ckanext-toolbelt 0.2.0

No project description provided

ckanext-toolbelt

Collection of different entities that are useful sometimes.

Requirements

CKAN version	Compatible?
2.8 and earlier	no
2.9	yes
master	yes

Decorators (ckanext.toolbelt.decorators)

Collector

Creates a decorator that can collect functions and return them in a dictionary. Originally designed for actions, auth functions, validators and helpers.

Can be used as decorator. Call Collector.get_collection when you need dictionary with names of helpers mapped to helper functions

helper = Collector()

@helper
def func():
	pass

###
# ITemplateHelpers
def get_helpers(self):
	return helper.get_collection()

Collector.split allows you to visually separate decorator from the method, that returns collection

action, get_actions = Collector().split()

@action
def func():
	pass

###
# IActions
def get_actions(self):
	return get_actions()

If you want your functions prefixed by the plugin name, provide this prefix as a first argument to the Collector's constructor. If particular items must remain unprefixed, you can specify what name to use, when decorating an item

validator, get_validators = Collector("toolbelt").split()

@validator
def func():
	"""I am toolbelt_func
	"""
	pass

@validator("custom_func")
def func():
	"""I am custom_func
	"""
	pass

###
# IValidators
def get_validators(self):
	return get_validators()

Cache

Cache for functions.

Cache()
def func(v):
    return v * v

By default, cache is based on:

• module, where function is defined
• name of the function
• positional arguments
• named arguments

That means that following two invocations cached separately:

func(10)
func(v=10)

Cached data stored in redis as a JSON serialized structure. In order to use different serializers, you can specify dumper and loader parameters when creating Cache instance. Any function that accepts single value and returns str or bytes can be used as a dumper. Any function that accepts str or bytes and returns unserialized value can be used as loader.

from pickle import dumps, loads

@Cache(dumper=dumps, loader=loads)
def func(v):
    return v

As mentioned before, cache key computed using module, name of the function and parameters. It can be changed by passing a function as key argument to the Cache constructor. Expected signature is key_strategy(func, *args, **kwargs).

# this function will be called only once, because cache key is based on its name.
# And name will never change. Unless you change it
@Cache(key=lambda f, *a, **k: f.__name__)
def func(v):
    return v

Cache duration(in seconds) can be configured via duration parameter of the constructor(which can be a callable that returns comuted duration).

cache = Cache(duration=3600)

@cache
def func(v):
    return v + v

---

Magic

Don't use it, really! But, in case you have to, here are some things that can temporarily solve your problems:

Application is slow, but only for some users.

Change queries for user activities. Use it if only particular users(especially ones, who follows a lot of entities) become really slow.

from ckan.toolbelt import magic
magic.conjure_fast_group_activities()

---

Plugins

---

CLI

As soon as you've installed ckanext-toolbelt, it will register ckan toolbelt route for CLI. You don't have to add toolbelt to the list of enabled plugins. But depending on the list of enabled plugins, extra subroutes will be added to the ckan toolbelt route.

In addition, there is global ctb command that allows to use this package without CKAN installed or without CKAN config file. But in this way some of commands (search-index for example) are not available, because they use CKAN core.

Global commands, available via ctb and ckan toolbelt routes:

# Print to stdout basic Makefile for ckan-deps-installer
make deps-makefile

# Start mail server that will catch outcomming mails.
dev mail-server

Commands that depends on CKAN core and available only via ckan toolbelt route:

# Drop packages that are only in search index but not in DB.
search-index clear-missing

# Clean the DB, optionally keeping data in the given tables.
db clean --yes [-k user] [-k group] [-k ...]