Spaces:
Sleeping
Sleeping
Mohit Raj commited on
Sync from GitHub via hub-sync
Browse files- config.py +2 -1
- rag_chain.py +10 -12
- requirements.txt +1 -0
config.py
CHANGED
|
@@ -10,10 +10,11 @@ load_dotenv()
|
|
| 10 |
|
| 11 |
# βββ API Keys ββββββββββββββββββββββββββββββββββββββββββββββββ
|
| 12 |
GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY", "").strip()
|
|
|
|
| 13 |
|
| 14 |
# βββ Model Configuration ββββββββββββββββββββββββββββββββββββ
|
| 15 |
EMBEDDING_MODEL = os.getenv("EMBEDDING_MODEL", "all-MiniLM-L6-v2")
|
| 16 |
-
LLM_MODEL = os.getenv("LLM_MODEL", "
|
| 17 |
|
| 18 |
# βββ Paths βββββββββββββββββββββββββββββββββββββββββββββββββββ
|
| 19 |
BASE_DIR = os.path.dirname(os.path.abspath(__file__))
|
|
|
|
| 10 |
|
| 11 |
# βββ API Keys ββββββββββββββββββββββββββββββββββββββββββββββββ
|
| 12 |
GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY", "").strip()
|
| 13 |
+
GROQ_API_KEY = os.getenv("GROQ_API_KEY", "").strip()
|
| 14 |
|
| 15 |
# βββ Model Configuration ββββββββββββββββββββββββββββββββββββ
|
| 16 |
EMBEDDING_MODEL = os.getenv("EMBEDDING_MODEL", "all-MiniLM-L6-v2")
|
| 17 |
+
LLM_MODEL = os.getenv("LLM_MODEL", "llama-3.3-70b-versatile")
|
| 18 |
|
| 19 |
# βββ Paths βββββββββββββββββββββββββββββββββββββββββββββββββββ
|
| 20 |
BASE_DIR = os.path.dirname(os.path.abspath(__file__))
|
rag_chain.py
CHANGED
|
@@ -92,23 +92,21 @@ def build_rag_chain(vector_store: FAISS) -> RetrievalQA:
|
|
| 92 |
search_kwargs={"k": config.TOP_K_RESULTS},
|
| 93 |
)
|
| 94 |
|
| 95 |
-
#
|
|
|
|
|
|
|
|
|
|
| 96 |
import re
|
| 97 |
-
raw_key = str(config.
|
| 98 |
api_key = re.sub(r'[^a-zA-Z0-9_\-]', '', raw_key)
|
| 99 |
|
| 100 |
-
print(f"
|
| 101 |
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
| 105 |
-
google_api_key=api_key,
|
| 106 |
temperature=config.TEMPERATURE,
|
| 107 |
-
|
| 108 |
-
convert_system_message_to_human=True,
|
| 109 |
-
# Use v1beta for Gemini 2.0 models
|
| 110 |
-
api_version="v1beta",
|
| 111 |
-
transport="rest"
|
| 112 |
)
|
| 113 |
|
| 114 |
# Build the RAG chain
|
|
|
|
| 92 |
search_kwargs={"k": config.TOP_K_RESULTS},
|
| 93 |
)
|
| 94 |
|
| 95 |
+
# Initialize Groq LLM (Super Fast)
|
| 96 |
+
from langchain_groq import ChatGroq
|
| 97 |
+
|
| 98 |
+
# Aggressive cleaning of the Groq key
|
| 99 |
import re
|
| 100 |
+
raw_key = str(config.GROQ_API_KEY)
|
| 101 |
api_key = re.sub(r'[^a-zA-Z0-9_\-]', '', raw_key)
|
| 102 |
|
| 103 |
+
print(f" β‘ Groq Engine active. Key length: {len(api_key)}")
|
| 104 |
|
| 105 |
+
llm = ChatGroq(
|
| 106 |
+
model_name=config.LLM_MODEL,
|
| 107 |
+
groq_api_key=api_key,
|
|
|
|
| 108 |
temperature=config.TEMPERATURE,
|
| 109 |
+
max_tokens=config.MAX_OUTPUT_TOKENS,
|
|
|
|
|
|
|
|
|
|
|
|
|
| 110 |
)
|
| 111 |
|
| 112 |
# Build the RAG chain
|
requirements.txt
CHANGED
|
@@ -3,6 +3,7 @@ langchain-community==0.2.12
|
|
| 3 |
langchain-core==0.2.38
|
| 4 |
langchain-google-genai==1.0.10
|
| 5 |
langchain-huggingface==0.0.3
|
|
|
|
| 6 |
faiss-cpu>=1.8.0
|
| 7 |
sentence-transformers==3.0.1
|
| 8 |
google-generativeai==0.7.2
|
|
|
|
| 3 |
langchain-core==0.2.38
|
| 4 |
langchain-google-genai==1.0.10
|
| 5 |
langchain-huggingface==0.0.3
|
| 6 |
+
langchain-groq
|
| 7 |
faiss-cpu>=1.8.0
|
| 8 |
sentence-transformers==3.0.1
|
| 9 |
google-generativeai==0.7.2
|