Feature/groq gemini fallback

.gitignore

@@ -1,2 +1,45 @@
+# ─── Secrets (NEVER commit real keys) ───────────────────────────
+.env
+backend/.env
+frontend/.env.local
+# Keep .env.example — it has placeholder values only
+
+# ─── Database ───────────────────────────────────────────────────
+*.db
 codechat.db
-*.log
+
+# ─── Logs ───────────────────────────────────────────────────────
+*.log
+logs/
+
+# ─── Python ─────────────────────────────────────────────────────
+__pycache__/
+*.py[cod]
+*.pyo
+*.pyd
+*.egg-info/
+dist/
+build/
+.eggs/
+*.egg
+.venv/
+venv/
+env/
+
+# ─── FAISS indexes (can be large, regenerated on demand) ────────
+backend/indexes/
+
+# ─── Node / Next.js ─────────────────────────────────────────────
+node_modules/
+frontend/.next/
+frontend/out/
+frontend/build/
+.npm
+
+# ─── OS / Editor ─────────────────────────────────────────────────
+.DS_Store
+Thumbs.db
+.idea/
+.vscode/
+*.swp
+*.swo

README.md

@@ -35,7 +35,7 @@ In a separate terminal:
 
 ```bash
 ollama serve
-ollama pull llama3.2:1b
+ollama pull qwen2.5-coder:1.5b-instruct-q4_K_M
 ```
 
 ## 5) Frontend setup

backend/.env.example

@@ -9,13 +9,23 @@ DATABASE_URL=sqlite:///./codechat.db
 
 # ─── Ollama ───────────────────────────────────────────────────────────────────
 OLLAMA_BASE_URL=http://localhost:11434
-OLLAMA_REQUEST_TIMEOUT=240
-OLLAMA_MODEL=llama3.2:3b
+OLLAMA_REQUEST_TIMEOUT=45
+OLLAMA_MODEL=qwen2.5-coder:1.5b-instruct-q4_K_M
 LLM_MAX_TOKENS=1200
-RAG_TOP_K=8
+RAG_TOP_K=3
 EMBEDDING_LOCAL_ONLY=true
 
 # ─── Optional cloud LLM keys ─────────────────────────────────────────────────
+# Fallback chain: Ollama → Gemini → Groq (automatic, requires keys below)
+
+# Gemini (primary cloud fallback) — get key at aistudio.google.com
+GEMINI_API_KEY=your_gemini_api_key_here
+GEMINI_MODEL=gemini-2.0-flash
+
+# Groq (secondary fallback, free tier) — get key at console.groq.com
+GROQ_API_KEY=your_groq_api_key_here
+GROQ_MODEL=llama-3.1-8b-instant
+
 # Only needed if using llm_provider=openai or anthropic
 # OPENAI_API_KEY=
 # ANTHROPIC_API_KEY=

backend/README.md

@@ -149,8 +149,8 @@ JWT_SECRET_KEY=
 JWT_ALGORITHM=HS256
 ACCESS_TOKEN_EXPIRE_MINUTES=30
 OLLAMA_BASE_URL=http://localhost:11434
-OLLAMA_REQUEST_TIMEOUT=240
-OLLAMA_MODEL=llama3.2:3b
+OLLAMA_REQUEST_TIMEOUT=45
+OLLAMA_MODEL=qwen2.5-coder:1.5b-instruct-q4_K_M
 LLM_MAX_TOKENS=1200
 RAG_TOP_K=8
 EMBEDDING_LOCAL_ONLY=true

backend/app/rag/gemini_client.py

@@ -0,0 +1,16 @@
+import os
+from typing import Optional
+from google import genai
+
+class GeminiClient:
+    def __init__(self, api_key: str, model_name: str = "gemini-1.5-flash"):
+        self.client = genai.Client(api_key=api_key)
+        self.model_name = model_name
+
+    def generate(self, prompt: str, system_prompt: Optional[str] = None) -> str:
+        full_prompt = f"{system_prompt}\n\n{prompt}" if system_prompt else prompt
+        response = self.client.models.generate_content(
+            model=self.model_name,
+            contents=full_prompt
+        )
+        return response.text