Feature/issue 1090

[riyanshigupta890-cloud]

Summary [required]

This PR implements configurable recommendation scoring weights, allowing non-developers to tune the recommendation engine without modifying code. Previously, scoring weights were hardcoded as magic numbers in utils/recommender.py, making it impossible to adjust them without code changes or test different scoring strategies.

The solution adds a JSON configuration file and a config loader in src/config.py that reads weights at runtime. Weights are cached for performance, and can be overridden via environment variables for different deployments.

Related Issue [required]

Closes #1090

Type of Change [required]

• Bug fix
• Feature — adds new functionality
• Data
• Documentation
• Style
• Refactor
• Test

What Was Changed [required]

File	Change made
src/config.py	Added load_recommendation_weights() method and get_recommendation_weights() function to load weights from JSON config file with validation and caching
utils/recommender.py	Removed hardcoded SCORING_WEIGHTS dictionary; updated score_single_project() to load weights from config at runtime
config/recommendation_weights.json	NEW FILE — stores configurable scoring weights with default values (skill: 3, level: 2, interest: 2, time: 1)

How to Test This PR [required]

1. Clone this branch: git checkout feature/issue-1090
2. Verify config loads correctly: python verify_config.py
3. Expected output:

Expected test output:

27 passed, 0 failed out of 27 tests

Test Results [required]

All existing tests continue to pass. The new config loading was verified with the verification script:
✓ Testing config loading...

✓ Successfully imported get_recommendation_weights

✓ Loaded weights: {'skill': 3, 'level': 2, 'interest': 2, 'time': 1}

✓ Weights are correct!

✅ ALL CHECKS PASSED - Configuration is working!

Screenshots (if UI change)

N/A — Backend feature only

Self-Review Checklist [required]

• I have read CONTRIBUTING.md and followed all guidelines
• My branch name follows convention: feature/issue-1090
• Config loading verified with verify_config.py — all checks pass
• No debug print statements introduced
• All new functions have docstrings
• Changes are scoped to Issue [Feature]: Make Recommendation Scoring Configurable #1090 only
• Weights are cached to avoid file I/O on every request
• Environment variable override supported for different deployments

Notes for Reviewer

• The weights are loaded once on first call and cached in a module-level variable for performance
• The config file location can be overridden via the RECOMMENDATION_WEIGHTS_PATH environment variable, enabling different weight configurations per deployment
• The implementation validates that all required weight keys exist and that all values are positive numbers
• No breaking changes — all existing functionality preserved

[vercel]

@riyanshigupta890-cloud is attempting to deploy a commit to the komalsony234-1530's projects Team on Vercel.

A member of the Team first needs to authorize it.