Spaces:
Paused
Paused
File size: 2,558 Bytes
910140e 6272795 910140e 6272795 910140e 2b1d596 910140e 6272795 910140e 6272795 910140e 6272795 910140e 6272795 910140e 6272795 6bc2eb0 6272795 6bc2eb0 6272795 6bc2eb0 6272795 | 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 | import os
import json
from groq import Groq
import google.generativeai as genai
# ==========================
# Setup API Clients
# ==========================
groq_client = Groq(api_key=os.getenv("GROQ_API_KEY"))
genai.configure(api_key=os.getenv("GEMINI_API_KEY"))
gemini_model = genai.GenerativeModel("gemini-2.5-flash")
# ==========================
# Llama (Groq)
# ==========================
def call_llama(messages, temperature=0.7):
"""
messages = [
{"role": "system", "content": "..."},
{"role": "user", "content": "..."}
]
"""
response = groq_client.chat.completions.create(
model="llama-3.1-8b-instant",
messages=messages,
temperature=temperature
)
return response.choices[0].message.content
# ==========================
# Gemini
# ==========================
def call_gemini(messages, temperature=0.7):
"""
Gemini expects combined prompt.
We convert messages into a single string.
"""
formatted_prompt = ""
for m in messages:
formatted_prompt += f"{m['role'].upper()}: {m['content']}\n"
response = gemini_model.generate_content(
formatted_prompt,
generation_config={"temperature": temperature}
)
return response.text
# ==========================
# AI Classifier
# ==========================
def classify_prompt(prompt):
system_prompt = """
You are an AI intent classifier.
Return ONLY valid JSON in this format:
{
"intent": "chat | reasoning | live_data | greeting",
"needs_search": true or false,
"complexity": "low | medium | high"
}
Rules:
- Logical, math, proof, step explanation -> reasoning
- Current events, weather, news, price, today -> live_data
- Hello/hi/how are you -> greeting
- Everything else -> chat
"""
messages = [
{"role": "system", "content": system_prompt},
{"role": "user", "content": prompt}
]
raw = call_llama(messages, temperature=0)
try:
return json.loads(raw)
except:
return {
"intent": "chat",
"needs_search": False,
"complexity": "low"
}
# ==========================
# LLM Judge (for dual answers)
# ==========================
def judge_answers(answer1, answer2):
judge_prompt = f"""
Choose which answer is better.
Reply ONLY with:
1
or
2
Answer 1:
{answer1}
Answer 2:
{answer2}
"""
result = call_llama(
[{"role": "user", "content": judge_prompt}],
temperature=0
).strip()
if result == "2":
return 2
return 1
|