Run the below script to build the Docker image and run all the services:
$ ./scripts/dev.sh [project]
Or run the following command to build the Docker image:
$ docker-compose build .
Run the following command to run all the services:
$ docker-compose up
Open http://localhost:3000/ in your browser to use the app.
Jupyter Lab is running on http://127.0.0.1:8888/lab.
Look at your terminal and find a corresponding URL with a token in the URL parameter;
e.g. http://127.0.0.1:8888/lab?token=
.
The backend server running Tornado and an iPython kernel is running at http://localhost:6789/api/?value=test.
Add the package to ./requirements.txt
and run:
$ ./scripts/server/setup.sh
Add the package to ./mage_ai/frontend/package.json
and run:
$ ./scripts/frontend/setup.sh
Instead of using breakpoint()
, add the following line to your code where you
want a debug:
import pdb; pdb.set_trace()
Attach to running container to use debugger. To get the container ID, run docker ps
and look in the NAMES
column.
$ docker attach [container_id]
Open the example.ipynb notebook for an interactive Python environment and connect your data to the app.
You can run the tool inside a Jupyter notebook cell iFrame using the method:
mage_ai.launch()
within a single cell.
Optionally, you can use the following arguments to change the default host and port that the iFrame loads from:
To stop the tool, run this command: mage_ai.kill()
mage_ai.launch(iframe_host='127.0.0.1', iframe_port=1337)
Directions at brew.sh.
More details at Homebrew website.
Note: We are currently using Node 14 as of 11/1/21, but Node 16 became the latest LTS version as of 10/26/21. You can try using Node 16, but if there are issues running the app, revert to Node 14.
# Node v14
$ brew install node@14
# Node v16
$ brew install node@16
If switching between Node versions is needed, use nvm, but uninstall any Node versions installed with Homebrew first. Refer to nvm's docs for details on installing and using different Node versions.
# Install nvm
$ curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.0/install.sh | bash
# Uninstall any existing node versions (e.g. through Homebrew) FIRST.
$ nvm install 17.9.0
$ nvm use 16 17.9.0
$ nvm alias default 17.9.0
More details at Yarn website.
$ npm install -g yarn
Change directory into the front-end folder.
$ cd mage-ai/mage_ai/frontend
Install Node modules using yarn
.
$ yarn install
While in the directory mage-ai/mage_ai/frontend
,
run the following command to launch the UI locally.
$ yarn run dev
Now visit http://localhost:3000/datasets to view the tool.
Change directory into the root folder.
$ cd mage-ai
Install Python packages
$ pip3 install -r requirements.txt
This server is used in the data preparation tool. It leverages ZMQ and WebSocket.
$ python3 mage_ai/server/server.py
You can optionally set the host or port environment variables:
$ export HOST=127.0.0.1
$ export PORT=5789
Or, you can set the host and port at runtime (see below).
While in the root directory, run the following command to launch the backend locally.
$ python3 mage_ai/server/app.py
or
$ python3 mage_ai/server/app.py --host 127.0.0.2 --port 1337
Now visit http://localhost:5789 to make HTTP requests to the backend server.
In your notebook or interactive Python environment, run the following code to use a local
version of the mage_ai
library that reads from your local repository:
import sys
sys.path.append('/absolute_path_to_repo/mage-ai')
import mage_ai
Load sample datasets to test and play with.
import mage_ai
from mage_ai.sample_datasets import list_dataset_names, load_dataset
dataset_names = list_dataset_names()
df = load_dataset('titanic_survival.csv')
- [WIP] How to add a chart for visualization
- [WIP] How to add a report
- [WIP] How to add a cleaning action
- [WIP] How to add a suggested cleaning action