saylaan / ai_navigation_app_2022

App for a Navigation Request using AI

Getting Started

The travel order resolver was an application created containing three components: The voice, the Natural Language Processing (NLP) and the pathfinder. The voice component records the voice of a user to a file. The NLP component transcribed this file and saved it to a text file. Later the text is analyzed for a valid train travel request and the departure and destination extracted. The pathfinder component took these two locations and used a graph structure to determine the fastest route between them. Later an application was created to use this functionality. Keywords: Natural Language Processing, Pathfinder, Artificial Intelligence, train

For having more information about the project and the methodology, please see the document "Report.pdf" in the root of the project`

The components

You can see all the three component necessary for the project in the folder /docs/jupyter-components There you will have :

• Audio recorder This component will record an audio sample, and save it to an audio file.

• NLP component This component takes an audio file and writes its content to a text file

• Transcript component This component takes a text file and extracts a travel request destination and departure

• Pathfinder component This component takes a departure and destination, and calculates the fastest path between the nearest train stations to these locations.

Launch interface

The infrastructure is made in Python with Flask framework

You have two possibility to use the app, with a container docker or a basic installation on your environment.

Container docker

1. First, you need to go to the infrastructure directory cd /app

2. Run the script ./launch_server.sh

3. Open a browser and go to the page http://localhost:5000/

Basic installation

1. First, you need to go to the infrastructure directory cd /app

2. Then, you need to install all dependencies pip install -r requirements.txt

3. Install all this module : pip install langdetect pip install icecream pip install spacy python -m spacy download fr_core_news_md

4. If you have any missing module during the installtion. Just run the command pip install name_module

5. After the project installation is complete, you can launch the app by going in cd /app/back and run the command python app.py

Troubleshooting

Make sure that your current working directory for the project is at the root of the repo