Spaces:
Sleeping
Sleeping
Commit
·
9b3a839
1
Parent(s):
fce8396
Update Chat_with_pdf_OpenAI
Browse files- Chat_with_pdf_OpenAI +50 -49
Chat_with_pdf_OpenAI
CHANGED
@@ -10,54 +10,55 @@ import os
|
|
10 |
import openai
|
11 |
from streamlit_chat import message
|
12 |
|
13 |
-
|
14 |
-
|
15 |
-
|
16 |
-
|
17 |
-
# upload file
|
18 |
-
pdf = st.file_uploader("Upload your PDF", type="pdf")
|
19 |
-
|
20 |
-
# extract the text
|
21 |
-
if pdf is not None:
|
22 |
-
pdf_reader = PdfReader(pdf)
|
23 |
-
text = ""
|
24 |
-
for page in pdf_reader.pages:
|
25 |
-
text += page.extract_text()
|
26 |
-
|
27 |
-
# split into chunks
|
28 |
-
text_splitter = CharacterTextSplitter(
|
29 |
-
separator="\n",
|
30 |
-
chunk_size=1000,
|
31 |
-
chunk_overlap=200,
|
32 |
-
length_function=len
|
33 |
-
)
|
34 |
-
chunks = text_splitter.split_text(text)
|
35 |
-
|
36 |
-
# create embeddings
|
37 |
-
embeddings = OpenAIEmbeddings()
|
38 |
-
knowledge_base = FAISS.from_texts(chunks, embeddings)
|
39 |
-
|
40 |
-
if 'generated' not in st.session_state:
|
41 |
-
st.session_state['generated'] = []
|
42 |
-
if 'past' not in st.session_state:
|
43 |
-
st.session_state['past'] = []
|
44 |
-
|
45 |
-
# show user input
|
46 |
-
user_question = st.text_input("Ask a question about your PDF:")
|
47 |
-
if user_question:
|
48 |
-
docs = knowledge_base.similarity_search(user_question)
|
49 |
-
|
50 |
-
llm = OpenAI()
|
51 |
-
chain = load_qa_chain(llm, chain_type="stuff")
|
52 |
-
with get_openai_callback() as cb:
|
53 |
-
response = chain.run(input_documents=docs, question=user_question)
|
54 |
-
print(cb)
|
55 |
|
56 |
-
#
|
57 |
-
st.
|
58 |
-
st.session_state.generated.append(response)
|
59 |
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
import openai
|
11 |
from streamlit_chat import message
|
12 |
|
13 |
+
def openai_pdf():
|
14 |
+
OPENAI_API_KEY = st.text_input("Input your OpenAI API key", "")
|
15 |
+
os.environ["OPENAI_API_KEY"] = OPENAI_API_KEY
|
16 |
+
# st.header("Ask your PDF 💬")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
17 |
|
18 |
+
# upload file
|
19 |
+
pdf = st.file_uploader("Upload your PDF", type="pdf")
|
|
|
20 |
|
21 |
+
# extract the text
|
22 |
+
if pdf is not None:
|
23 |
+
pdf_reader = PdfReader(pdf)
|
24 |
+
text = ""
|
25 |
+
for page in pdf_reader.pages:
|
26 |
+
text += page.extract_text()
|
27 |
+
|
28 |
+
# split into chunks
|
29 |
+
text_splitter = CharacterTextSplitter(
|
30 |
+
separator="\n",
|
31 |
+
chunk_size=1000,
|
32 |
+
chunk_overlap=200,
|
33 |
+
length_function=len
|
34 |
+
)
|
35 |
+
chunks = text_splitter.split_text(text)
|
36 |
+
|
37 |
+
# create embeddings
|
38 |
+
embeddings = OpenAIEmbeddings()
|
39 |
+
knowledge_base = FAISS.from_texts(chunks, embeddings)
|
40 |
+
|
41 |
+
if 'generated' not in st.session_state:
|
42 |
+
st.session_state['generated'] = []
|
43 |
+
if 'past' not in st.session_state:
|
44 |
+
st.session_state['past'] = []
|
45 |
+
|
46 |
+
# show user input
|
47 |
+
user_question = st.text_input("Ask a question about your PDF:")
|
48 |
+
if user_question:
|
49 |
+
docs = knowledge_base.similarity_search(user_question)
|
50 |
+
|
51 |
+
llm = OpenAI()
|
52 |
+
chain = load_qa_chain(llm, chain_type="stuff")
|
53 |
+
with get_openai_callback() as cb:
|
54 |
+
response = chain.run(input_documents=docs, question=user_question)
|
55 |
+
print(cb)
|
56 |
+
|
57 |
+
#st.write(response)
|
58 |
+
st.session_state.past.append(user_question)
|
59 |
+
st.session_state.generated.append(response)
|
60 |
+
|
61 |
+
if st.session_state['generated']:
|
62 |
+
for i in range(len(st.session_state['generated'])-1, -1, -1):
|
63 |
+
message(st.session_state["generated"][i], key=str(i))
|
64 |
+
message(st.session_state['past'][i], is_user=True, key=str(i) + '_user')
|