 31, 2024 09:57
Fix issue causing large memory consumption in pred_dist()
* Add support for python 3.12

* Drop support for python 3.9

* Bump flake8 and pylint
Release 0.5.0
* Implemented NormalFixedMean

* Added NormalFixedVar and NormalFixedMean to test_distns.py

* Match log(sigma) parameterization of Normal class

* Update __init__.py

Linter is demanding a blank line for ending a file (sigh)
Makefile updates and adding support for FixedNormalMean
feat: Add estimator type with reg and clf for NGBClassifier and NGBRegressor
* Update partial_fit to respect validation data

Existing logic ignores passed validation data, the if erarly_stopping_rounds control sequence is updated to respect the passed validation data.

Co-authored-by: jerome-f <irudayaf@sc1nc008is02.eth.rsiec.sc1.science.roche.com>
* v0.5.5 release
The two parameters `validation_fraction` and `early_stopping_rounds` have been added to the `params` dictionary.
Update parameters in `get_params`
Co-authored-by: Mahmoud <mhegab@raisaenergy.com>
feat: added Weibull distribution
* Enabling `python=3.13`

* Increase lower bounds of `numpy` and `scipy` for `python=3.13` environments.
* Update dev dependencies and improve pyproject.toml configuration

- Update dev dependencies to latest versions:
  - pytest: ^6.1.2 → ^8.0.0
  - black: ^22.8.0 → ^24.0.0
  - pre-commit: ^2.0 → ^4.0.0
  - isort: ^5.6.4 → ^5.13.0
  - pylint: ^3.0.3 → ^3.2.0
  - flake8: ^7.0.0 → ^7.1.0

- Migrate to modern Poetry format: [tool.poetry.dev-dependencies] → [tool.poetry.group.dev.dependencies]
- Add [tool.black] configuration with explicit line length and target versions
- Improve Python version markers for numpy/scipy using PEP 508 markers

Closes #379

* Fix Black target-version configuration

Remove py312 and py313 from target-version as they're not supported
by the current Black version in CI environment

* Update pre-commit configuration

- Update Black version from 22.8.0 to 24.10.0 to match pyproject.toml
- Fix deprecated default_stages: commit -> pre-commit

* Apply linting fixes

- Apply Black formatting changes (blank lines after docstrings)
- Update code formatting to match updated linter configuration

* Update pre-commit-hooks to v6.0.0

- Fix deprecated stage names warning
- Update from v2.2.3 to v6.0.0 to use modern pre-commit hooks

* update release
* Fix NumPy 2.x natural gradient compatibility issue

- Replace vectorized np.linalg.solve with sample-wise computation in scores.py
- Add robust dimension checking and error handling
- Handle singular matrix cases with pseudo-inverse fallback
- Add comprehensive test suite for NumPy 2.x compatibility
- Fixes GitHub issue #384: dimension mismatch in natural gradient computation

The original issue occurred when np.linalg.solve received incompatible
dimensions between metric and gradient arrays. The new implementation:
1. Validates dimensions before computation
2. Solves linear systems sample-by-sample for better stability
3. Provides clear error messages for dimension mismatches
4. Handles singular matrices gracefully

All existing tests pass and new compatibility tests verify the fix.

* fix veresion

* Fix linting issues in test_numpy2_compatibility.py

- Fix unbalanced tuple unpacking in make_regression call
- Replace broad Exception catching with specific ValueError and LinAlgError
- All linter checks now pass

* Fix trailing whitespace in scores.py

- Remove trailing whitespace that was causing GitHub CI to fail
- All linter checks now pass on Python 3.11

* tests: make_regression unpack (X, y) for sklearn compatibility across versions

* tests: request coef=True in make_regression and unpack 3 values to satisfy pylint unbalanced-tuple-unpacking across sklearn versions

* Fix natural gradient shape handling without performance hit

* Delete .evn
* Adding Sympy factory and example notebooks and doc

* cleanup helpers from local testing

* Update GitHub Actions and refactor code for consistency

- Upgrade actions/checkout from v2 to v4 and actions/setup-python from v1 to v5 in pythonpackage.yml
- Refactor list comprehensions in betabinomial.py and logitnormal.py for improved readability
- Clean up imports and formatting in sympy_utils.py and __init__.py
- Remove unnecessary blank lines in test_sympy_existing_distns.py

* Add sympy dependency to project configuration

* Fix Copilot review issues and add Normal mixture example

Address three numerical stability issues from PR #403 review:
- Make Y-clamping conditional on log(y) presence in mixture path
- Use nan_to_num with finite clamps instead of zeroing gradients
- Tighten _fit_mle lower bound from exp(-700) to exp(-150)

Add example notebook for 3-component Normal mixture demonstrating
that mixtures over R correctly handle negative observations.

* Run black formatter on sympy_utils.py

* Enhance MLE fitting by including extra parameters and simplify sampling method

* Validate class probabilities in make_distribution to ensure finiteness and positivity
* update to  Python 3.14
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

⤵️ pull merge-conflict Resolve conflicts manually

Projects

None yet

Development

Successfully merging this pull request may close these issues.