Skip to main content

Tackle box is a declarative DSL for building modular workflows and code generators. Tool is plugins based and can easily be extended by writing additional hooks or importing external providers creating a web of interoperable CLIs.

Project description

Tackle

pypi python codecov main-tests

Tackle is a language for building modular code generators and declarative CLIs. It can make any config file dynamic with both strong and weakly typed programmable flow control common to a general purpose programming language. Basically you can write a fully functional CLI and Turing-complete program in yaml. It's wild.

Warning: Tool is in it's very early phases and only now considered stable enough to use. It may never reach 1.x version as, if it gets enough stars, it will be converted to a spec and re-written in rust (ie give it a star if you'd like to see that).

Install

Note: tackle can install dependencies on its own. Check docs for advanced installation methods to isolate tackle from your system python.

python -m venv env && source env/bin/activate
pip install tackle-box

Hello world

To call tackle, create a yaml file and run tackle hello-world.yaml.

Simply use the print hook.

hw->: print Hello world!

Which using jinja templating can be called in four different ways.

words: Hello world!
expanded:
  ->: print
  objects: "{{words}}"
compact->: print {{words}}
jinja_extension->: "{{ print(words) }}"
jinja_filter->: "{{ words | print }}"

And can also have interactive prompt hooks.

name->: input
target:
  ->: select Say hi to who?
  choices:
    - world
    - universe
hello->: print My name is {{name}}. Hello {{target}}!

Hooks can have loops, conditionals, and other base methods.

words:
  - Hello
  - cruel
  - world!
expanded:
  ->: print {{item}}
  for: words
  if: item != 'cruel'
compact->: print {{item}} --for words --if "item != 'cruel'"

Hooks can be written in python.

from tackle import BaseHook

class Greeter(BaseHook):
    hook_type: str = "greeter"
    target: str
    args: list = ['target']
    def exec(self):
        print(f"Hello {self.target}")

Or new hooks can be declaratively created with tackle.

greeter<-:
  target: str
  args: ['target']
  exec:
    hi->: print Hello {{target}}

And both can be called in the same way.

hello: world!
compact->: greeter {{hello}}
expanded:
  ->: greeter
  target: "{{hello}}"
jinja_extension->: "{{ greeter(hello) }}"
jinja_filter->: "{{ hello | greeter }}"

Declarative hooks are strongly typed objects with many declarative fields.

words<-:
  hi:
    type: str
    regex: ^(Bonjour|Hola|Hello)
  target: str

p->: print {{item}} --for values(words(hi="Hello",target="world!"))

Which can have methods that extend the base.

words<-:
  hi: Wadup
  say<-:
    target: str
    exec:
      p->: print {{hi}} {{target}}

p->: words.say --hi Hello --target world!

And also support inheritance.

base<-:
  hi:
    default: Hello

words<-:
  extends: base
  say<-:
    target: str
    exec:
      p->: print {{hi}} {{target}}

p->: words.say --target world!

And will later support extending common schemas like OpenAPI, protobuf, and json schema

Hooks and fields can have documentation embedded in them.

greeter<-:
  help: Something that greets
  hi:
    default: Yo
    description: Initial greeting.
  target:
    type: str
    description: Thing to greet.
  exec:
    p->: print {{hi}} {{target}}

That [will] drive a help screen by running tackle hello-world.yaml --help -> Coming soon

And last, everything can be imported / called remotely from github repos.

import-hello_>: import robcxyz/tackle-hello-world
call->: greeter world!
# Or
local-call->: tackle hello-world.yaml
remote-call->: tackle robcxyz/tackle-hello-world --version v0.1.0

Creating a web of declarative CLIs.

Use Cases

Topics

Roadmap

  • Declarative schemas
    • json schema
    • openapi
    • protobuf
  • Cached providers
  • State management

Contributing

Contributions welcome particularly for additional hooks and suggestions around their interfaces. If contributing to the core parser, please make sure any changes are non-breaking by running tests (make test). For writing providers / hooks, please make sure to use the testing conventions laid out in the existing providers whereby the test uses pytest fixtures to change the directory into the test's directory. This is so the tests can be run from both the root of the project and from within the test directory to facilitate debugging.

Code of Conduct

Everyone interacting in the tackle-box project's codebases, issue trackers, chat rooms, and mailing lists is expected to follow the PyPA Code of Conduct.

Credit

Special thanks to the cookiecutter community for laying the basis for this project.

Project details


Download files

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

Source Distribution

tackle-box-0.3.0b5.tar.gz (157.9 kB view hashes)

Uploaded Source

Built Distribution

tackle_box-0.3.0b5-py2.py3-none-any.whl (226.0 kB view hashes)

Uploaded Python 2 Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page