lglass / ImageLabeller

Simple Flask app to allow users to label images and save results in a database

ImageLabeller

Simple Flask app to allow users to label images and save results in a database. Can be run either via docker-compose or directly. For deploying as a web-app, we recommend the docker-compose option - see instructions for this here.

Running with Docker(compose)

If you have Docker and docker-compose installed, just run the commands

docker-compose build
docker-compose up

Then point your browser to localhost:80/.

Requirements for running locally (native)

It is recommended to use a virtualenv or conda environment to run this, in order to avoid the risk of clashes with other python packages.

pip install -r requirements.txt

Run the flask app

From this directory, run

python il_app.py

Then navigate your browser to http://localhost:5000

Upload data

First, put the data you want to load into a zip archive.

• Put all the files into a single directory and cd to that directory.
• Do zip -f <zipfilename> . - it is recommended that you name the zipfile with "" in the filename to enable the coordinates and date to be inferred and saved along with the image.
• Login to ImageLabeller as an admin user, and go to the base URL plus /admin/upload then select your zipfile and click "Save".

Retrieve the data

If you are a user with admin rights, navigate to http://localhost:5000/admin/download to retrieve the assigned labels as a .csv or .json file.