pyconll / pyconll / 18862881474

98%

push

github

Initial Implementation of Structural Parsing (#146)

* Step 1 via cc. Change parsing logic into a class.

* Move parsing out of Token class.

* Lint fixes.

* Move everything up to 3.12 and small start on structural schema logic.

* Some wip.

* Attempt before reformat.

* Update formatting logic properly.

* Pretty good progress I'd say.

* MFer, it works.

* Still need to do more hardening with handling and surfacing errors, but it works for ConlluToken unit and int tests.

* Getting a move on.

* Fix version number and add more to CHANGELOG.

* Working through it.

* More implementation improvements.

* Handle more errors and inconsistencies, improve error handling distinctions, etc.

* Annoying string reformatting.

* More improvements.

* Fix all tests, make an opt-out for all the file hashing, and add a __repr__ for the generated classes.

* Work on restructuring modules.

* Proper annotation of public API, and fix ABC issue on Conllable so that type will be able to be piped in from top level.

* Revert load API to avoid breaking change, and change __init__ imports, and add custom schema descriptor.

* Small linting changes.

* Change minimum required version to 3.13 and start moving around code for parsing consolidation.

* Get mypy to shut-it.

* Update loader.

* Change tests back.

* Move back and consider consolidated and time to bring up quality before next checkpoint.

* Add new command.

* Update versions again.

* Big linting changes.

* Rename to better match concepts and fix syntax error.

* Remove legacy parsing.

* Almost there.

* More safe name generation.

* Move up python version to take advantage of template strings and increase safety.

* Change things to t-strings for safety and nail down the design finally.

* Try to address linting errors.

* Maybe python 3.14 was a mistake :).

* Update documentation build version.

* Add tests, small optimization, and fix line length and bug.

* Resign to the fact that Parse and FormatErr... (continued)

372 of 383 new or added lines in 12 files covered. (97.13%)

754 of 773 relevant lines covered (97.54%)

0.98 hits per line