Update app.py
Browse files
app.py
CHANGED
@@ -10,142 +10,149 @@ import os
|
|
10 |
import google.generativeai as genai
|
11 |
from dotenv import load_dotenv
|
12 |
|
13 |
-
|
|
|
14 |
genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
|
15 |
|
16 |
|
17 |
-
def
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
|
|
|
|
|
|
24 |
|
25 |
|
26 |
def get_text_chunks(text):
|
27 |
-
|
28 |
-
|
29 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
30 |
|
31 |
|
32 |
-
def get_vector_store(text_chunks):
|
33 |
-
embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
|
34 |
-
vector_store = FAISS.from_texts(text_chunks, embedding=embeddings)
|
35 |
-
vector_store.save_local("faiss_index")
|
36 |
-
|
37 |
-
|
38 |
-
def get_conversational_chain():
|
39 |
-
prompt_template = """
|
40 |
-
Answer the question as detailed as possible from the provided context, make sure to provide all the details, if the answer is not in
|
41 |
-
provided context just say, "answer is not available in the context", don't provide the wrong answer\n\n
|
42 |
-
Context:\n {context}?\n
|
43 |
-
Question: \n{question}\n
|
44 |
-
|
45 |
-
Answer:
|
46 |
-
"""
|
47 |
-
|
48 |
-
model = ChatGoogleGenerativeAI(model="gemini-pro", temperature=0.3)
|
49 |
-
prompt = PromptTemplate(template=prompt_template, input_variables=["context", "question"])
|
50 |
-
chain = load_qa_chain(model, chain_type="stuff", prompt=prompt)
|
51 |
-
|
52 |
-
return chain
|
53 |
-
|
54 |
-
|
55 |
-
def user_input(user_question):
|
56 |
-
embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
|
57 |
-
new_db = FAISS.load_local("faiss_index", embeddings, allow_dangerous_deserialization=True)
|
58 |
-
docs = new_db.similarity_search(user_question)
|
59 |
-
chain = get_conversational_chain()
|
60 |
-
response = chain({"input_documents": docs, "question": user_question}, return_only_outputs=True)
|
61 |
-
st.write("Reply: ", response["output_text"])
|
62 |
-
|
63 |
-
|
64 |
-
## function to load Gemini Pro model and get responses
|
65 |
-
model = genai.GenerativeModel("gemini-pro-vision")
|
66 |
-
def get_gemini_response(input, image=None):
|
67 |
-
if image is not None and model.is_image_model:
|
68 |
-
response = model.generate_content([input, image])
|
69 |
-
else:
|
70 |
-
response = model.generate_content(input)
|
71 |
-
return response.text
|
72 |
-
|
73 |
-
|
74 |
-
## Initialize our Streamlit app
|
75 |
-
|
76 |
-
st.set_page_config(page_title='Combined Streamlit Application')
|
77 |
-
st.header("Streamlit Application")
|
78 |
-
|
79 |
-
# Define the different applications
|
80 |
-
applications = {
|
81 |
-
"PDF Chat": "pdf_chat",
|
82 |
-
"Image Chat": "image_chat",
|
83 |
-
"Q&A Chat": "qa_chat"
|
84 |
-
}
|
85 |
-
|
86 |
-
# Render the dropdown menu
|
87 |
-
selected_app = st.sidebar.selectbox("Select Application", list(applications.keys()))
|
88 |
-
|
89 |
-
# Function to display PDF Chat application
|
90 |
-
def pdf_chat():
|
91 |
-
st.header("PDF Chat Application")
|
92 |
-
user_question = st.text_input("Ask a Question from the PDF Files")
|
93 |
-
if user_question:
|
94 |
-
user_input(user_question)
|
95 |
-
|
96 |
-
pdf_docs = st.file_uploader("Upload your PDF Files and Click on the Submit & Process Button", accept_multiple_files=True)
|
97 |
-
if st.button("Submit & Process"):
|
98 |
-
with st.spinner("Processing..."):
|
99 |
-
raw_text = get_pdf_text(pdf_docs)
|
100 |
-
text_chunks = get_text_chunks(raw_text)
|
101 |
-
get_vector_store(text_chunks)
|
102 |
-
st.success("Done")
|
103 |
-
|
104 |
-
# Function to display Image Chat application
|
105 |
-
def image_chat():
|
106 |
-
st.header("Image Chat Application")
|
107 |
-
input_text = st.text_input("Input for Gemini Pro:", key="input_gemini")
|
108 |
-
