Skip to content

Commit 797b396

Browse files
sanjay-kvsanjay-kv
authored
Merge pull request #1314 from J-B-Mugundh/feature-1
Added Beyond GPS Navigational Chatbot
2 parents 2c2342c + 92d873c commit 797b396

File tree

5 files changed

+361
-0
lines changed

5 files changed

+361
-0
lines changed

‎Algorithms and Deep Learning Models/Beyond-GPS-Navi-Bot/README.md

Lines changed: 36 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,36 @@
1+
# NaviBot-Voice-Assistant
2+
3+
To get started with NaviBot-Voice-Assistant, follow these steps:
4+
5+
1. Navigate to the project directory:
6+
7+
```bash
8+
cd NaviBot-Voice-Assistant
9+
```
10+
11+
3. Activating venv (optional)
12+
13+
```bash
14+
conda create -n venv python=3.10+
15+
conda activate venv
16+
```
17+
18+
4. Install dependencies:
19+
20+
```python
21+
pip install -r requirements.txt
22+
```
23+
24+
5. Configure environment variables
25+
```
26+
Rename `.env-sample` to `.env` file
27+
Replace the API your Google API Key,
28+
```
29+
Kindly follow refer to this site for getting [your own key](https://ai.google.dev/tutorials/setup)
30+
<br/>
31+
32+
6. Run the chatbot:
33+
34+
```bash
35+
streamlit run app.py
36+
```

‎Algorithms and Deep Learning Models/Beyond-GPS-Navi-Bot/app.py

Lines changed: 133 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,133 @@
1+
import streamlit as st
2+
import pyttsx3
3+
import speech_recognition as sr
4+
from PyPDF2 import PdfReader
5+
from langchain.text_splitter import RecursiveCharacterTextSplitter
6+
import os
7+
from langchain_google_genai import GoogleGenerativeAIEmbeddings
8+
import google.generativeai as genai
9+
from langchain_community.vectorstores import FAISS
10+
from langchain_google_genai import ChatGoogleGenerativeAI
11+
from langchain.chains.question_answering import load_qa_chain
12+
from langchain.prompts import PromptTemplate
13+
from dotenv import load_dotenv
14+
15+
load_dotenv()
16+
os.getenv("GOOGLE_API_KEY")
17+
genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
18+
19+
# Initialize pyttsx3 for voice output
20+
engine = pyttsx3.init()
21+
22+
# Function to speak the text
23+
def speak(text):
24+
engine.say(text)
25+
engine.runAndWait()
26+
27+
# Function to listen to voice input
28+
def listen():
29+
r = sr.Recognizer()
30+
with sr.Microphone() as source:
31+
st.write("Listening...")
32+
r.adjust_for_ambient_noise(source)
33+
audio = r.listen(source)
34+
35+
try:
36+
user_input = r.recognize_google(audio)
37+
st.write(f"You said: {user_input}")
38+
return user_input
39+
except sr.UnknownValueError:
40+
st.write("Sorry, I could not understand what you said.")
41+
return None
42+
except sr.RequestError as e:
43+
st.write(f"Could not request results from Google Speech Recognition service; {e}")
44+
return None
45+
46+
47+
def get_pdf_text(pdf_docs):
48+
text=""
49+
for pdf in pdf_docs:
50+
pdf_reader= PdfReader(pdf)
51+
for page in pdf_reader.pages:
52+
text+= page.extract_text()
53+
return text
54+
55+
56+
def get_text_chunks(text):
57+
text_splitter = RecursiveCharacterTextSplitter(chunk_size=10000, chunk_overlap=1000)
58+
chunks = text_splitter.split_text(text)
59+
return chunks
60+
61+
62+
def get_vector_store(text_chunks):
63+
embeddings = GoogleGenerativeAIEmbeddings(model = "models/embedding-001")
64+
vector_store = FAISS.from_texts(text_chunks, embedding=embeddings)
65+
vector_store.save_local("faiss_index")
66+
67+
68+
def get_conversational_chain():
69+
70+
prompt_template = """
71+
Answer the question as detailed as possible from the provided context, make sure to provide all the details, if the answer is not in
72+
provided context just say, "answer is not available in the context", don't provide the wrong answer\n\n
73+
Context:\n {context}?\n
74+
Question: \n{question}\n
75+
76+
Answer:
77+
"""
78+
79+
model = ChatGoogleGenerativeAI(model="gemini-pro",
80+
temperature=0.3)
81+
82+
prompt = PromptTemplate(template = prompt_template, input_variables = ["context", "question"])
83+
chain = load_qa_chain(model, chain_type="stuff", prompt=prompt)
84+
85+
return chain
86+
87+
88+
def user_input(user_question):
89+
embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
90+
91+
# Load the local FAISS index with dangerous deserialization allowed
92+
new_db = FAISS.load_local("faiss_index", embeddings, allow_dangerous_deserialization=True)
93+
docs = new_db.similarity_search(user_question)
94+
95+
chain = get_conversational_chain()
96+
97+
response = chain(
98+
{"input_documents": docs, "question": user_question},
99+
return_only_outputs=True
100+
)
101+
102+
speak(response["output_text"]) # Speak the response
103+
st.write("Reply: ", response["output_text"])
104+
105+
106+
def main():
107+
st.set_page_config("Beyond GPS Navigation")
108+
st.header("Beyond GPS Navigator for Blind")
109+
110+
user_question = st.text_input("Ask your query")
111+
voice_input_button = st.button("Voice Input")
112+
113+
if voice_input_button:
114+
user_question = listen() # Listen to voice input
115+
if user_question:
116+
user_input(user_question)
117+
118+
if user_question:
119+
user_input(user_question)
120+
121+
with st.sidebar:
122+
st.title("Menu:")
123+
pdf_docs = st.file_uploader("Upload your route data and Click on the Submit & Process Button", accept_multiple_files=True)
124+
if st.button("Submit & Process"):
125+
with st.spinner("Processing..."):
126+
raw_text = get_pdf_text(pdf_docs)
127+
text_chunks = get_text_chunks(raw_text)
128+
get_vector_store(text_chunks)
129+
st.success("Done")
130+
131+
132+
if __name__ == "__main__":
133+
main()

