Gravity Spy Plus 2.0

Prerequsities

Login to LIGO Server with SSH

• Register for a LIGO account
• Follow the instruction in LIGO account settings to set up SSH key

ssh your_user_name@ssh.ligo.org        

• Enter the password and login
• Choose LHO or LLO
  • When creating subject sets for the glitches, make sure you're on the correct LDG site:
    • If you are creating subject sets for Livingston glitches, select LLO
    • If you are creating subject sets for Hanford glitches, select LHO
• Choose any machine

Install Gravityspy-ligo-pipeline dependency package.

git clone https://github.com/Gravity-Spy/gravityspy-ligo-pipeline.git
cd gravityspy-ligo-pipeline
CONDA_OVERRIDE_CUDA="11.2" conda create -c conda-forge --name GS-plus tensorflow python=3.10 cudatoolkit=11.2 django python-nds2-client psycopg2 
conda activate GS-plus
python -m pip install .

Install Gravity Spy Plus

cd ..
git clone https://github.com/Gravity-Spy/gravityspy-plus.git

Login to Zooniverse

• Register for a Zooniverse account
• Add login credentials below into shell run control file (normally ~/.bashrc or ~/.zshrc)

#Panoptes Login
export PANOPTES_USERNAME=(your zooniverse username)
export PANOPTES_PASSWORD=(your zooniverse password)

Make Gravity Spy subject and upload to Zooniverse

This pipeline generate omega scans from given event time on both main channel and auxiliary channels. Then it concatenate main channels and auxiliary channels PNGs vertically into a subject. Finally it upload all subjects to Zooniverse subject sets.

Under the hood, given an event time of Gravity Spy event at which an excess noise event occurred, the interferometer id, the round number of Hveto rounds and number of auxiliary channels, it returns a set of combination images for main channel and auxiliary channels with highest correlation in multiple durations and upload it to the zooniverse project set.

• ifo (str): What interferometer had this an excess noise event.
• manual_list_of_auxiliary_channel_names (list)[optional, None]: This will override any auxiliary channel list that might have been supplied by the auxiliary_channel_correlation_algorithm and force this to be the auxiliary channels that are associated with this Subject.
• subject_set_id: This is the unique five digit ID from the Zooniver subject set

Manually run make_gspy_subject on a pre-selected set of auxiliary channels

cd gravityspy-plus
./manage.py make_gspy_subject --event-time 1238303043.3125 --ifo H1 --subject_set_id 12345 --manual-list-of-auxiliary-channel-names H1:ASC-AS_A_RF45_Q_YAW_OUT_DQ H1:LSC-REFL_A_LF_OUT_DQ H1:ASC-AS_A_RF45_Q_PIT_OUT_DQ

Use Hveto to select top N auxiliary channels within a period of event time with LLO

./manage.py make_gspy_subject --ifo L1 --subject_set_id 12345 --start-time 1253102469 --end-time 1253103469 

Or with LHO

./manage.py make_gspy_subject --ifo H1 --subject_set_id 12345 --start-time 1238303043 --end-time 1238304043

Or with automation script

• Download the required csv files (follow intructions in readme.txt in /detector_data)
• DQ.py = Data Quality (DQ) flagged glitches refer to a segment of data that has been marked as potentially affected by noise. These glitches are identified and flagged. In this script, we will be creating subjects based on the DQ flag.
• O3_L1.py and O3_H1.py = Automates the process of manually running make_gspy_subject on a pre-selected set of auxiliary channels. The O3 GPS times will be coming from the csv files.
• O4.py = Automates the process of using Hveto to select top N auxiliary channels within selected range of O4 GPS times.

cd gravityspy-plus/subject/management/commands/automate_gs_subjects
nohup python -u O3_L1.py > current_date.out 2>&1 &
#output will be put in .out file. this is helpful for when the code breaks and you need to rerun/need the gps time.

Parse Classifications from Zooniverse

This pipeline parse classifications from Zooniverse that are annotated by volunteers. The classification is whether the glitch in auxiliary channel is related to glitch in main/h(t) channel. Then it will save each classification as an item in psql database table, which records the annotation and subject information, such as channel names and event time.

Add login credentials below into ~/.bashrc

#Django Login
export GRAVITYSPYPLUS_DATABASE_NAME="gravityspyplus"
export GRAVITYSPYPLUS_DATABASE_USER="django_gravityspyplus_user"
export GRAVITYSPYPLUS_DATABASE_PASSWORD="(contact lab for password)"
export GRAVITYSPYPLUS_DATABASE_HOST="gravityspyplus.ciera.northwestern.edu"
export GRAVITYSPYPLUS_DATABASE_PORT="5432"

Parse classifications from Zooniverse and save into psql database

./manage.py parse_gspy_classifications --workflow-id=23089

Login to psql database

psql -U django_gravityspyplus_user -d gravityspyplus -h gravityspyplus.ciera.northwestern.edu

View parsed classifications table in psql database

\d
select * from classification_classfication;