Releases: amir-zeldes/DepEdit
Releases · amir-zeldes/DepEdit
V3.4.0.1 - bugfix
This is a bugfix release
- Did not retain ,= values on append when other key names were already present in the input for the targeted field
V3.4.0.0 - add ,= operator
- Add ,= action operator for concatenating multiple values to a nested key (e.g. Cxn,=Ditransitive)
- Bugfix: allow = in definition values
V3.3.1 - bug fix for capturing groups
Bug fix capturing groups
V3.3.0 - Rudimentary split support, bugfixes and speedups
- Experimental support for splitting tokens (adding a token) using split= action
- Added storage3 field for additional hidden storage
- Faster execution by earlier abandonment of non-matching rules and definition priority sorting
- Bugfixes
V3.2.1 - bug fix
This is a bug fix release addressing cases in which existing enhanced dependencies could not be overwritten by additional enhanced dependencies between the same pair of tokens.
- multiple edits to the same token pair are now possible in sequence
- only the latest write operation is retained in the rule output
V3.2.0 - feature subtraction operator and minor fixes
- Added
-=
subtraction operator for removing features from FEATS and MISC - Better ordering for comment lines
- Performance improvements and bug fixes
V3.1.0 - more MWT support, --text option and module invocation
- Add support for making supertokens (MWTs) with
><
operator - Add option
-t, --text
to generate sentence# text =
comment - Add storage and storage2 convenience fields (can be used to save intermediate manipulations, do not get serialized)
- Add operators for querying and adding multiple concurrent enhanced dependencies (
#1~#2
, fieldsedep
andedom
, see documentation) - Make invokable as a module (
python -m depedit
) when installed
V3.0.0 - improved enhanced dependencies, piped values and more
V3.0.0
- better enhanced edge manipulation with
edep
,edom
and~
operator - support pipe-separated features using
+=
action - implement UD validator-conforming feature and edep sorting
- use
once
directive to apply rules only once per sentence - added
storage
pseudo field for temporary value storage - added examples for English morphology and enhanced dependency enrichment
- minor bugfixes and updated documentation
V2.2.0.1 - sentence annotations as conditions + bugfix
Added conditions on sentence annotations, for example:
tok=/!/;#S:type=imperative #2>#1 #1:xpos=IMP_MARKER
i.e. find a sentence with a comment # type = imperative
and only then apply the rule to the contained token '!'
This adds a bugfix to 2.2.0.
V2.2.0 - sentence annotations as conditions
Added conditions on sentence annotations, for example:
tok=/!/;#S:type=imperative #2>#1 #1:xpos=IMP_MARKER
i.e. find a sentence with a comment # type = imperative
and only then apply the rule to the contained token '!'