Thami0011/Transcript_Darija_Chatbot

# Darija Transcript chatbot 🤖 This project is an intelligent chatbot designed to understand and respond to user queries in Darija (Moroccan Arabic). It combines automatic speech recognition (ASR), natural language understanding, and AI-powered classification to assist users in a banking context **_AttijariWafaBank_**. #### The chatbot is capable of: - Understanding different languages like French, English, Standard Arabic, Morrocan Darija Arabic ... - Taking as an input both Text messages and Audio messades. _wav format_ - Interpreting data to respond within a banking context. - Denying every inference not related to the banking system with a user friendly prompt. #### The app uses : - FastAPI to deploy the app with different endpoints. - Ollama to send requests via its API. - ATLASAI as an LLM model to understand and respond in Darija Dialect. - Python-Based ASR models like Whisper and Wav2Vac2 to transcribe audio files. As a basic understanding of how the app works, it accepts both text and voice messages via : #### 🚀 Endpoints > **_/chat_** > POST method that accepts a body request JSON: > `{"prompt":"textPrompt"}` > **_/voice_** > POST method that acepts as a body request a JSON: > `{"record":MultipartFile}` --- ## 📄 Installation ###### 📍 Local Installation Requirements : - Python3.12 - FFMPEG - OLLAMA 1. **_(OPTIONNAL)_ Create a Python Virtual Environment :** > Start by creating a python virtual environment to separate diffenrent project depedencies from your local Python dependencies, with the command `python -m venv path_and_name_of_your_venv` 2. **Copy the project to you local environment :** > Inside your working directory, run the command `git clone github.com` to download the project locally. 3. **Install dependencies :** > To install them all at one, use the command: > `pip install -r requirements.txt` or `python3 -m pip install -r requirements.txt` 4. **Add ATLASAI model to OLLAMA** …