fix: safe weight parsing to prevent ValueError and remove redundant typecasts (#804)

[suhaniiz]

📋 PR Checklist

Thank you for contributing to PDF-Assistant-RAG! 🎉
Please fill out this template before submitting. PRs without it filled in will be closed.

---

🔗 Related Issue

Closes #804

---

📝 What does this PR do?

This PR resolves an inconsistent and fragile weight accumulation pattern inside get_entity_context that left the graph pipeline vulnerable to unexpected crashes.

Changes introduced:

• Replaced the fragile int(edge.get("weight", 1)) assignment with a robust parsing pattern (int(float(raw_weight))) wrapped inside a structured try-except block. This safely handles float inputs, float strings (e.g., "2.5"), or arbitrary missing properties without raising a ValueError or TypeError.
• Simplified the cumulative logic to use a clean += operator on the safely handled scalar.
• Removed the redundant int() typecasting overhead from the final sorted() key lambda loop, as dictionary weight properties are now guaranteed numerical entities.

---

🗂️ 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 (uvicorn app.main:app --reload)
• Tested the affected API endpoints manually with varied edge weight types
• Added / updated tests

---

📸 Screenshots (if UI change)

N/A

---

⚠️ Anything to flag for reviewers?

The defensive try-except fallback assumes a standard default weight value of 1 if an unparseable data variant appears inside the knowledge graph attributes, keeping context generation intact without bringing down the core RAG server.

---

✅ 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

[vercel]

@suhaniiz is attempting to deploy a commit to the param20h's projects Team on Vercel.

A member of the Team first needs to authorize it.