Ai/codex search relevance

[SmileJune]

Summary by CodeRabbit

• Documentation

  • Added service link to the TechCase application in the README.

• New Features

  • Added "OpenAI Codex" as a searchable technology with support for common aliases like Codex CLI and related keywords for improved discoverability.

• Tests

  • Expanded test coverage for search functionality, including query matching, expansion, and intelligent fuzzy matching behavior.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: bc84cb7d-0b37-448b-91de-886acf96f53e

📥 Commits

Reviewing files that changed from the base of the PR and between ff44339 and 911a795.

📒 Files selected for processing (4)

• README.md
• apps/backend/app/keywords/dictionaries.py
• apps/backend/app/search/evaluation/queries.json
• apps/backend/tests/test_search_query_rules.py

---

📝 Walkthrough

Walkthrough

This PR adds OpenAI Codex as a searchable technology keyword. It defines a keyword rule recognizing "codex" and variants, adds evaluation test data with expected search results, and provides unit tests validating keyword matching and query behavior.

Changes

OpenAI Codex Search Support

Layer / File(s)	Summary
OpenAI Codex keyword definition apps/backend/app/keywords/dictionaries.py	KEYWORD_RULES is extended with a new KeywordRule entry for "OpenAI Codex" (type: "technology") and four aliases: "openai codex", "codex", "codex cli", "openai codex cli".
Search evaluation test data apps/backend/app/search/evaluation/queries.json	A new query object technology-openai-codex is inserted into the evaluation dataset with keyword "codex", category "technology", Korean intent description, and two expected result entries defining baseline search behavior.
Unit tests for search query rules apps/backend/tests/test_search_query_rules.py	Three test functions verify that "codex" matches "OpenAI Codex", query expansion works correctly, "codex" searches omit fuzzy matching, and "code" searches retain it. A recursive contains_key helper enables checking for keys in nested data structures.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Poem

🐰 A codex hops through keyword rules,
Aliases dance like woodland pools,
Tests confirm our fuzzy friend knows
Where OpenAI's great wisdom flows! ✨

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch ai/codex-search-relevance

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}