SalesSage is a product developed for TikTok TechJam 2024, integrating frontend (React) and backend (Flask) technologies. The application leverages pre-trained models such as snips-nlu and whisperX for tasks including sentimental analysis, ToDo list recommendations, and Automatic Speech Recognition (ASR). Furthermore, we implemented this on lark web application to host our web application on lark messanger.

1. ngrok
2. python 3.7.1
3. python 3.9
4. cuda 11.8
5. HuggingFace

Clone this repo before setting up

git clone https://github.com/Darkenles/SalesSage.git

Note: Remember to replace your HF Secret Key in 'run_whisperX.py' and 'test_whisperX.py' at the line where 'diarize_model' is being called, and subscribe to the repository provided in WhisperX for HF (WhisperX GitHub Repo).

Ensure Python is linked to Python version 3.9 before running the command below:

cd Backend
python -m venv venvone
venvone\Scripts\activate

Replace venvone with the name that your virtual environment to be called

Within your git clone repo, run the following commands

mkdir uploads
mkdir model_data

pip install torch==2.0.0+cu118 torchaudio==2.0.0+cu118 -f https://download.pytorch.org/whl/torch_stable.html
pip install -r requirements.txt

git clone https://github.com/m-bain/whisperX.git
cd whisperX
pip install -e .
cd ..

After accessing your virtual environment, cd into your repo clone

python server.py

cd snips-nlu

conda create --name myenv python=3.7.1

conda activate myenv

Replace myenv with the desired name for your environment.

Navigate to the snips-nlu folder. This is where the second flask app will be executed. Following which run the two commands below to set up the virtual environment.

pip install -r requirements.txt

conda install --yes --file conda_requirements.txt

python -m snips_nlu download en

python App.py

The version of React being used in this project is 18.0.0

1. Run npm install

cd Frontend
npm install

2. Run npm start to start the app in the development mode.

npm start

The web application browser should run on your http://localhost:3000

1. Create a new web app in the Lark Developer Console
2. Setup your ngrok and follow the instructions here ngrok guide
3. Run the frontend and backend scripts
4. Setup your web application to a link using ngrok by running the follow command

ngrok http 3000

5. With the link which is highlighted by the boarder as seen from this figure
   
   After getting the link, modify the web app configuration in ur lark developer for desktop homepage to the link.
   
   
6. After releasing the version on lark, This web application will be able to access by user through the lark messanger.

The frontend of our web application will display the results in the following format

Our users could easily read and understand the meaning of the various output.