‎Algorithms and Deep Learning Models/Beyond-GPS-Navi-Bot/main.py

Lines changed: 88 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,88 @@
1+
import os
2+
import streamlit as st
3+
from dotenv import load_dotenv
4+
import google.generativeai as gen_ai
5+
import speech_recognition as sr
6+
from gtts import gTTS
7+
from tempfile import TemporaryFile
8+
9+
# Load environment variables
10+
load_dotenv()
11+
12+
# Configure Streamlit page settings
13+
st.set_page_config(
14+
page_title="Chat with Gemini-Pro!",
15+
page_icon=":brain:", # Favicon emoji
16+
layout="centered", # Page layout option
17+
)
18+
19+
GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY")
20+
21+
# Set up Google Gemini-Pro AI model
22+
gen_ai.configure(api_key=GOOGLE_API_KEY)
23+
model = gen_ai.GenerativeModel('gemini-pro')
24+
25+
# Function to translate roles between Gemini-Pro and Streamlit terminology
26+
def translate_role_for_streamlit(user_role):
27+
if user_role == "model":
28+
return "assistant"
29+
else:
30+
return user_role
31+
32+
# Function to recognize speech input
33+
def recognize_speech():
34+
r = sr.Recognizer()
35+
with sr.Microphone() as source:
36+
st.write("Speak now...")
37+
audio = r.listen(source)
38+
39+
try:
40+
user_prompt = r.recognize_google(audio)
41+
st.write("You said:", user_prompt)
42+
return user_prompt
43+
except sr.UnknownValueError:
44+
st.write("Sorry, I could not understand your audio.")
45+
return ""
46+
except sr.RequestError as e:
47+
st.write("Could not request results from Google Speech Recognition service; {0}".format(e))
48+
return ""
49+
50+
# Function to output voice
51+
def speak(text):
52+
tts = gTTS(text=text, lang='en')
53+
with TemporaryFile(suffix=".wav", delete=False) as f:
54+
tts.write_to_fp(f)
55+
filename = f.name
56+
st.audio(filename, format='audio/wav')
57+
58+
59+
# Initialize chat session in Streamlit if not already present
60+
if "chat_session" not in st.session_state:
61+
st.session_state.chat_session = model.start_chat(history=[])
62+
63+
# Display the chatbot's title on the page
64+
st.title("🤖 Gemini Pro - ChatBot")
65+
66+
# Display the chat history
67+
for message in st.session_state.chat_session.history:
68+
with st.chat_message(translate_role_for_streamlit(message.role)):
69+
st.markdown(message.parts[0].text)
70+
71+
# Input field for user's message
72+
voice_input = st.checkbox("Voice Input")
73+
if voice_input:
74+
user_prompt = recognize_speech()
75+
else:
76+
user_prompt = st.text_input("Ask Gemini-Pro...")
77+
78+
if user_prompt:
79+
# Add user's message to chat and display it
80+
st.chat_message("user").markdown(user_prompt)
81+
82+
# Send user's message to Gemini-Pro and get the response
83+
gemini_response = st.session_state.chat_session.send_message(user_prompt)
84+
85+
# Display Gemini-Pro's response
86+
with st.chat_message("assistant"):
87+
st.markdown(gemini_response.text)
88+
speak(gemini_response.text)

