Skip to content

Conversation

@vankesteren
Copy link
Member

No description provided.

vankesteren and others added 30 commits November 25, 2024 13:25
* Improve error messaging for variable fitting

* Fix erroneous mypy error
* Split varspec parameter into config

* Update documentation and mypy/ruff
* Also add the schema to the CI, so we know when the schema has changed.
* Add minimum 95% testing coverage (currently at 95.03%).
* Improved testing.
* Variable type can now be inferred from the distribution even if the distribution supports multiple variable types.
* MetaVar.get_var_type now returns a TypeError instead of a ValueError for unsupported types.
* A more appropriate error is thrown when a distribution is created with missing parameters.
* The UniqueDistributionMixin now implements __new__ instead of __init__, so that the mixin does not cause the input parameters to change to *args, **kwargs. This should not have any effect on the user, except that the documentation and signature will now give the correct parameters for the unique faker and regex distributions.
* Add option to set encoding

* Fix ruff issue
* Check the filename before synthesizing

* Make mypy happy
Also greatly improve performance of synthesis for most numerical distributions.
* Adds write_* functions to write a dataframe to a file.
* Add distribution module to exports for type completion

* also add privacy to module exports

* also add file to module exports
qubixes and others added 11 commits October 2, 2025 13:59
- Pyreadstat > 1.3 now reads polars dataframes, so we are not transforming to pandas for .sav and .dat files.
-  Require Python >= 3.10. Python 3.9 will be deprecated tomorrow, and pyreadstat is not available for 3.9 anymore.
-  Fix the coverage testing, now at almost 96%.
-  Fix the typing for the unique key distribution.
* Check whether we can remove pandas

* Keep pandas for testing
- update to 3.13
- add dockerignore for much smaller size
- add --no-compile and --no-cache-dir to pip call
* Small updates to the documentation

* add sphinx-autobuild for documentation writing

uv run sphinx-autobuild is nice

* Update docs: index, quickstart, and what is page

* updates to datafree and improve_synth

* last small updates

* Rename extension -> plugin

Plus plugin -> plug-in

* Fix more issues in rst files

* Remove optional

* Remove all , optional in docstring

It seems to give issues

* Fix reference

---------

Co-authored-by: Raoul Schram <r.d.schram@uu.nl>
* Rework fitter/distribution finding

* Remove old code

* Remove debug statement

* Fix distribution uniqueness

* Update metasyn/registry.py

Co-authored-by: Erik-Jan van Kesteren <e.vankesteren1@uu.nl>

* Rework behavior and better error message

* Update docstring

---------

Co-authored-by: Erik-Jan van Kesteren <e.vankesteren1@uu.nl>
* Rename distribution registries to plugins

Only for user facing functions and methods.

* Fix dist_registries key
* Improve the provenance information in the GMF

* Fix ruff

* Remove comments

* Update example GMF files

* Update example_gmf_titanic.json

---------

Co-authored-by: Erik-Jan van Kesteren <erikjanvankesteren@pm.me>
@vankesteren vankesteren merged commit c2a21de into main Oct 16, 2025
22 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

Development

Successfully merging this pull request may close these issues.

3 participants