IDA aims to become an all-in-one natural language based interface for Data Science Workflows. It's core purpose is to ease the task of a Data Scientist by providing a chat interface. At the moment, IDA is capable of handling the following tasks:
Here's a video demonstration of IDA: https://youtu.be/7FQTiFdqyDE
Coming soon:
IDA currently consists of three main components:
ida-ws
: A Java Spring server that provides the IDA REST-API.ida-chatbot
: The Angular webclient for IDA.librarian
: A general purpose toolkit to programmatically work with software libraries.You will need at least the following tools:
Docker is not directly supported by Windows 10 Home and Windows <= 8.1. Use Docker Toolbox if you are using an unsupported OS.
IDA is set up via bash-scripts. If you are using Windows you will have to install bash, e.g. by installing cmder, by installing Cygwin or via WSL.
To get everything up and running follow these steps:
docker run -d -p 5000:5000 --name registry registry:2
docker swarm init --advertise-addr 127.0.0.1
./services/deploy-dev.sh
.http://127.0.0.1:3080/
.ng serve
and spring-boot-devtools
internally. Updates to the sources (.class
-files in case of Spring) should be reflected automatically.docker stack rm ida-stack-dev
.IDA does not yet support distributed environments and is therefore not ready for big production use-cases. To run IDA in a single-node production environment you need:
To start the production stack:
docker run -d -p 5000:5000 --name registry registry:2
docker swarm init --advertise-addr [SERVER_IP]
where [SERVER_IP]
is the IP on which IDA should be accessible../services/deploy.sh
http://[SERVER_IP]/
.docker stack rm ida-stack
.