Update classification_chain.py
Browse files- classification_chain.py +10 -2
classification_chain.py
CHANGED
@@ -2,8 +2,6 @@
|
|
2 |
import os
|
3 |
from langchain.chains import LLMChain
|
4 |
from langchain_groq import ChatGroq
|
5 |
-
|
6 |
-
# We'll import the classification_prompt from prompts.py
|
7 |
from prompts import classification_prompt
|
8 |
|
9 |
def get_classification_chain() -> LLMChain:
|
@@ -22,3 +20,13 @@ def get_classification_chain() -> LLMChain:
|
|
22 |
prompt=classification_prompt
|
23 |
)
|
24 |
return classification_chain
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
import os
|
3 |
from langchain.chains import LLMChain
|
4 |
from langchain_groq import ChatGroq
|
|
|
|
|
5 |
from prompts import classification_prompt
|
6 |
|
7 |
def get_classification_chain() -> LLMChain:
|
|
|
20 |
prompt=classification_prompt
|
21 |
)
|
22 |
return classification_chain
|
23 |
+
|
24 |
+
def classify_with_history(query: str, chat_history: list) -> str:
|
25 |
+
"""
|
26 |
+
Classifies a user query based on the context of previous conversation (chat_history).
|
27 |
+
"""
|
28 |
+
# Add the history into the query context if needed (depending on the type of model)
|
29 |
+
context = "\n".join([f"User: {msg['content']}" for msg in chat_history]) + "\nUser: " + query
|
30 |
+
# Update the prompt with both the context and the query
|
31 |
+
classification_result = get_classification_chain().run({"query": context})
|
32 |
+
return classification_result
|