Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -11,7 +11,6 @@ from langchain.indexes import VectorstoreIndexCreator
|
|
11 |
from langchain import OpenAI, VectorDBQA
|
12 |
|
13 |
import os
|
14 |
-
openai_api_key = os.environ["OPENAI_API_KEY"]
|
15 |
|
16 |
|
17 |
def pdf_to_text(pdf_file, query):
|
@@ -40,15 +39,13 @@ def pdf_to_text(pdf_file, query):
|
|
40 |
llm = HuggingFaceHub(repo_id="google/flan-t5-xl", model_kwargs={"temperature":0, "max_length":512})
|
41 |
loaders = UnstructuredPDFLoader(pdf_file)
|
42 |
|
43 |
-
index =
|
44 |
-
embedding=HuggingFaceEmbeddings(),
|
45 |
-
text_splitter= CharacterTextSplitter(chunk_size=1000, chunk_overlap=0).from_loaders(loaders))
|
46 |
#inference
|
47 |
qa = VectorDBQA.from_chain_type(llm=llm, chain_type="stuff", vectorstore=vectorstore)
|
48 |
from langchain.chains import RetrievalQA
|
49 |
chain = RetrievalQA.from_chain_type(llm=llm,
|
50 |
chain_type="stuff",
|
51 |
-
retriever=index
|
52 |
input_key="question")
|
53 |
return chain.run(query)
|
54 |
|
|
|
11 |
from langchain import OpenAI, VectorDBQA
|
12 |
|
13 |
import os
|
|
|
14 |
|
15 |
|
16 |
def pdf_to_text(pdf_file, query):
|
|
|
39 |
llm = HuggingFaceHub(repo_id="google/flan-t5-xl", model_kwargs={"temperature":0, "max_length":512})
|
40 |
loaders = UnstructuredPDFLoader(pdf_file)
|
41 |
|
42 |
+
index = vectorstore.as_retriever()
|
|
|
|
|
43 |
#inference
|
44 |
qa = VectorDBQA.from_chain_type(llm=llm, chain_type="stuff", vectorstore=vectorstore)
|
45 |
from langchain.chains import RetrievalQA
|
46 |
chain = RetrievalQA.from_chain_type(llm=llm,
|
47 |
chain_type="stuff",
|
48 |
+
retriever=index,
|
49 |
input_key="question")
|
50 |
return chain.run(query)
|
51 |
|