Overview

Author: dan@1howardcapital.com | daniel.dube@annalect.com

Pronunciation: ˈfiɡər > ˈfiɡyər == "figure"

Summary: Zero-dependency python framework for object oriented development. Implement once, document once, in one place.

With fgr, you will quickly learn established best practice... or face the consequences of runtime errors that will break your code if you deviate from it.

Experienced python engineers will find a framework that expects and rewards intuitive magic method usage, consistent type annotations, and robust docstrings.

Implement pythonically with fgr and you will only ever need to: implement once, document once, in one place.

Mission Statement

Ultimately, fgr seeks to capture and abstract all recurring patterns in application development with known, optimal implementations, so engineers can focus more on clever implementation of application-specific logic and good documentation than on things like how to query X database most efficiently, whether or not everything important is being logged correctly, where to put what documentation, and how to implement an effective change management scheme with git in the first place.

Getting Started

Installation

pip install fgr

Basic Usage

import fgr


class Pet(fgr.Object):
    """A pet."""

    id_: fgr.Field[int]
    name: fgr.Field[str]
    type_: fgr.Field[str] = {
        'default': 'dog',
        'enum': ['cat', 'dog'],
        'nullable': False,
        'required': True,
        }
    is_tail_wagging: fgr.Field[bool] = fgr.Field(
        default=True,
        enum=[True, False],
        nullable=False,
        required=True,
        )

Best Practice - Guard Rails at a Bowling Alley

fgr has been designed from the outset to teach best practice to less experienced python engineers, without compromising their ability to make effective and timely contributions.

To fgr, it is more important developers are able to make effective contributions while learning, rather than sacrifice any contribution at all until the developer fully understands why something that could be done many ways should only ever be done one way.

Exceptions

This is achieved primarily through the raising of exceptions. In many cases, if a developer inadvertently deviaties from a known best practice, fgr will raise a code-breaking error (informing the developer of the violation) until the developer implements the optimal solution.

Logging

fgr will commandeer your application's log.

• It will automatically redact sensitive data inadvertently introduced to your log stream that would have made your application fail audits.
• It will intercept, warn once, and subsequently silence print statements, debug statements, and other errant attempts at logging information in ways certain to introduce a known anti-pattern, vulnerability, or otherwise pollute your log stream.

In short, if fgr raises an error or otherwise does not support the thing you are trying to do: it is because the way in which you are trying to do it contains at least one anti-pattern to a known, optimal solution.

Advanced Usage

import fgr


class Flea(fgr.Object):
    """A nuisance."""

    name: fgr.Field[str] = 'FLEA'


class Pet(fgr.Object):
    """A pet."""

    id_: fgr.Field[str]
    _alternate_id: fgr.Field[int]

    name: fgr.Field[str]
    type_: fgr.Field[str] = {
        'default': 'dog',
        'enum': ['cat', 'dog'],
        'nullable': False,
        'required': True,
        }

    in_: fgr.Field[str]
    is_tail_wagging: fgr.Field[bool] = fgr.Field(
        default=True,
        enum=[True, False],
        nullable=False,
        required=True,
        )

    fleas: fgr.Field[list[Flea]] = [
        Flea(name='flea1'),
        Flea(name='flea2')
        ]


# Automatic case handling.
request_body = {
    'id': 'abc123',
    'alternateId': 123,
    'name': 'Bob',
    'type': 'dog',
    'in': 'timeout',
    'isTailWagging': False
    }
pet = Pet(request_body)

assert pet.is_snake_case == Pet.is_snake_case is True
assert pet.isCamelCase == Pet.isCamelCase is False
assert pet['alternate_id'] == pet._alternate_id == request_body['alternateId']
assert dict(pet) == {k: v for k, v in pet.items()} == pet.to_dict()

# Automatic, mutation-safe "default factory".
dog = Pet(id='abc321', alternate_id=321, name='Fido')
assert pet.fleas[0] is not dog.fleas[0]

# Automatic memory optimization.
assert Flea().__sizeof__() == (len(Flea.__slots__) * 8) + 16 == 24

class Flet(Flea, Pet):
    ...

class Pea(Pet, Flea):
    ...

assert Flet().__sizeof__() == (len(Flet.__base__.__slots__) * 8) + 16 == 72
assert Pea().__sizeof__() == (len(Pea.__base__.__slots__) * 8) + 16 == 72
assert Flet().name == 'FLEA' != Pea().name

# Intuitive, database agnostic query generation.
assert isinstance(Pet.is_tail_wagging, fgr.Field)
assert isinstance(Pet.type_, fgr.Field)

assert dog.type_ == Pet.type_.default == 'dog'

query = (
    (
        (Pet.type_ == 'dog')
        & (Pet.name == 'Fido')
        )
    | Pet.name % ('fido', 0.75)
    )
query += 'name'
assert dict(query) == {
    'limit': None,
    'or': [
        {
            'and': [
                {
                    'eq': 'dog',
                    'field': 'type',
                    'limit': None,
                    'sorting': []
                    },
                {
                    'eq': 'Fido',
                    'field': 'name',
                    'limit': None,
                    'sorting': []
                    }
                ],
            'limit': None,
            'sorting': []
            },
        {
            'field': 'name',
            'like': 'fido',
            'limit': None,
            'sorting': [],
            'threshold': 0.75
            }
        ],
    'sorting': [
        {
            'direction': 'asc',
            'field': 'name'
            }
        ]
    }

Local Logging

import fgr


class AgentFlea(fgr.Object):
    """Still a nuisance."""

