Skip to main content

Nothing helps coder be more smarter & less dumber.

Project description

Nothing (not)

Nothing helps coders be more smarter, some cooler, less dumber, and much faster. not.

Take hold of the key 🔑 to gradual automation.

Installation

nothing-cli is very young, only an infant. 🐣

If you'd like to give it a try, it's available in alpha on PyPi.

pip install nothing-cli

In part due to its youth, only Python 3.7 and above are supported. 😬

The command for interacting with the tool is not. Get the overview of its subcommands like so:

not --help

Get a quick first taste by running:

not sample

You should now have a sample Procedure available, aptly named nothing. Get an overview of what it does and where it lives with the describe subcommand:

not describe nothing

You can invoke it with the do subcommand:

not do nothing

You'll be walked through the Procedure for doing... nothing. Enjoy! Folks don't do enough nothing, in my opinion.

Overview

A Realistic Example

The central concept of nothing-cli is the Procedure. A Procedure is simply a yaml file in a .nothing directory. We interact with them with the not command.

Procedures are not quite todo lists, not quite instructions, and not quite forms. The idea is to use them for infrequent, "toilsome" tasks that are easy to get lost in, annoying to document, and hard ––but very enticing–– to automate.

Here's a simple Procedure for a developer's personal checklist before starting work on a new feature branch:

---
title: Get started with a feature branch
description: A few preflight checks before you start coding
context:
  - dev_goal: Briefly, what do you want to accomplish with this branch?
  - __feature_branch_name: What name did you come up with for the new branch?
knowns:
  - main_branch: master
steps: |-
  Check out the main branch and get the latest changes:
  git checkout {main_branch} && git pull

  Think of a short, kebab-cased name that captures your goal:
  "{dev_goal}"

  Create the new branch:
  git checkout -b {__feature_branch_name}

  Push the new branch to the remote:
  git push -u {__feature_branch_name}

If you copy and paste the above into a file called preflight-checks.yml and save it in a directory called .nothing (either in your home or working directory), then you can "do" the Procedure by calling:

not do preflight-checks

For those without a terminal handy, it'd look like this:

asciicast

Explanation

The required keys of a Procedure yaml file are:

  • title
  • description
  • steps

Optionally, you can also add:

  • context
  • knowns

Let's break each key down.

title

This one is self-explanatory: It's the title of your Procedure! When creating a new Procedure with not new, the filename defaults to a slugified-version of this value, but you can change it during the process.

description

Also self-explanatory. The description is required because, well, since when has extra documentation hurt anyone?

steps

This is potentially the most unusual piece of a Procedure. You'll notice that the block of text is prepended by a |- and a newline.

That specifies a block scalar, a lesser-known feature of yaml. The very existence of this thing was a great inspiration to create this tool in the first place. It functions as a sort of lawless playground for plain text, supporting:

  • Multiline text and indentation.
  • A clear visual delineation of content from configuration (content here being your steps)
  • The specification of each step as a paragraph, which is is both obvious to read and easy to edit.

In line with the fun flirtiness of this tool, the last line of any step paragraph is highlighted, since the last line is a very common place to include a command to copy-paste, plus it adds gravitas. 😉

Within a step, any variable defined in context or knowns can be referenced in curly braces with a Python f-string-style template. Which brings us to:

context

Context points to a list that contains any mix of plain strings and dictionaries. The former is "simple" and the latter "complex."

context:
  - simple_variable
  - complex_variable: Answer this prompt to give a value to this context variable.

For simple_value, the user would be prompted at the start of the run for the value with a default phrase: Please provide a value for [variable name].

For a complex_variable, the dictionary key will be the variable name, and the value is the prompt used to ask for it. This is generally preferable, but simple context is there if you're in a hurry.

There's one more form for context known as "lazy context." Sometimes, it's not advantageous to prompt the user for a value at the very beginning of a run. With a lazy context variable, the user is prompted right before the first step that references it.

context:
  - __lazy_variable

The dunder prefix is what lets nothing-cli know the variable is lazy. You'll notice that __feature_branch_name in the Procedure above is lazy. Lazy context variables can simple or complex and you can use them all you want.

knowns

Knowns are hardcoded context variables, essentially. They look a lot like complex context. The subtle differences are:

  1. Knowns can only be specified as dictionaries.
  2. The value of a known dictionary item is its value.

This is useful if you want to reference something changeable over and over again (like a build URL) but the user doesn't necessarily need to know it off the top of their head. This way, authors can disseminate useful information and change it later without making major edits.

Inspo & Rationale

Once upon a time, we all read this article about "do-nothing scripting" by Dan Slimmon.

It's an incredible concept. I've personally wasted hours of my life trying to turn ever-more-complex aliases into shell functions. I've spent days attempting to automate scripts to do extremely infrequent –but highly toilsome– tasks that probably would not even have taken hours manually.

"Toil," as Slimmon terms it, sucks. We'll do anything to get away from it. Do-nothing scripts are a place to meet halfway: Nowhere near the cognitive overhead of writing an actual automation script, but more interactive and dynamic than pure documentation.

But, even logic-free & sugary sweet do-nothing scripts are written in the language of your choice can be fragile. Your stylistic decisions rot. There's no formality to creation or maintenance. Hardcoded strings sprinkled on off-the-cuff implementation. Suddenly, toil returns. Utility vanishes.

