feat: implement semantic sub-query decomposition and parallel retrieval with RRF merge

[knoxiboy]

📋 PR Checklist

---

🔗 Related Issue

Closes #566

---

📝 What does this PR do?

Implements semantic sub-query decomposition for multi-part user questions:

1. Updates transform_query inside retriever.py to use a dedicated prompt asking the LLM to identify distinct semantic components in a complex query (e.g. comparing topics) and return them as a list of sub-queries.
2. Executes dense vector and BM25 queries in parallel (using concurrent.futures.ThreadPoolExecutor) for each sub-query.
3. Merges candidate lists from all sub-queries using a generalized Reciprocal Rank Fusion (RRF) algorithm.

---

🗂️ Type of Change

• 🐛 Bug fix
• ✨ New feature
• 🔧 Refactor / code cleanup
• 📝 Documentation update
• 🎨 UI / styling change
• ⚙️ CI / tooling / config change
• 🧪 Tests

---

🧪 How was this tested?

• Ran the backend locally
• Verified retrieval pipeline manually

---

📸 Screenshots (if UI change)

---

⚠️ Anything to flag for reviewers?

None.

---

✅ Self-Review Checklist

• My branch is based on dev, not main
• I have not added any secrets / API keys
• I have not modified main branch or any HuggingFace deployment config
• My code follows the existing style (no unnecessary formatting changes)
• I have updated relevant docs / comments if needed