tsq 0.0.18

A Friendly Task Scheduler Frontend

A Friendly Task Scheduler Frontend

Introduction

Installation

pip install tsq

Cutting edge version from the repository:

pip install git+https://github.com/admk/tsq.git

Usage

Here’s an instruction on how to use tq, the front-end CLI for task-spooler:

Basic Usage

tq [options] [command]

Common Commands:

1. list (ls): Shows job information in a compact format. Or simply call tq without any arguments.

   tq list
   

2. ids (id): Displays the job IDs.

   tq ids
   

3. info: Provides detailed job information.

   tq info
   

4. commands (cmd): Shows the commands of all jobs.

   tq commands
   

5. outputs (out): Displays job outputs.

   tq outputs
   

6. add (a): Add a new job to the queue.

   tq add [options] <command>
   

   • -G: Specify the number of GPUs required.
   • -N: Specify the number of slots required.
   • -u: Ensure the command is unique.
   • -f: Add jobs from a file or standard input.

7. wait (w): Wait for jobs to finish.

   tq wait
   

8. kill (k): Kill running jobs.

   tq kill <job_id>
   

9. remove (rm): Remove jobs from the queue.

   tq remove <job_id>
   

10. rerun (rr): Rerun a job.

tq rerun <job_id>

11. requeue (rq): Requeue a job.

tq requeue <job_id>

Additional Options:

• -V: Print version info.
• -rc: Use a specific configuration file.
• -b: Specify the backend (dummy, ts).
• -g: Assign a group to a job.

For adding a job with multiple resources:

tq add -G 2 -N 4 "python train_model.py"

For reading jobs from a file:

tq add -f job_list.txt