pulseox 0.1.4

Tools to create, update, and monitor a system pulse and dashboard.

Introduction

The pulseox project provides python tools to create, update, and monitor a system pulse and dashboard using GitHub.

You can have your clients update a project page on GitHub with current status and have a unified dashboard which collects together the status of all projects (including projects which should have reported but have not).

PulseOx can use the notifier package to automatically notify you when project status changes.

Install

You can install as usual via something like

pip install pulseox

or

uv add pulseox

or if you want to develop in a fresh environment you can do something like

python3 -m venv .venv       # Create venv to get pip.
source .venv/bin/activate   # Activate venv.
pip install uv              # Install uv
uv venv --seed --clear      # Recreate venv since uv likes that
source .venv/bin/activate   # Source new venv
pip install uv              # Add uv to venv if you don't have global uv
uv sync                     # Sync dependencies.

Usage

Client

You can instantiate a pulseox client and use it to update a page in a GitHub project. The client will use the GitHub API and an access token to post content as shown below:

>>> from pulseox.client import PulseOxClient
>>> client = PulseOxClient(token=YOUR_GITHUB_PAT)
>>> content = ('Some content\n'
...            '- data I want to keep track of\n'
...            '- for dashboard monitoring\n')
>>> result = client.post(
...   owner, repo, 'example_project_page.md', content,  # required arguments
...   report='GOOD', note='')  #  these can be omitted and have defaults
>>> result.status_code in (200, 201)
True
>>> _ = client.post(owner, repo, 'alt_example.md', content, report='BAD',
...       note='We can report BAD runs as well')

The client will automatically include metadata at the bottom of the file indicating the following:

• report: (as provided to the client)
• updated: (timestamp for when the last update occurred)
• note: (note provided to client)

Dashboard

You can then use a dashboard creation tool to collect together all the posted content to get a summary using something like:

>>> import datetime
>>> from pulseox.specs import PulseOxSpec
>>> from pulseox.dashboard import PulseOxDashboard
>>> spec_list = [  # Example of job specifications we are tracking
...     PulseOxSpec(owner=owner, repo=repo, path='example_project_page.md',
...	      schedule=datetime.timedelta(minutes=10)),
...     PulseOxSpec(owner=owner, repo=repo, path='alt_example.md',
...	      schedule=datetime.timedelta(hours=10)),
...     PulseOxSpec(owner=owner, repo=repo, path='missing.md',
...       schedule='* * * * *')]  # can use cron string for schedule
>>> dashboard = PulseOxDashboard(
...     token=YOUR_GITHUB_PAT, owner=owner, repo=repo, spec_list=spec_list
... ).write_summary(force_refresh=True)  # Write summary to github
>>> print(dashboard.summary.text)  # Get text summary for local view
# Changes
- [alt_example.md](alt_example.md) None --> ERROR ...
- [missing.md](missing.md) None --> MISSING None
- [example_project_page.md](example_project_page.md) None --> OK ...
# ERROR
- [alt_example.md](alt_example.md) We can report BAD runs as well ...
# MISSING
- [missing.md](missing.md) error: (status_code=404) NOT FOUND None
# OK
- [example_project_page.md](example_project_page.md) ...

The specification list is also saved on GitHub. So after you have written the summary at least once, you can read the summary remotely via something like:

>>> dashboard = PulseOxDashboard(
...     token=YOUR_GITHUB_PAT, owner=owner, repo=repo).get_remote_data(
... ).write_summary(force_refresh=True)
>>> print(dashboard.summary.text)  # Get text summary for local view
# ERROR
- [alt_example.md](alt_example.md) We can report BAD runs as well ...
# MISSING
- [missing.md](missing.md) error: (status_code=404) NOT FOUND None
# OK
- [example_project_page.md](example_project_page.md) ...

Notification

You can use the notifiers package to send you notifications on status changes by providing a dictionary of keyword arguments when creating your dashboard. For example you could do something like:

dashboard = PulseOxDashboard(... notify={'telegram': {
    'token': YOUR_TELEGRAM_TOKEN, 'chat_id': YOUR_CHAT_ID}})

when creating your dashboard. See docs for notifiers for details on available providers and keyword arguments accepted by each notifier.

Summary Details

The summary will have sections for the following:

• ERROR: This comes first and consists of all client posts with status of BAD or posts which were reported but had problems..
• MISSING: This comes second and consists of all PulseOxSpec instances where an update has not been provided within the given schedule.
• OK: This comes last and consists of all PulseOxSpec instances which were posted with status OK within the required schedule.

The summary will be in markdown format if mode was set to 'md' and org-mode format if mode was set to 'org'. Empty sections will be omitted.

Within each section, there will be an entry like:

- <path_to_file> <note> <update_time>

with <path_to_file> being both the (relative) path to the posted file formatted as a link in either markdown or org format so if the user clicks on it, they will be taken to the give file.