langchain/templates/rag-matching-engine/rag_matching_engine/chain.py
2024-01-03 13:28:05 -08:00

79 lines
2.4 KiB
Python

import os
from langchain_community.embeddings import VertexAIEmbeddings
from langchain_community.llms import VertexAI
from langchain_community.vectorstores import MatchingEngine
from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import PromptTemplate
from langchain_core.pydantic_v1 import BaseModel
from langchain_core.runnables import RunnableParallel, RunnablePassthrough
# you need to preate the index first, for example, as described here:
# https://github.com/GoogleCloudPlatform/generative-ai/blob/main/language/use-cases/document-qa/question_answering_documents_langchain_matching_engine.ipynb
expected_variables = [
"project_id",
"me_region",
"gcs_bucket",
"me_index_id",
"me_endpoint_id",
]
variables = []
for variable_name in expected_variables:
variable = os.environ.get(variable_name.upper())
if not variable:
raise Exception(f"Missing `{variable_name}` environment variable.")
variables.append(variable)
project_id, me_region, gcs_bucket, me_index_id, me_endpoint_id = variables
vectorstore = MatchingEngine.from_components(
project_id=project_id,
region=me_region,
gcs_bucket_name=gcs_bucket,
embedding=VertexAIEmbeddings(),
index_id=me_index_id,
endpoint_id=me_endpoint_id,
)
model = VertexAI()
template = (
"SYSTEM: You are an intelligent assistant helping the users with their questions"
"on research papers.\n\n"
"Question: {question}\n\n"
"Strictly Use ONLY the following pieces of context to answer the question at the "
"end. Think step-by-step and then answer.\n\n"
"Do not try to make up an answer:\n"
"- If the answer to the question cannot be determined from the context alone, "
'say \n"I cannot determine the answer to that."\n'
'- If the context is empty, just say "I do not know the answer to that."\n\n'
"=============\n{context}\n=============\n\n"
"Question: {question}\nHelpful Answer: "
)
prompt = PromptTemplate.from_template(template)
retriever = vectorstore.as_retriever(
search_type="similarity",
search_kwargs={
"k": 10,
"search_distance": 0.6,
},
)
chain = (
RunnableParallel({"context": retriever, "question": RunnablePassthrough()})
| prompt
| model
| StrOutputParser()
)
# Add typing for input
class Question(BaseModel):
__root__: str
chain = chain.with_types(input_type=Question)