Postpwn

A smart task rescheduler for Todoist that optimally distributes your tasks based on rules you set.

Features

• Supports custom rules based on task labels
• Reschedules tasks based on customizable weight limits
• Can run as a one-off command or on a schedule

Usage

Currently the best way to use this project is through the docker image published to GHCR. Eventually, a Python package and standalone binary will be provided.

Keep in mind that this is still in development and the API is subject to change. I recommend pinning the version/SHA of the image you use until I get around to adding tests and proper versioning.

Example docker-compose file:

services:
  reschedule-tasks:
    init: true
    image: ghcr.io/alexzasorin/postpwn/app:latest
    command:
      [
        # Reschedule tasks that are not assigned to others, have a date, are not
        # recurring, have no deadline, and are not highest priority
        "--filter",
        "!assigned to:others & !no date & !recurring & no deadline & !p1",
        # Path to JSON rules file inside the container
        "--rules",
        "/app/rules.json",
        # Time zone for scheduling
        "--time-zone",
        "US/Pacific",
        # Cron string to run this at midnight every day 
        "--schedule",
        "0 0 * * *",
        # Your Todoist API token
        "--token",
        "${TODOIST_USER_TOKEN}",
      ]
    volumes:
      # Mount your rules file into the container
      - ~/.config/rescheduler/rules.json:/app/rules.json
  reschedule-recurring:
    init: true
    image: ghcr.io/alexzasorin/postpwn/app:latest
    container_name: reschedule-recurring
    command:
      [
        # Reschedule tasks that are not assigned to others, have a date, are overdue,
        # are recurring, have no deadline, and are not highest priority
        "--filter",
        "!assigned to:others & !no date & overdue & recurring & no deadline & !p1",
        "--time-zone",
        "US/Pacific",
        "--schedule",
        "0 0 * * *",
        "--token",
        "${TODOIST_USER_TOKEN}",
      ]
    volumes:
      - ~/.config/rescheduler/rules.json:/app/rules.json

Options

• --filter: Todoist filter to select tasks (default: "!assigned to:others & !no date & !recurring & no deadline")
• --rules: Path to JSON rules file
• --dry-run: Simulate changes without applying them
• --token: Todoist API token (can also be set via TODOIST_USER_TOKEN environment variable)
• --time-zone: Time zone for scheduling (default: "Etc/UTC")
• --schedule: Cron string for running on a schedule

Configuration

Rules File

Weights can be assigned to tasks based on their labels. The rescheduler will optimally distribute tasks based on these weights, with higher priority tasks being valued more.

Create a JSON file to define weights for different task labels and daily capacity:

{
  // Maximum weight of tasks allowed per day.
  "max_weight": 10,
  // Alternatively, you can set different weights for each day
  // "max_weight": {
  //   "sunday": 5,
  //   "monday": 10,
  //   "tuesday": 10,
  //   "wednesday": 10,
  //   "thursday": 10,
  //   "friday": 8,
  //   "saturday": 5
  // },
  "rules": [
    // filter - Todoist task label
    // weight - Weight of the task with this label
    { "filter": "@< 15 min", "weight": 2 },
    { "filter": "@< 60 min", "weight": 4 },
    { "filter": "@< 3 hrs", "weight": 8 },
    { "filter": "@> 3 hrs", "weight": 10 },
  ],
}

TODO

• Publish docker image to GHCR
• Update deployment to use new scheme
• Add proper logging
• Add proper README
• Update to latest Todoist API
• Add tests
  • Create pytest fixture for the reschedule params kwargs
  • Parametrize the reschedule params fixture?
  • Create pytest fixture for the Fake TodoistAPI
  • Create pytest fixture for the current datetime
  • Refactor FakeTodoistAPI to create "distribution" of final task dates for easier comparison, refactor test cases to use this
• Make logs not look like ass lol
• Switch to toml config???
  • Merge CLI args and rules config into a unified configuration
• Allow disabling "smart" rescheduling
• Consider retrying individual Todoist API calls instead of the entire
• Use Session from requests to implement retry logic? rescheduling function
• Catch improper cron string
• Try using V4 API of scheduler for typing
• Add limits as alternative to weights
• Add semantic release
• Publish executable using PyInstaller
• Add value to WeightedTask and increase value for older tasks, make optional
• Allow overriding the default values for each priority
• Move my deployment out of the main code
• Allow "punting" of tasks further than today

License

GPL-3.0 License