eatyourgreens / Zooniverse_SLSN

Classify super luminous super novae based on their light curves

Lasair-Zooniverse Integration

This repository demonstrates integration of lasair, a broker for astronomers studying transient and variable astrophysical sources and the Zooniverse citizen science platform. The code here was used to prepare daily data uploads based on new alerts from lasair for the Superluminous Supernova citizen science project hosted on Zooniverse.

Run with Docker

You'll need to have Docker Desktop installed and running.

Copy docker-compose.template.yml to docker-compose.yml and add your Panoptes secret and Lasair token.

Build the Docker image.

docker compose build

This will install all the Python dependencies, then run the run.py script to build and upload Zooniverse subjects. Read the Dockerfile in this repo to see exactly what the build runs.

Run with GitHub Actions

The 'build subjects' action will run a Docker build, installing the Python dependencies and running run.py.

You can run it manually from the Actions tab. It is scheduled to run weekly on Wednesday mornings from the main branch, to add new subjects to the active workflow. Run it manually from the json-subjects branch to upload new JSON light curves.

Installation

Clone this repo

$ git clone https://github.com/rafia17/Zooniverse_SLSN.git

Change directory into Zooniverse_SLSN and install dependencies (e.g. wget, matplotlib, panstamps, panoptes_client, confluent-kafka, lasair).

$ pip install -r requirements.txt

Run example

To run the example do

$ python run.py

To get help, try:

$ python run.py --help

Add the LASAIR_TOKEN variable to your environment. The config.ini file contains important parameters for this example. The example script queries the lasair_6ZooniverseTestSN-likecandidatesinlast14days TOPIC from the lasair kafka stream at KAFKA_SERVER i.e. kafka.lsst.ac.uk:9092. The kafka stream returns all alerts added to that stream since the stream was last queried by the GROUP_ID, 'Test' in this case. For this example the number of alerts returned is limited (RECORDS_LIMIT) to 20. The script downloads lightcurve data for each alert using the LASAIR_TOKEN and saves these to DATA_DIR. From the downloaded lightcurve data, the script produces a lightcurve plot and grabs a PanSTARRS-1 (PS1) image at the location of each alert (see example below) which are also saved to DATA_DIR. Note: If the kafka stream does not return any alerts it may be because the *** GROUP_ID *** in the config.ini has already queried all the latest alerts from the stream. In this case try changing GROUP_ID *** to something else and try running the script again *** OR execute the code with the --test option, which will add a random UUID to the end of the *** GROUP_ID ***.

The plots and PS1 image are required to produce subjects for upload to a specific Zooniverse workflow and project identified by WORKFLOW_ID and PROJECT_ID located at the Zooniverse endpoint (ENDPOINT). Note: The example script will fail to upload to the Zooniverse project at this point unless you have the correct premissions for that project. To test the upload to Zooniverse build a Zooniverse project of your own and update the WORKFLOW_ID and PROJECT_ID in config.ini for your project.

Example Subject data for ZTF20acufbmq

Customisation

To customise integration between lasair and your own Zooniverse project create your own lasair_zooniverse_class take a look here to see how this was implemented for the Superluminous Supernovae project.

In development

Realtime processing of volunteer classifications with Caesar

Installation

conda env create --file environment.yml

You will also need to install miclaraia/caesar_external in the conda environment.

conda activate lasair
git clone https://github.com/miclaraia/caesar_external.git
cd caesar_external
pip install -e .

Next setup caesar_external following these steps relating to caesar config new: https://zooniverse.github.io/swap/setup.html#caesar-config-command-reference

python caesar_consumer.py

We note that posting directly to lasair has been disabled in caesar_consumer.py. If you would like to post your results back to the Lasair database, please get in touch with the Lasair team ([e-mail]. ) to discuss this possibility. If the team agrees, you will receive the custom urls/api requests that need to be included in the send_lasair function at Line 295 of caesar_consumer.py