Dynamic Choice Detection in convert.py

[abhijit9040]

The ConvertVars class in scripts/convert.py uses hardcoded lists to define valid languages and versions:

LANGUAGE_CHOICES = ["all", "en", "es", ...]
VERSION_CHOICES = ["all", "latest", "1.0", "1.1", ...]

This creates several issues:

• Manual maintenance required whenever a new language or version is added.
• Risk of inconsistency if new files are added to source/ but not reflected in the script.
• Low discoverability, as contributors may not realize the script must also be updated.

Proposed Solution

• Refactor ConvertVars to dynamically detect available editions, versions, and languages at runtime by scanning project directories.

Requirements

• Scan source/ directory for .yaml files.
• Extract metadata from filenames (edition-layout-version-language pattern) or from YAML meta blocks.
• Populate choice lists dynamically during class initialization.
• Preserve special logic for all and latest.

Files Involved

• scripts/convert.py (specifically the ConvertVars class)

[abhijit9040]

Hi @sydseter , can you review this issue ?

[abhijit9040]

HI @sydseter , can i work on this issue ?

[Arunodoy18]

Hi 👋

I’d like to work on this.

My plan is to refactor ConvertVars to dynamically detect available editions, versions, and languages by scanning the source/ directory at runtime instead of relying on hardcoded lists.

Proposed approach:

• Scan source/ for .yaml files
• Parse filename structure (edition-layout-version-language)
• Optionally validate against YAML meta blocks for correctness
• Dynamically populate LANGUAGE_CHOICES and VERSION_CHOICES during initialization
• Preserve special handling for all and latest

This will remove manual maintenance requirements and prevent inconsistencies when new files are added.

Please let me know if this aligns with the intended direction before I begin implementing.

[sydseter]

@abhijit9040 Sorry for the late response. This actually something that I have been looking at trying to fix. The meta field in mapping file is the best solution that I have been looking at. E.g defining language and version there and generate the correct files based on the mapping.

[abhijit9040]

Hi @sydseter, thanks for the feedback!
Using the meta block as the source of truth sounds like a better approach.

I can update convert.py to read the meta data from the mapping files and build the choice lists dynamically, while keeping all and latest working as before.