mongo_queue

Task queue built on mongo with channels and unique job id.

Website • autobotAI Automation Platform

Inspired by kapilt/mongoqueue

Change Log:

v0.2.1

• Added optional locking when finding job by id, so it can be marked completed.

v0.1.9

• Added method to get pending jobs by channels

v0.1.8

• Added delay while queuing a new job. This allows user to delay the job execution by x seconds. It is similar to sleep but this allows the functionality for the fresh jobs.

v0.1.7

• Added index for repair operation.

v0.1.6

• Added utility methods for getting running jobs and pending jobs count.

v0.1.5

• Fixes issue with the next job method, it was picking up jobs with dependency

v0.1.4

• Improved the job.next method. Removed the lookup and removed double operation.

v0.1.3

• Added dependency index for faster lookup. Update complete method to retry 3 times while pulling the dependencies.
• Corrected pull dependency query by adding filter. It was updating all the documents.

v0.1.2

• Added diskUsage for larger queue dependency resolution.

v0.1.1

• Added find_and_update for finding the next job added process to pick the next job if the previous is already locked with multiple retries..

v0.1.0

• Added optional inc_attempt parameter for job.release. This will allow user to choose if they want to increment the attempt when releasing a job.

v0.0.9

• Added method find_by_id to find a job by it's id.

v0.0.7

• Added mongo backward compatibility. The aggregate function was using lookup which is only available after Mongo 3.6 (Not avaialble in the DocumentDB), Modified lookup to use old syntax.

v0.0.6

• Added sleep and state feature while releasing a job. This provides a way to not pickup job until provided seconds and store state for long running jobs.

v0.0.5

• Added depends_on feature. You can create dependency between jobs by supplying depends_on[] with previously created job ids.

v0.0.3

• Added unique index with job_id and channel. This is to make sure that the same job is not added multiple times. If not job id provided an unique id generated by default.

Usage

Install the package.

pip install mongo_queue

Usage Example:

• Create Queue Object

from mongo_queue.queue import Queue
from pymongo import MongoClient

queue = Queue(MongoClient('localhost', 27017).task_queue, consumer_id="consumer-1", timeout=300, max_attempts=3)

• Add task to queue default channel

queue.put({"task_id": 1})

• Add task to queue with priority to default channel

queue.put({"task_id": 1}, priority=1)

• Add task to queue in a specific channel

queue.put({"task_id": 1}, priority=1, channel="channel_1")

• Add task to queue with unique job_id

queue.put({"task_id": 1}, priority=1, channel="channel_1", job_id="x_job")

• Add task with dependency

job1 = queue.put({"task_id": 1}, priority=1, channel="channel_1", job_id="x_job")
job2 = queue.put({"task_id": 2}, priority=1, channel="channel_1", job_id="x_job", depends_on=[job1])

• Get the next job to be executed from the default channel

job = queue.next()

• Get the next job to be executed from a specific channel

job = queue.next(channel="channel_1")

• Update job progress for long-running jobs

job.progress(count=10)

• Put the job back in queue, this will be picked up again later, this will update attempts after max attempts the job will not be picked up again.
• You can also set state and sleep while releaseing a job
• sleep in seconds. The job will not be picked up again till the sleep time expires.
• state you can store state in the job for long running jobs.

job.release()
# or
job.release(sleep=10, state={"some": "state"})

• Put the job back in queue with error, this will be picked up again later, this will update attempts after max attempts the job will not be picked up again.

job.error("Some error occured")

• Complete the job. This will delete job from the database.

job.complete()

Build Steps

# Setup venv of python version 3.6 and above
python3.9 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
pip install wheel
pip install --upgrade twine
rm -rf dist
# Goto https://pypi.org/manage/account/token/ and capture the `$HOME/.pypirc` config and create the file.
vim $HOME/.pypirc
python3 setup.py sdist bdist_wheel
python3 -m twine upload -r pypi dist/* 

Local Development and Testing

export MONGO_URI=mongodb+srv://username:pwd@mongourl/test?retryWrites=true&w=majority
cd mong_queue # Root directory of the package
python3.9 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
python3 -m unittest mongo_queue.test