ChatSentiment

ChatSentiment is a Streamlit web application for analyzing the sentiment of chat messages. It uses the Hugging Face Llama model for generating responses and TextBlob for sentiment analysis.

Try the streamlit app: https://chatsentiment.streamlit.app/

Features

Sentiment Analysis: Analyzes the sentiment of chat messages and generates a report showing the distribution of positive, negative, and neutral sentiments.
Keyword Frequency Analysis: Extracts keywords from chat messages and displays the top 10 most frequent keywords.
Word Cloud Visualization: Generates a word cloud visualization based on the extracted keywords.
Clear Conversation: Allows users to clear the conversation history.

Installation

Clone the repository:

git clone https://github.com/VenkateshSoni/ChatSentiment.git

Navigate to the project directory:
```
cd ChatSentiment
```
Install the required dependencies:
```
pip install -r requirements.txt
```
Run following command in terminal to start the app.
```
streamlit run app.py
```

Usage

Running with Streamlit

Run the following command to start the Streamlit app:

streamlit run app.py

Once the app is running, you can input your questions in the chat interface and view the sentiment analysis report, keyword frequency analysis, and word cloud visualization in real-time.

Running with Docker

Alternatively, you can run the application using Docker. First, pull the Docker image from the repository:

docker pull venkateshsoni/chatsentiment:1.1

Then, run the Docker container:

docker run -p 80:80 venkateshsoni/chatsentiment:1.1

This will start the Streamlit app inside a Docker container, and you can access it by navigating to http://localhost in your web browser.

Documentation

Functions

init_page(): Initializes the Streamlit page configuration, setting the page title to "ChatSentiment", adding a header "ChatSentiment", and setting the title of the sidebar to "Options".
init_messages(): Initializes the messages session state by adding a clear conversation button to the sidebar. If the button is clicked or the session state does not contain messages, it adds a default system message to the messages session state.
generate_answer(question): Generates an answer for the given question using the Hugging Face Llama model.
analyze_sentiment(text): Analyzes the sentiment of the given text using TextBlob.
extract_keywords(messages): Extracts keywords from the given messages.
render_sentiment_report(messages): Renders a sentiment analysis report based on the given messages, including sentiment analysis, keyword frequency analysis, and word cloud visualization.
main(): Main function to run the Streamlit app, which initializes the page, messages, gets user input, generates responses, and displays the sentiment report and chat messages.

Contributing

Contributions are welcome! If you have any suggestions, enhancements, or bug fixes, please feel free to open an issue or create a pull request.

License

This project is licensed under the MIT License.


With these instructions, users can choose to run the application either directly using Streamlit or through Docker.