File size: 9,212 Bytes
2379054 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 | """
OPTIMIZED ENGINE - User Selectable: Gemini / Local / Hybrid
"""
import os
import sys
from pathlib import Path
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
try:
from dotenv import load_dotenv
load_dotenv()
except Exception as e:
pass
from openai import OpenAI
from typing import List, Dict
import chromadb
from sentence_transformers import SentenceTransformer
from mcp.cache_manager import _cache
from mcp.token_optimizer import _optimizer
class OptimizedEngine:
def __init__(self, engine_type: str = "hybrid", model_name: str = "gemini-3.5-flash"):
"""
Initialize optimized engine
Args:
engine_type: "gemini", "local", or "hybrid"
model_name: Gemini model name
"""
self.engine_type = engine_type
self.model_name = model_name
self.api_key = os.environ.get("GEMINI_API_KEY", "")
self.use_gemini = False
# Setup Gemini if available
if self.api_key:
try:
self.client = OpenAI(
api_key=self.api_key,
base_url="https://generativelanguage.googleapis.com/v1beta/openai/"
)
self.use_gemini = True
print("β
Gemini API connected")
except Exception as e:
print(f"β οΈ Gemini connection failed: {e}")
self.use_gemini = False
else:
if engine_type in ["gemini", "hybrid"]:
print("β οΈ GEMINI_API_KEY not set. Falling back to local.")
self.engine_type = "local"
# Initialize RAG
self.embedding_model = SentenceTransformer('all-MiniLM-L6-v2')
self.chroma_client = chromadb.PersistentClient(path="knowledge/chroma_db")
try:
self.collection = self.chroma_client.get_collection("track_design_knowledge")
print("β
Connected to knowledge base")
except Exception as e:
self.collection = None
print("β οΈ Knowledge base not found")
# Show active engine
if engine_type == "gemini" and self.use_gemini:
print("π Engine: Gemini (Cloud) - Fast, Accurate")
elif engine_type == "local":
print("π Engine: Local LLM (Offline) - Free, Works without internet")
elif engine_type == "hybrid":
if self.use_gemini:
print("π Engine: Hybrid (Gemini + Local Fallback)")
else:
print("π Engine: Hybrid (Local only - no API key)")
def search(self, question: str, top_k: int = 3) -> List[Dict]:
if not self.collection:
return []
q_embed = self.embedding_model.encode([question])
results = self.collection.query(
query_embeddings=q_embed.tolist(),
n_results=top_k * 2
)
formatted = []
if results['documents'] and results['documents'][0]:
for i in range(len(results['documents'][0])):
formatted.append({
"content": results['documents'][0][i],
"metadata": results['metadatas'][0][i],
"score": 1 - results['distances'][0][i] if results['distances'] else 0
})
return _optimizer.select_best_chunks(formatted, question)
def ask(self, question: str, top_k: int = 3, force_refresh: bool = False) -> Dict:
# 1. Check cache
if not force_refresh:
cached = _cache.get(question, top_k)
if cached:
return {
"answer": cached.get('answer', ''),
"sources": cached.get('sources', []),
"confidence": cached.get('confidence', 0),
"cached": True,
"model_used": cached.get('model_used', 'cache')
}
# 2. Search and optimize
results = self.search(question, top_k)
if not results:
return {
"answer": "No relevant information found.",
"sources": [],
"confidence": 0,
"cached": False
}
# 3. Condense context
context = _optimizer.condense_context(results, max_chars=2000)
# 4. Get answer based on engine type
answer, confidence, model_used = self._get_answer(question, context)
# 5. Extract sources
sources = []
for r in results[:3]:
sources.append({
"source": r['metadata'].get('source', 'Unknown'),
"type": r['metadata'].get('type', 'Unknown'),
"score": round(r['score'], 3)
})
result = {
"answer": answer,
"sources": sources,
"confidence": confidence,
"cached": False,
"model_used": model_used
}
_cache.set(question, result, top_k)
return result
def _get_answer(self, question: str, context: str) -> tuple:
"""Get answer based on engine type"""
# HYBRID: Try Gemini first, fallback to local
if self.engine_type == "hybrid" or self.engine_type == "gemini":
if self.use_gemini:
answer, confidence = self._ask_gemini(question, context)
if "ERROR" not in answer and "Error" not in answer:
return answer, confidence, "gemini-3.5-flash"
# LOCAL or HYBRID fallback
return self._ask_local(question, context)
def _ask_gemini(self, question: str, context: str) -> tuple:
if not self.use_gemini or not self.api_key:
return "ERROR: Gemini not available", 0.0
prompt = f"""
You are a Track Design Engineer. Extract exact values from the context.
QUESTION: {question}
CONTEXT:
{context}
RULES:
1. Extract EXACT numbers with units
2. If formula exists, use it
3. Cite source: "From [source]: [value]"
4. If not found, say "NOT FOUND"
ANSWER:
"""
try:
response = self.client.chat.completions.create(
model=self.model_name,
messages=[
{"role": "system", "content": "You are a Track Design Engineer. Be concise and accurate."},
{"role": "user", "content": prompt}
],
temperature=0.1,
max_tokens=300
)
answer = response.choices[0].message.content
confidence = 0.85
except Exception as e:
answer = f"ERROR: {e}"
confidence = 0.0
return answer, confidence
def _ask_local(self, question: str, context: str) -> tuple:
try:
import ollama
prompt = f"""
Q: {question}
Context: {context}
Answer with specific numbers. Cite sources.
A:"""
response = ollama.chat(
model="qwen2:1.5b",
messages=[{"role": "user", "content": prompt}],
options={
"num_predict": 250,
"temperature": 0.1,
"top_k": 20,
}
)
answer = response['message']['content']
confidence = 0.60
except Exception as e:
answer = f"ERROR: {e}"
confidence = 0.0
return answer, confidence, "qwen2:1.5b"
def ask_with_sources(self, question: str, top_k: int = 3, force_refresh: bool = False) -> str:
result = self.ask(question, top_k, force_refresh)
output = f"""
{'='*60}
π― OPTIMIZED ENGINE
{'='*60}
π {question}
π‘ {result['answer']}
π Sources:
"""
for s in result['sources']:
output += f" - {s['source']} ({s['type']}, {s['score']:.2%})\n"
output += f"""
β‘ Model: {result.get('model_used', 'Unknown')}
{'='*60}
"""
return output
_engine = None
def get_engine(engine_type: str = "hybrid") -> OptimizedEngine:
global _engine
if _engine is None:
_engine = OptimizedEngine(engine_type=engine_type)
return _engine
def ask(question: str, top_k: int = 3) -> str:
engine = get_engine()
return engine.ask_with_sources(question, top_k)
if __name__ == "__main__":
print("=" * 60)
print(" π OPTIMIZED ENGINE")
print("=" * 60)
print("")
print(" Select engine:")
print(" 1: Gemini (Cloud)")
print(" 2: Local (Offline)")
print(" 3: Hybrid (Gemini + Fallback)")
choice = input(" Enter (1, 2, 3): ").strip()
engine_map = {"1": "gemini", "2": "local", "3": "hybrid"}
engine_type = engine_map.get(choice, "hybrid")
engine = OptimizedEngine(engine_type=engine_type)
while True:
q = input("\nπ Ask: ").strip()
if q.lower() in ['exit', 'quit']:
break
if not q:
continue
print("\nβ³ Processing...")
print(engine.ask_with_sources(q))
|