toadstool 0.1.0

Python Load Tools Suite

Toadstool

Python load tools -- An opinionated solution for cleanly loading files directly into Python objects.

This library works by creating custom Python finders and loaders and injecting them into the Python meta search path.

The Python import system is explained in great detail here.

Consider a json file sample.json

{
    "employee": {
        "name":       "sonoo",
        "salary":      56000,
        "married":    true
    },
    "menu": {
        "id": "file",
        "value": "File",
        "popup": {
          "menuitem": [
            {"value": "New", "onclick": "CreateDoc()"},
            {"value": "Open", "onclick": "OpenDoc()"},
            {"value": "Save", "onclick": "SaveDoc()"}
          ]
        }
      }
}

can be loaded and used in the following way:

import toadstool
import sample

# Or import using ToadContext
from toadstool.utils.utils import ToadContext
from toadstool.loaders.json_loader import JsonLoader
with ToadContext(JsonLoader):
  import sample

>>> sample.
sample.employee  sample.json      sample.menu
print(sample.menu)
> {'id': 'file', 'value': 'File', 'popup': {'menuitem': [{'value': 'New', 'onclick': 'CreateDoc()'}, {'value': 'Open', 'onclick': 'OpenDoc()'}, {'value': 'Save', 'onclick': 'SaveDoc()'}]}}

Requirements

Requires Python 3.10

Install

Basic Install

$ python3 -m pip install toadstool

Advanced Install

$ python3 -m pip install toadstool[options] Where options are any of the following (or combinations thereof):

• toml For .toml support using toml library for Python versions prior to Python 3.11, which introduced tomllib into the standard library
• yaml For .yaml support using PyYaml
• gql For .gql or .graphql query support using graphql-core library

Usage

To use this module, simply import toadstool before any import you wish to use the included loaders for. This can be at the top of your source file or just before any imports for the supported filetypes.

Loaders

GraphQL (.gql | .graphql)

Load graphql queries directly as graphql.language.ast.DocumentNode objects from the GraphQL Core Library. Allows direct importing of queries/mutations/subscriptions/fragments (aka GraphQL operations).

Example:

Given a GraphQL query file names queries.graphl or queries.gql with the following contents:

query HeroComparison($first: Int = 3) {
  leftComparison: hero(episode: EMPIRE) {
    ...comparisonFields
  }
  rightComparison: hero(episode: JEDI) {
    ...comparisonFields
  }
}

fragment comparisonFields on Character {
  name
  friendsConnection(first: $first) {
    totalCount
    edges {
      node {
        name
      }
    }
  }
}

mutation CreateReviewForEpisode($ep: Episode!, $review: ReviewInput!) {
  createReview(episode: $ep, review: $review) {
    stars
    commentary
  }
}

then you can import the contents of the file either as a whole module:

import queries
print(queries.__dict__)

>
{'__name__': 'queries', '__doc__': None, '__package__': '', '__loader__': GqlImporter('queries.gql'), '__spec__': ModuleSpec(name='queries', loader=GqlImporter('queries.gql')), 'HeroComparison': DocumentNode, 'operations': {'HeroComparison': DocumentNode}, '__file__': 'queries.gql'}

or using specific query names:

from queries import HeroComparison
print(HeroComparison.definitions)

> (OperationDefinitionNode at 0:180, FragmentDefinitionNode at 182:339)

Also tracks all operations in a module dict as queries.operations

JSON (.json)

Loads JSON objects using builtin json library. The top-level JSON keys are stored as attirbutes for the module and the whole json converted dict is stored as imported_name.json For example, the following file sample.json

{
    "employee": {
        "name":       "sonoo",
        "salary":      56000,
        "married":    true
    },
    "menu": {
        "id": "file",
        "value": "File",
        "popup": {
          "menuitem": [
            {"value": "New", "onclick": "CreateDoc()"},
            {"value": "Open", "onclick": "OpenDoc()"},
            {"value": "Save", "onclick": "SaveDoc()"}
          ]
        }
      }
}

can be loaded and used in the following way:

import toadstool
import sample

>>> sample.
sample.employee  sample.json      sample.menu
print(sample.menu)
> {'id': 'file', 'value': 'File', 'popup': {'menuitem': [{'value': 'New', 'onclick': 'CreateDoc()'}, {'value': 'Open', 'onclick': 'OpenDoc()'}, {'value': 'Save', 'onclick': 'SaveDoc()'}]}}

You also have the json root object available at sample.json:

import toadstool
import sample

print(sample.json)
>{'employee': {'name': 'sonoo', 'salary': 56000, 'married': True}, 'menu': {'id': 'file', 'value': 'File', 'popup': {'menuitem': [{'value': 'New', 'onclick': 'CreateDoc()'}, {'value': 'Open', 'onclick': 'OpenDoc()'}, {'value': 'Save', 'onclick': 'SaveDoc()'}]}}}

TOML (.toml)

Loads TOML files such that each top-level table becomes an attribute of the imported module. Also loads the whole TOML file as a dictionary under the toml attirbute (which will overwrite any table from the file with the name toml as well). For example, if you have example.toml with the following contents:

[project]
name = 'Toadstool'
description = 'Python Load Tools Suite'
readme = 'README.md'
requires-python = "~=3.10"
license = { file = 'LICENSE' }
version = '0.1.0'
authors = [{ name = 'Andrés Alejos', email = '[e-mail]' }]
classifiers = [
    'License :: OSI Approved :: MIT Licens',
    'Programming Language :: Python :: 3',
    'Topic :: Software Development',
    'Topic :: Utilities',
]
keywords = ["import", "loader", "meta", "sys"]
urls = { Home = "https://github.com/acalejos/toadstool" }