    name: fgr.Field[str] = 'FLEA'
    apiKey: fgr.Field[str] = '9ac868264f004600bdff50b7f5b3e8ad'
    awsAccessKeyId: fgr.Field[str] = 'falsePositive'
    sneaky: fgr.Field[str] = 'AKIARJFBAG3EGHFG2FPN'


# Automatic log configuration, cleansing, and redaction.

print(AgentFlea())
# >>>
# {
#   "level": WARNING,
#   "time": 2024-02-26 18:50:20.317 UTC,
#   "log": fgr.core.log,
#   "data":   {
#     "message": "Calls to print() will be silenced by fgr."
#   }
# }
# {
#   "apiKey": "[ REDACTED :: API KEY ]",
#   "awsAccessKeyId": "falsePositive",
#   "name": "FLEA",
#   "sneaky": "[ REDACTED :: AWS ACCESS KEY ID ]"
# }

print(AgentFlea())
# >>>
# {
#   "apiKey": "[ REDACTED :: API KEY ]",
#   "awsAccessKeyId": "falsePositive",
#   "name": "FLEA",
#   "sneaky": "[ REDACTED :: AWS ACCESS KEY ID ]"
# }

Deployed Logging

import os
os.environ['ENV'] = 'DEV'

import fgr

assert (
    fgr.core.constants.PackageConstants.ENV
    in {
        'dev', 'develop',
        'qa', 'test', 'testing',
        'uat', 'stg', 'stage', 'staging',
        'prod', 'production',
        }
    )


class AgentFlea(fgr.Object):
    """Still a nuisance."""

    name: fgr.Field[str] = 'FLEA'
    apiKey: fgr.Field[str] = '9ac868264f004600bdff50b7f5b3e8ad'
    awsAccessKeyId: fgr.Field[str] = 'falsePositive'
    sneaky: fgr.Field[str] = 'AKIARJFBAG3EGHFG2FPN'


print(AgentFlea())
# >>>
# {
#   "level": WARNING,
#   "time": 2024-02-26 19:02:29.020 UTC,
#   "log": fgr.core.log,
#   "data":   {
#     "message": "Call to print() silenced by fgr.",
#     "printed": "{\n  \"apiKey\": \"[ REDACTED :: API KEY ]\",\n  \"awsAccessKeyId\": \"falsePositive\",\n  \"name\": \"FLEA\",\n  \"sneaky\": \"[ REDACTED :: AWS ACCESS KEY ID ]\"\n}"
#   }
# }

print(AgentFlea())
# >>>

fgr.log.info(AgentFlea())
# >>>
# {
#   "level": INFO,
#   "time": 2024-02-26 19:13:21.726 UTC,
#   "log": fgr.core.log,
#   "data":   {
#     "AgentFlea": {
#       "apiKey": "[ REDACTED :: API KEY ]",
#       "awsAccessKeyId": "falsePositive",
#       "name": "FLEA",
#       "sneaky": "[ REDACTED :: AWS ACCESS KEY ID ]"
#     }
#   }
# }

Planned Features

• Wiki / Sphinx Documentation Support Done!

  • fgr should support a simple interface for generating wiki / sphinx style documentation for packages.

• RESTful Framework / OpenAPI Support

  • fgr should support all aspects of an OpenAPI specification and provide corresponding framework functionality for HTTP request handling.

• Template Packages

  • fgr should include a Pet shop style demo API and python package as a template for developers to copy / paste from.

• Database Parse & Sync

  • fgr should be able to generate a python package with fully enumerated and optimized Objects (and a corresponding fgr API package) when supplied with access to a database for which at least one schema may be inferred.
    • CLI commands like $ fgr-api-from-sql ${api_name} ${sql_conn_string} . should instantly output two ideally structured package repositories for a RESTful python API and corresponding object management package.
    • The package could use any supplied credentials to either query a database directly or make requests to a deployed API. This means the same package used to power the API can be distributed and pip installed across an organization so business intelligence, data science, and other technical team members can manipulate data for their needs, while leaning on the package to optimize queries and stay informed around permission boundaries and request limits.

• Repo Generation

  • fgr should be expanded to optionally wrap any generated packages in a repository pre-configured with essentials and CI that should:
    • implement an ideal trunk-based branch strategy, inline with current best practices for change management and developer collaboration
    • enforce python code style best practices through automated linting and formatting
    • type-check python code and generate a report with mypy
    • run tests automatically, generate reports, and prevent commits that break tests
    • automatically prevent commits that do not adhere to standardized commit message conventions
    • using those conventions, automatically semantically version each successful PR and automatically generate and update a CHANGELOG.md file
    • automatically generate and publish secure wiki documentation
  • Generated repos may contain up to all of the following:
    • CHANGELOG.md
    • CODEOWNERS
    • CONTRIBUTING.md
    • .git
      • .git/hooks/
    • .github/workflows/
      • Support planned for gitlab and bamboo.
    • .gitignore
    • LICENSE
    • .pre-commit-config.yaml
    • pyproject.toml
    • README.md
    • /src
      • /package
      • /tests

• Async

  • Everything should be runnable as coroutines.

Credits

• @sol.courtney

  • Teaching me the difference between chicken-scratch, duct tape, and bubble gum versus actual engineering, and why it matters.

• pydantic

  • A portion of whose code for dealing with aggravating things like handling ForwardRefs I shamelessly copy, pasted, and re-purposed.

• python-semantic-release

  • Much of whose CI I shamelessly copy, pasted, and re-purposed.