‎Algorithms and Deep Learning Models/Beyond-GPS-Navi-Bot/map.py

Lines changed: 99 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,99 @@
1+
import os
2+
import streamlit as st
3+
from dotenv import load_dotenv
4+
import google.generativeai as gen_ai
5+
import speech_recognition as sr
6+
from gtts import gTTS
7+
from tempfile import TemporaryFile
8+
import webbrowser
9+
10+
# Load environment variables
11+
load_dotenv()
12+
13+
# Configure Streamlit page settings
14+
st.set_page_config(
15+
page_title="Beyond GPS Navigator!",
16+
page_icon=":brain:", # Favicon emoji
17+
layout="centered", # Page layout option
18+
)
19+
20+
GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY")
21+
22+
# Set up Google Gemini-Pro AI model
23+
gen_ai.configure(api_key=GOOGLE_API_KEY)
24+
model = gen_ai.GenerativeModel('gemini-pro')
25+
26+
# Function to translate roles between Gemini-Pro and Streamlit terminology
27+
def translate_role_for_streamlit(user_role):
28+
if user_role == "model":
29+
return "assistant"
30+
else:
31+
return user_role
32+
33+
# Function to recognize speech input
34+
def recognize_speech():
35+
r = sr.Recognizer()
36+
with sr.Microphone() as source:
37+
st.write("Speak now...")
38+
audio = r.listen(source)
39+
40+
try:
41+
user_prompt = r.recognize_google(audio)
42+
st.write("You said:", user_prompt)
43+
return user_prompt
44+
except sr.UnknownValueError:
45+
st.write("Sorry, I could not understand your audio.")
46+
return ""
47+
except sr.RequestError as e:
48+
st.write("Could not request results from Google Speech Recognition service; {0}".format(e))
49+
return ""
50+
51+
# Function to output voice
52+
def speak(text):
53+
tts = gTTS(text=text, lang='en')
54+
with TemporaryFile(suffix=".wav", delete=False) as f:
55+
tts.write_to_fp(f)
56+
filename = f.name
57+
st.audio(filename, format='audio/wav')
58+
59+
# Get user's current location
60+
current_location = st.text_input("What is your current location?")
61+
62+
# Ask the user for their destination
63+
destination = recognize_speech()
64+
65+
# Initialize chat session in Streamlit if not already present
66+
if "chat_session" not in st.session_state:
67+
st.session_state.chat_session = model.start_chat(history=[])
68+
69+
# Display the chatbot's title on the page
70+
st.title("🤖 Gemini Pro - ChatBot")
71+
72+
# Display the chat history
73+
for message in st.session_state.chat_session.history:
74+
with st.chat_message(translate_role_for_streamlit(message.role)):
75+
st.markdown(message.parts[0].text)
76+
77+
# Input field for user's message
78+
voice_input = st.checkbox("Voice Input")
79+
if voice_input:
80+
user_prompt = recognize_speech()
81+
else:
82+
user_prompt = st.text_input("Ask Gemini-Pro...")
83+
84+
if user_prompt:
85+
# Add user's message to chat and display it
86+
st.chat_message("user").markdown(user_prompt)
87+
88+
# Send user's message to Gemini-Pro and get the response
89+
gemini_response = st.session_state.chat_session.send_message(user_prompt)
90+
91+
# Display Gemini-Pro's response
92+
with st.chat_message("assistant"):
93+
st.markdown(gemini_response.text)
94+
speak(gemini_response.text)
95+
96+
# If the response contains directions, open them in Google Maps
97+
if "directions" in gemini_response.text:
98+
directions_url = "https://www.google.com/maps/dir/?api=1&origin=" + current_location + "&destination=" + destination
99+
webbrowser.open(directions_url)

‎Algorithms and Deep Learning Models/Beyond-GPS-Navi-Bot/requirements.txt

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,5 @@
1+
streamlit
2+
dotenv
3+
google-generativeai
4+
speech_recognition
5+
gtts

0 commit comments

Comments
 (0)