Merging Apertium Systems

[AsH1605]

PR: Merging Apertium Systems with Sample Implementation

Summary of Changes

• Implemented a solution to merge two Apertium systems using the sample systems spa-cat and eng-deu as test cases.
• Created the following scripts to facilitate the merging process:
  1. merge_apertium_systems.py: Handles the overall merging logic for systems.
  2. merge_json.py: Merges JSON configuration files related to Apertium systems.
  3. merge_modes.py: Deals with mode-specific merging of configurations.
  4. merge_configure.py: Merges system-specific configuration files.
  5. merge_makefile.py: Handles merging of Makefiles for the build process.

Testing

• The implementation was tested using the sample systems:
  • spa-cat: Spanish to Catalan system.
  • eng-deu: English to German system.
• Successfully generated merged configurations and scripts for these sample systems.

Request for Feedback

1. Is the Implementation Correct?
   Could you review the current implementation to confirm if it aligns with the expected workflow for merging Apertium systems?

2. Next Steps:
   If this approach is correct, how should I proceed for integration?

Files Added

• merge_apertium_systems.py
• merge_json.py
• merge_modes.py
• merge_configure.py
• merge_makefile.py

Sample Systems Used for Testing

• spa-cat: Spanish ↔ Catalan
• eng-deu: English ↔ German

[mr-martian]

Could you please explain what problem this feature solves?

[AsH1605]

Thank you for your question, @mr-martian . This PR is not directly linked to a specific issue but is based on the discussion in this thread with @TinoDidriksen.

[TinoDidriksen]

What I mentioned in #36 was merging the codebases of https://github.com/TinoDidriksen/regtest and this repo, not merging test suite datasets.

I don't think we would want to merge datasets. If anything, we would like to split datasets into their own repositories, so they don't clog up cloning languages and pairs, and so their history can be trimmed independently. Regression datasets get very large, very fast.

[AsH1605]

@TinoDidriksen Thank you for the clarification regarding #36! I’ll focus on this and make the necessary adjustments.