[project.optional-dependencies]
gql = ['graphql_core>=3.2.3']
yaml = ['pyyaml >= 5.3.1']
toml = ['toml >= 0.10.2;python_version < "3.11"']
all = ['toadstool[gql]', 'toadstool[yaml]', 'toadstool[toml]']

[sample]
name = 'Sample'

[[Root]]
name = 'Root'

can be loaded and used in the following way:

import toadstool
import example

>>> example.
example.Root     example.project  example.sample   example.toml

YAML (.yaml | .yml)

Loads a YAML file, assigning each top-level key as a module attribute (similar to the JSON loader). Also loads the whole YAML definition as a dict into the yaml attirbute. If the YAML file has more than 1 YAML definition (which is legal within the YAML definition), then only the yaml attribute is set, and is a list[dict]. Suppose you have a YAML file called sample.yaml with the following contents:

document: 1
name: 'erik'

then you could do the following:

import toadstool
import sample

>>> sample.
sample.document sample.name sample.yaml

and suppose you had a similar multifile YAML file:

document: 1
name: 'erik'
---
document: 2
name: 'config'

then you could do the following:

import toadstool
import sample

>>> sample.
sample.yaml

Config (.ini | .cfg | .config)

Uses the configparser package from the standard library to load the file and assign any top-level key to the module's attribute, and the entire ConfigParser object to the config attribute. Suppose you have the following config file sample.ini:

[http]
port=8080
username=httpuser
[https]
port=8043
username=httpsuser
[FTP]
port=8043
username=ftpuser
[database]
driverclass   = com.mysql.jdbc.Driver
dbName        = mydatabase
port          = 3306
username      = root

then you could do the following

import toadstool
import sample

>>> sample.
sample.http  sample.https sample.FTP sample.database sample.config

CSV (.csv)

Loads a CSV file and loads the contents into rows and columns attributes as list[list[str]]. This will attempt to determine if the CSV contains a header as the first row, and will do the following according to that determination:

If Header:

• Creates attributes named_rows and named_columns
  • named_columns: Each column is a dict keyed on the column header and contains the whole column
  • named_rows: Each row contains a dict with each column entry keyed from its header to the value at that row.
• Populates each column name as an attribute name that will correspond to its respective column
• Populates an attribute fieldnames with the header values.

For any CSV:

• Creates a rows attirbute which contains a list[list[str]] with the CSV content rows (skips header if exists)
• Creates a columns attirbute which contains a list[list[str]] with the CSV content columns

Suppose you have a file called sample.csv with the following contents:

Index,Customer Id,First Name
1,DD37Cf93aecA6Dc,Sheryl
2,1Ef7b82A4CAAD10,Preston
3,6F94879bDAfE5a6,Roy
4,5Cef8BFA16c5e3c,Linda
5,053d585Ab6b3159,Joanna

then you can do the following:

import toadstool
import sample

>>> sample.
sample.First_Name         sample.named_rows         sample.rows
sample.Index              sample.columns            sample.named_columns
sample.Customer_Id        sample.fieldnames

ToadContext

Located in toadstool.utils.utils, ToadContext is a Context Manager to allow imports with toadstool loaders without permanently changing the sys.meta_path. Using this, you can import your files the same way as you would using import toadstool from within a context without having to actually import toadstool. You must explicitly pass loaders you wish to use as arguments.

This context manager does not yield anything, so the proper usage is:

with ToadContext(Loader | list(Loaders)):
    import my_module

Limitations

Toadstool works by injecting the supported Loaders into the sys.meta_path during Toadstool's module init. The sys.meta_path is a list of class instances that contain the find_spec and exec_module methods. This occurs the first time that Toadstool is imported (upon init of Toadstool), after which toadstool is cached in the sys.modules data structure, which is what is checked when any module is referenced before deferring to the loaders in sys.meta_path and searching for the module. Therefore, after the first import toadstool that occurs per Python interpreter session, the module __init__.py might not be run again and thus if you do anything to alter the sys.meta_path after import toadstool then the Loaders may no longer be in the search path.

One possible solution is to use importlib.reload(toadstool) which reruns the module __init__.py code (only code not contained in if __name__ == '__main__' block).

Another possible solution is to use the ToadContext when you wish to import using the Toadstool Loaders, This could prove to be redundant, but it also ensures that the sys.meta_path is only altered while importing the modules that require toadstool and removes the Loaders upon existing the context.

Contributing

Feel free to open feature requests for new features or simply supported file extensions. If you feel so inclined, you can also open up a pull request to merge in your own loader. Creating a new loader is very straight forwards and can be accomplished by:

• Creating a file in toadstool/loaders with the name {filetype}_loader.py

• Implementing a class that inherits from the Loader class located in toadstool/loaders/base_loader. Your class must implement the following:

  • Add a class attribute file_exts specifying your file extension(s) either as a str for a single extension or list[str] for multiple extensions. For example:

  class CsvLoader(Loader):
    file_exts="csv"
  

  class GqlLoader(Loader):
    file_exts=["gql","graphql"]
  

  • Define def exec_module(self, module): where your functionality will be implemented. This will typically look like updating the module.__dict__ with values according to how you load the file

• If your loader has dependencies:

  • Specify a new optional dependency in pyrpoject.toml under [project.optional-dependencies]
  • Add your dependency to the all optional dependency by adding it as toadstool[your_dep]

• Add your loader to the list all_loaders in the root __init__.py file as a 3-tuple with:

  • Path to your module in dotted notation
  • Name of your loader class
  • Name of dependency from pyrpoject.toml if it exists or None

Refer to any of the existing loaders in toadstool/loaders for examples