uploaded_file = st.file_uploader("Choose an image...", type="jpg")
|
109 |
-
if uploaded_file is not None:
|
110 |
-
image = Image.open(uploaded_file)
|
111 |
-
st.image(image, caption="Uploaded Image", use_column_width=True)
|
112 |
-
|
113 |
-
submit_gemini = st.button("Ask Gemini Pro")
|
114 |
-
|
115 |
-
if submit_gemini:
|
116 |
-
response_gemini = get_gemini_response(input_text, image)
|
117 |
-
st.subheader("Gemini Pro Response:")
|
118 |
-
st.write(response_gemini)
|
119 |
-
|
120 |
-
# Function to display Q&A Chat application
|
121 |
-
def qa_chat():
|
122 |
-
st.header("Q&A Chat Application")
|
123 |
-
# Initialize session state for chat history if it doesn't exist
|
124 |
-
if 'chat_history' not in st.session_state:
|
125 |
-
st.session_state['chat_history'] = []
|
126 |
-
|
127 |
-
input_qa = st.text_area("Input for Q&A:", key="input_qa")
|
128 |
-
submit_qa = st.button("Ask the question")
|
129 |
-
|
130 |
-
if submit_qa and input_qa:
|
131 |
-
response_qa = get_gemini_response(input_qa)
|
132 |
-
# Add user query and response to session state chat history
|
133 |
-
st.session_state['chat_history'].append(("You", input_qa))
|
134 |
-
st.subheader("Q&A Response:")
|
135 |
-
for chunk in response_qa:
|
136 |
-
st.write(chunk.text)
|
137 |
-
st.session_state['chat_history'].append(("Gemini Pro", chunk.text))
|
138 |
-
|
139 |
-
st.subheader("Q&A Chat History:")
|
140 |
-
for role, text in st.session_state['chat_history']:
|
141 |
-
st.write(f"{role}: {text}")
|
142 |
-
|
143 |
# Map selected application to corresponding function
|
144 |
selected_app_func = {
|
145 |
-
|
146 |
-
|
147 |
-
|
148 |
}
|
149 |
|
150 |
# Run the selected application function
|
151 |
-
selected_app_func[selected_app]()
|
|
|
10 |
import google.generativeai as genai
|
11 |
from dotenv import load_dotenv
|
12 |
|
13 |
+
# Load environment variables
|
14 |
+
load_dotenv()
|
15 |
genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
|
16 |
|
17 |
|
18 |
+
def process_pdfs(pdf_files):
|
19 |
+
"""
|
20 |
+
Extracts text from uploaded PDFs and splits them into chunks.
|
21 |
+
"""
|
22 |
+
text = ""
|
23 |
+
for pdf in pdf_files:
|
24 |
+
reader = PdfReader(pdf)
|
25 |
+
for page in reader.pages:
|
26 |
+
text += page.extract_text()
|
27 |
+
return get_text_chunks(text)
|
28 |
|
29 |
|
30 |
def get_text_chunks(text):
|
31 |
+
"""
|
32 |
+
Splits the text into manageable chunks for processing.
|
33 |
+
"""
|
34 |
+
splitter = RecursiveCharacterTextSplitter(chunk_size=10000, chunk_overlap=1000)
|
35 |
+
return splitter.split_text(text)
|
36 |
+
|
37 |
+
|
38 |
+
def create_vector_store(text_chunks):
|
39 |
+
"""
|
40 |
+
Creates a vector store for efficient text retrieval.
|
41 |
+
"""
|
42 |
+
embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
|
43 |
+
return FAISS.from_texts(text_chunks, embedding=embeddings).save_local("faiss_index")
|
44 |
+
|
45 |
+
|
46 |
+
def build_qa_chain():
|
47 |
+
"""
|
48 |
+
Defines the prompt template and loads the question-answering chain.
|
49 |
+
"""
|
50 |
+
prompt_template = """
|
51 |
+
Answer the question in detail, considering the provided context.
|
52 |
+
If the answer is not available, state "answer unavailable".\n\n
|
53 |
+
Context: {context}?\n
|
54 |
+
Question: {question}\n
|
55 |
+
|
56 |
+
Answer:
|
57 |
+
"""
|
58 |
+
model = ChatGoogleGenerativeAI(model="gemini-pro", temperature=0.3)
|
59 |
+
prompt = PromptTemplate(template=prompt_template, input_variables=["context", "question"])
|
60 |
+
return load_qa_chain(model, chain_type="stuff", prompt=prompt)
|
61 |
+
|
62 |
+
|
63 |
+
def answer_question(question, embeddings, vector_store):
|
64 |
+
"""
|
65 |
+
Searches the vector store and retrieves relevant documents for answering.