Realizing this as my team and my friends experimented more with the practice, I did the logical thing: I wrote an entire piece of software to automate the process of writing the code we write to prevent ourselves from writing too much code to automate stuff when we want to stop toiling and just write some freaking code.

It's a noble cause, I think. [Insert infrequent terrible process] always goes faster when you can just ask your teammate for that really specific command and copy-paste. Cut them some slack. Get the command from a terminal robot.

Pudding (The proof is in it)

Here is the original do-nothing script example from Slimmon

import sys

def wait_for_enter():
    raw_input("Press Enter to continue: ")

class CreateSSHKeypairStep(object):
    def run(self, context):
        print("Run:")
        print("   ssh-keygen -t rsa -f ~/{0}".format(context["username"]))
        wait_for_enter()

class GitCommitStep(object):
    def run(self, context):
        print("Copy ~/new_key.pub into the `user_keys` Git repository, then run:")
        print("    git commit {0}".format(context["username"]))
        print("    git push")
        wait_for_enter()

class WaitForBuildStep(object):
    build_url = "http://example.com/builds/user_keys"
    def run(self, context):
        print("Wait for the build job at {0} to finish".format(self.build_url))
        wait_for_enter()

class RetrieveUserEmailStep(object):
    dir_url = "http://example.com/directory"
    def run(self, context):
        print("Go to {0}".format(self.dir_url))
        print("Find the email address for user `{0}`".format(context["username"]))
        context["email"] = raw_input("Paste the email address and press enter: ")

class SendPrivateKeyStep(object):
    def run(self, context):
        print("Go to 1Password")
        print("Paste the contents of ~/new_key into a new document")
        print("Share the document with {0}".format(context["email"]))
        wait_for_enter()

if __name__ == "__main__":
    context = {"username": sys.argv[1]}
    procedure = [
        CreateSSHKeypairStep(),
        GitCommitStep(),
        WaitForBuildStep(),
        RetrieveUserEmailStep(),
        SendPrivateKeyStep(),
    ]
    for step in procedure:
        step.run(context)
    print("Done.")

Here it is translated to a nothing-cli Procedure:

---
title: Provision New User Account
description: Create and distribute an SSH key for a new user.
context:
  - new_user_username
  - __email: Find and copy the new user's email and paste here
knowns:
  - dir_url: http://example.com/directory
  - build_url: http://example.com/builds/user_keys
steps: |-
  Run:
    ssh-keygen -t rsa -f ~/{new_user_username}

  Copy ~/new_key.pub into the `user_keys` Git repository, then run:
    git commit {new_user_username}; git push

  Wait for the build job at {build_url} to finish.

  Go to 1Password
  Paste the contents of ~/new_key into a new document
  Share the document with {__email}

I know it's impolite to talk about LOC, but the nothing-cli procedure version weighs in at 17 to the original 43 (excluding blank lines). The drastically improved readability and editability should speak for itself. The Procedure above is functionally identical to the original do-nothing script. Run them side by side and see!

Planned Features

The alpha version of nothing-cli was meant to be as focused as possible. For this reason, some mechanics of the original do-nothing script were omitted.

Promptless Knowns & Context

In an ideal world, users could specify context as usual:

---
# diet-review.yml
context:
  - favorite_food: What's your favorite thing to eat?

But then circumvent the prompt by running the Procedure like this:

not do diet-review --favorite-food 'french omelet'

Even more useful would be specifying a this for knowns:

---
# secret-stuff.yml
knowns:
  - secret_password

Specifying a known with no value (as a plain yaml list item) would require the Procedure to be run with the value as a command-line option.

not do secret-stuff # refuses to run
not do secret-stuff --secret_password 'n3veR $h4re tHis--'

Chaining Procedures

For the composition-minded, it could be a boon to write small, related Procedures and have them run directly into each other, maybe even sharing context.

This could be specified at runtime:

not do proc-1 --chain proc-2 --chain proc-3

Or in the Procedure itself:

---
# use-chains.yml
title: Use chains
description: Demonstration of chain usage
context:
  - name
steps: |-
  Love yourself and your colleagues, {name}.

  Choose composition.
chain:
  to: review-chaining
  pass_context: yes

Conditional Chaining

We're getting absolutely crazy here, but adding several features could allow for supercharged chains:

---
# ci-stuff.yml
title: CI/CD Toil
description: Boring CI thing, one day you'll automate. But not now.
context:
  # add syntax for limited choice context
  - deploy_phase[dev/prod]: What kind of deploy are you doing?
chain:
   # use context to determine where chain
  - if:
      deploy_phase_is: dev
      to: dev-deploy
  - if:
      deploy_phase_is: prod
      to: prod-deploy

This one is surely a moonshot but damn... I wanna do it 🤩

And more!

My head is full of ideas! Perhaps yours is too. If you believe in this tool, drop me a line, I'll be overjoyed and say nice things to you. 😸

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for nothing-cli, version 0.2.1
Filename, size File type Python version Upload date Hashes
Filename, size nothing_cli-0.2.1-py3-none-any.whl (38.0 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size nothing-cli-0.2.1.tar.gz (32.9 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page