abx

Module Contents

Classes

HookSpecDecoratorThatReturnsFirstResult	Type of a plugin method decorated with @hookspec(firstresult=True), which returns a single result (from the first plugin that implements the hook)
HookSpecDecoratorThatReturnsListResults	Type of a plugin method decorated with @hookspec(firstresult=False), which returns a list of results (one for each plugin that implements the hook)
TypedHookspecMarker	Improved version of pluggy.HookspecMarker that supports type inference of hookspecs with firstresult=True|False correctly https://github.com/pytest-dev/pluggy/issues/191
PluginInfo	Full Metadata Dictionary containing all info about a plugin, returned by abx.get_plugin()
ABXPluginManager	Patch to fix pluggy’s PluginManager to work with pydantic models. See: https://github.com/pytest-dev/pluggy/pull/536

Functions

is_valid_attr_name	Check if a string is a valid attribute name (used to validate hook method names on a plugin)
is_valid_module_name	Check if a string e.g. “some_pkg.some_plugin_name” is a valid module name (used to validate plugin IDs)
get_plugin_order	Get the order a plugin should be loaded in by reading its ./.plugin_order file or .order attr
get_plugin	Get the full PluginInfo metadata for a plugin, given its plugin ID, module, or class
get_all_plugins	Get the PluginInfo metadata for all the loaded plugins
get_all_hook_names	Get the names of all hookspec/hookimpl methods available across all loaded plugins
get_all_hook_specs	Get a set of all hookspec methods defined in all plugins (useful for type checking if a pm.hook.call() is valid)
find_plugins_in_dir	Find all the plugins in a given directory. Just looks for an init.py file.
get_pip_installed_plugins	replaces pm.load_setuptools_entrypoints(“abx”), finds plugins that registered entrypoints via pip
load_plugins	Load all the plugins from a dictionary of module names and directory paths.
get_plugin_hooks	Get all the functions marked with @hookimpl on a plugin module or class.
as_list	Flatten a list of lists returned by a pm.hook.call() into a single list of [result1, result2, …]
as_dict	Flatten a list of dicts returned by a pm.hook.call() into a single dict of {plugin_id1: result1, plugin_id2: result2, …}

Data

__id__
__label__
__author__
__homepage__
__order__
ParamsT
ReturnT
AttrName
PluginId
PluginSpec
pm

API

abx.__id__

‘abx’

abx.__label__

‘ABX’

abx.__author__

‘Nick Sweeting’

abx.__homepage__

‘https://github.com/ArchiveBox/ArchiveBox’

abx.__order__

0

abx.ParamsT

‘ParamSpec(…)’

abx.ReturnT

‘TypeVar(…)’

class abx.HookSpecDecoratorThatReturnsFirstResult

Bases: typing.Protocol

Type of a plugin method decorated with @hookspec(firstresult=True), which returns a single result (from the first plugin that implements the hook)

__call__(func: Callable[abx.ParamsT, abx.ReturnT]) → Callable[abx.ParamsT, abx.ReturnT]

class abx.HookSpecDecoratorThatReturnsListResults

Bases: typing.Protocol

Type of a plugin method decorated with @hookspec(firstresult=False), which returns a list of results (one for each plugin that implements the hook)

__call__(func: Callable[abx.ParamsT, abx.ReturnT]) → Callable[abx.ParamsT, List[abx.ReturnT]]

class abx.TypedHookspecMarker(project_name: str)

Improved version of pluggy.HookspecMarker that supports type inference of hookspecs with firstresult=True|False correctly https://github.com/pytest-dev/pluggy/issues/191

Initialization

__slots__

(‘project_name’,)