|
66 |
+
"""
|
67 |
+
docs = vector_store.similarity_search(question)
|
68 |
+
qa_chain = get_conversational_chain()
|
69 |
+
response = qa_chain({"input_documents": docs, "question": question}, return_only_outputs=True)
|
70 |
+
return response["output_text"]
|
71 |
+
|
72 |
+
|
73 |
+
def get_gemini_response(input_text, image=None):
|
74 |
+
"""
|
75 |
+
Generates a response using the Gemini Pro model, potentially with an image.
|
76 |
+
"""
|
77 |
+
model = genai.GenerativeModel("gemini-pro-vision")
|
78 |
+
if image is not None and model.is_image_model:
|
79 |
+
response = model.generate_content([input_text, image])
|
80 |
+
else:
|
81 |
+
response = model.generate_content(input_text)
|
82 |
+
return response.text
|
83 |
+
|
84 |
+
|
85 |
+
def pdf_chat_app():
|
86 |
+
"""
|
87 |
+
Handles the PDF chat functionality.
|
88 |
+
"""
|
89 |
+
st.header("PDF Chat Application")
|
90 |
+
user_question = st.text_input("Ask a Question from the PDF Files")
|
91 |
+
uploaded_files = st.file_uploader("Upload PDF Files", accept_multiple_files=True)
|
92 |
+
|
93 |
+
if uploaded_files:
|
94 |
+
with st.spinner("Processing..."):
|
95 |
+
text_chunks = process_pdfs(uploaded_files)
|
96 |
+
create_vector_store(text_chunks)
|
97 |
+
embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
|
98 |
+
vector_store = FAISS.load_local("faiss_index", embeddings)
|
99 |
+
st.success("Done!")
|
100 |
+
|
101 |
+
if user_question and uploaded_files:
|
102 |
+
answer = answer_question(user_question, embeddings, vector_store)
|
103 |
+
st.write("Reply:", answer)
|
104 |
+
|
105 |
+
def image_chat_app():
|
106 |
+
"""
|
107 |
+
Handles the image chat functionality.
|
108 |
+
"""
|
109 |
+
st.header("Image Chat Application")
|
110 |
+
input_text = st.text_input("Input for Gemini Pro:")
|
111 |
+
uploaded_image = st.file_uploader("Choose an image...", type="jpg")
|
112 |
+
if uploaded_image is not None:
|
113 |
+
image = Image.open(uploaded_image)
|
114 |
+
st.image(image, caption="Uploaded Image", use_column_width=True)
|
115 |
+
|
116 |
+
submit_gemini = st.button("Ask Gemini Pro")
|
117 |
+
|
118 |
+
if submit_gemini:
|
119 |
+
response_gemini = get_gemini_response(input_text, image)
|
120 |
+
st.subheader("Gemini Pro Response:")
|
121 |
+
st.write(response_gemini)
|
122 |
+
|
123 |
+
|
124 |
+
def qa_chat_app():
|
125 |
+
"""
|
126 |
+
Handles the Q&A chat functionality.
|
127 |
+
"""
|
128 |
+
st.header("Q&A Chat Application")
|
129 |
+
# Initialize session state for chat history if it doesn't exist
|
130 |
+
if 'chat_history' not in st.session_state:
|
131 |
+
st.session_state['chat_history'] = []
|
132 |
+
|
133 |
+
input_qa = st.text_area("Input for Q&A:", key="input_qa")
|
134 |
+
submit_qa = st.button("Ask the question")
|
135 |
+
|
136 |
+
if submit_qa and input_qa:
|
137 |
+
response_qa = get_gemini_response(input_qa)
|
138 |
+
# Add user query and response to session state chat history
|
139 |
+
st.session_state['chat_history'].append(("You", input_qa))
|
140 |
+
st.subheader("Q&A Response:")
|
141 |
+
for chunk in response_qa:
|
142 |
+
st.write(chunk.text)
|
143 |
+
st.session_state['chat_history'].append(("Gemini Pro", chunk.text))
|
144 |
+
|
145 |
+
st.subheader("Q&A Chat History:")
|
146 |
+
for role, text in st.session_state['chat_history']:
|
147 |
+
st.write(f"{role}: {text}")
|
148 |
|
149 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
150 |
# Map selected application to corresponding function
|
151 |
selected_app_func = {
|
152 |
+
"PDF Chat": pdf_chat,
|
153 |
+
"Image Chat": image_chat,
|
154 |
+
"Q&A Chat": qa_chat
|
155 |
}
|
156 |
|
157 |
# Run the selected application function
|
158 |
+
selected_app_func[selected_app]()
|