Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.

[1.6.1] - 2022-09-10

• Multi-word tokenization integrated with menard.parse/parse-start.
• Whitespace code formatting cleanup.

[1.6.0] - 2022-09-04

Core

• Support multi-word tokenization in menard.parse/parse
• Introduce new model .edn format with :name, :language, :morphology, :grammar :lexicon and models in this format for nl and en.
• Update a few dependencies.
• Remove use of cljslog; use new built-in menard.log instead
• Add new woordenlijst language model and lexicon in nl and en and intermediate source files.
• Move tenses from menard.nederlands to their own namespace menard.nederlands.tenses.

Server

• Update a few dependencies.
• Improve startup script: add MENARD_DIR where needed and add comments.

[1.5.0] - 2022-03-13

• Fix/refactor reflexive pronouns.
• Allow for reloading of models based on polling filesystem changes using babashka/fs.
• Incremental parsing so that partial parses can be shared between clients and servers.
• All keys taking only true/false are suffixed with '?'
• Move all linguistic structures (e.g. ug, subcat) to .edn files in resources.

[1.4.5] - 2021-08-7

• Fix typo

[1.4.4] - 2021-08-7

• More work on simple past verbs
• Refactor server-side support to support both ring and AWS Lambda

[1.4.3] - 2021-06-16

• Rewrite support for simple past verbs in Dutch.

[1.4.2] - 2021-06-05

• Rewrite lexical rules to unify all rules into a single file per language.

[1.4.1] - 2021-05-11

• Add lambda directory for deploying API to Amazon Lambda.
• Improve support for reflexive pronouns.
• Add more vocabulary.

[1.4.0] - 2021-01-07

• Move linguistic resources out of src/ and into resources/.

[1.3.5] - 2020-11-14

• Move all linguistic resources to resources/, to better emphasize that they are data, not executable code.

[1.3.4] - 2020-11-7

• Fix synchronization of model-loading, so a model is only added once rather than once per-thread, if the menard libary is used in a multi-threaded application or service.

[1.3.3] - 2020-09-13

• Upgrade to latest dag_unify release.
• Fix and refactor Dutch adjective morphology.

[1.3.2] - 2020-09-12

• Upgrade to latest dag_unify release.
• Add morphology for dik <-> dikke.

[1.3.1] - 2020-08-15

• Improve loading and reloading language models
  • block on model-creation if it doesn't exist yet.
  • only load model once unless doing reloading.
• Apostrophe-s pluralization applies to all Dutch nouns that end in 'y' or vowels (not just 'y'), but 'vowels' doesn't include 'e' for the purpose of this rule.

[1.3.0] - ??

• Release tag and commit for this release missed somehow.

[1.2.1] - 2020-07-26

• Fix bugs in Dutch morphology of adjectives and nouns
• Add new lexemes
• Separate intensifiers and exclamations into their own files for readability
• Improve code readability with variable name changes and value usages in menard/generate
• Improve readability of noun lexemes in Dutch by defaulting gender to :common if not specified

[1.2.0] - 2020-07-14

• Fix bugs related to dynamic model-loading
• Turn down logging

[1.1.0] - 2020-07-12

• Add new dynamic model-loading, so that a java process can update its linguistic state from the filesystem, without restarting the whole JVM to get an updated linguistic model.

[1.0.3] - 2020-07-09

• Fix noun pluralization errors in Dutch (e.g "been", "fiets", "plaats").

[1.0.2] - 2020-07-07

• Prevent overmatching on default pluralization parsing rule in Dutch.

[1.0.1] - 2020-07-07

• Reduce redunancy in lexical source code files
• Add more lexical entries

[1.0.0] - 2020-06-30

• Dutch adjective morphology fixes:
• eenzaam -> eenzame
• constrain m -> mm (e.g. 'stom' -> 'stomme') : 'm' must be preceded by 't'

[0.0.4] - 2020-06-29

• ~analyze: create lower,upper, and capitalized variants of all inputs
• ~parse: use (newly-modified) analyze for parsing's lookup function
• fix for one sense of irregular plural nouns: :quant should be :some, not :unspec
• all possessive pronouns need to have both :sing and :plur senses
• cleanup whitespace
• move demo output to separate file

[0.0.3] - 2020-06-21

• Rename from 'babylon' to 'menard'
• Update README.md with information about new name

[0.0.2] - 2020-06-20

• Add some new lexemes.
• Add :example key to nederlands/expressions.edn where appropriate.
• Upgrade to latest dag_unify version available.
• Add CHANGELOG.md

[0.0.1] - 2020-06-09

First release.