__call__(function: Callable[abx.ParamsT, abx.ReturnT] | None = None, firstresult: bool = False, historic: bool = False, warn_on_impl: Warning | None = None, warn_on_impl_args: Mapping[str, Warning] | None = None) → Callable[abx.ParamsT, List[abx.ReturnT]] | abx.HookSpecDecoratorThatReturnsListResults | abx.HookSpecDecoratorThatReturnsFirstResult

abx.is_valid_attr_name(x: str) → str

Check if a string is a valid attribute name (used to validate hook method names on a plugin)

abx.is_valid_module_name(x: str) → str

Check if a string e.g. “some_pkg.some_plugin_name” is a valid module name (used to validate plugin IDs)

abx.AttrName

None

abx.PluginId

None

class abx.PluginInfo

Bases: typing.TypedDict

Full Metadata Dictionary containing all info about a plugin, returned by abx.get_plugin()

Initialization

Initialize self. See help(type(self)) for accurate signature.

id: abx.PluginId

None

package: abx.AttrName

None

label: str

None

version: str

None

author: str

None

homepage: str

None

dependencies: List[str]

None

source_code: str

None

hooks: Dict[abx.AttrName, Callable]

None

module: types.ModuleType

None

abx.PluginSpec

‘TypeVar(…)’

class abx.ABXPluginManager

Bases: pluggy.PluginManager, typing.Generic[abx.PluginSpec]

Patch to fix pluggy’s PluginManager to work with pydantic models. See: https://github.com/pytest-dev/pluggy/pull/536

hook: abx.PluginSpec

None

create_typed_hookcaller(name: str, module_or_class: Type[abx.PluginSpec], spec_opts: pluggy.HookspecOpts) → pluggy.HookCaller

create a new HookCaller subclass with a modified signature so that the return type is correct and args are converted to kwargs

add_hookspecs(module_or_class: Type[abx.PluginSpec]) → None

Add HookSpecs from the given class, (generic type allows us to enforce types of pm.hook.call() statically)

parse_hookimpl_opts(plugin, name: str) → pluggy.HookimplOpts | None

abx.pm

‘ABXPluginManager(…)’

abx.get_plugin_order(plugin: abx.PluginId | pathlib.Path | types.ModuleType | Type) → Tuple[int, pathlib.Path]

Get the order a plugin should be loaded in by reading its ./.plugin_order file or .order attr

abx.get_plugin(plugin: abx.PluginId | types.ModuleType | Type) → abx.PluginInfo

Get the full PluginInfo metadata for a plugin, given its plugin ID, module, or class

abx.get_all_plugins() → Dict[abx.PluginId, abx.PluginInfo]

Get the PluginInfo metadata for all the loaded plugins

abx.get_all_hook_names() → Set[str]

Get the names of all hookspec/hookimpl methods available across all loaded plugins

abx.get_all_hook_specs() → Dict[str, Dict[str, Any]]

Get a set of all hookspec methods defined in all plugins (useful for type checking if a pm.hook.call() is valid)

abx.find_plugins_in_dir(plugins_dir: pathlib.Path) → Dict[abx.PluginId, pathlib.Path]

Find all the plugins in a given directory. Just looks for an init.py file.

abx.get_pip_installed_plugins(group: abx.PluginId = 'abx') → Dict[abx.PluginId, pathlib.Path]

replaces pm.load_setuptools_entrypoints(“abx”), finds plugins that registered entrypoints via pip

abx.load_plugins(plugins: Iterable[abx.PluginId | types.ModuleType | Type] | Dict[abx.PluginId, pathlib.Path])

Load all the plugins from a dictionary of module names and directory paths.

abx.get_plugin_hooks(plugin: abx.PluginId | types.ModuleType | Type | None) → Dict[abx.AttrName, Callable]

Get all the functions marked with @hookimpl on a plugin module or class.

abx.as_list(results: List[List[abx.ReturnT]]) → List[abx.ReturnT]

Flatten a list of lists returned by a pm.hook.call() into a single list of [result1, result2, …]

abx.as_dict(results: List[Dict[abx.PluginId, abx.ReturnT]]) → Dict[abx.PluginId, abx.ReturnT]

Flatten a list of dicts returned by a pm.hook.call() into a single dict of {plugin_id1: result1, plugin_id2: result2, …}