Python module to add support for ORM-style filtering to any list of items
QueryableList allows you to “filter” a list of items of varying types, simplifing code by replacing tedious for-loops with simple chaining.
It uses an interface common to some ORMs like Django, Flask, and IndexedRedis.
You can perform single filters on lists of data, or you can build queries and execute that query on any number of arbitrary data sets.
QueryableList also implements the boolean logic operators for lists (AND, OR, XOR) which can simplify your code.
What pattern does it replace?
Constant loops/getters to drill down data. If you are filtering data, displaying data by criteria, etc, your code will be FULL of these.
QueryableList simplifies and makes generic this common pattern of filtering.
def getOlderThan(people, minAge):
ret = 
for person in people:
if person.age and person.age > minAge:ret.append(person)
people = getAllPeople() # Get your data here
oldEnoughToRide = getOlderThan(people, 13)
notOldEnough = [person for person in people if person not in oldEnoughToRide]
people = QueryableListObjs( getAllPeople() ) # Transform data into QueryableList
oldEnoughToRide = people.filter(age__gt=13)
notOldEnough = people ^ oldEnoughToRide # The XOR of the filtered list to the parent is the NOT of the filter criteria
No function, no loop, and list comprehensions can get very messy or impossible with a large number of complicated filters applied.
The above example shows a one-time filtering of a list. You can also build reusable queries, and append different criteria based on conditions or through passing the query around different functions. See “Building Reusable Queries” section below for more info.
Perform one-time filters through one of the list-type extending classes:
QueryableListObjs - This assumes each item extends object [or implements __getattribute__].
QueryableListDicts - This assumes that each item is a dict [or implements __getitem__].
QueryableListMixed - QueryableList which can contain dict-like items or object-like item. (This is somewhat slower than using QueryableListObjs or QueryableListDicts directly, but use it if you need to mix, or need to support either type.)
The items within these lists do not need to be of the same type. If any fields are missing on the filtered objects, it will be assigned a value of “None” for filtering purposes.
You can filter the data within these objects through one of the following methods:
filterAnd - returns a QueryableList where each item matches ALL of the provided criteria.
filter - Alias for filterAnd
filterOr - returns a QueryableList where each item matches ANY of the provided criteria.
customFilter - Takes a lambda or a function as a parameter. Each element in the list is passed into this function, and if it returns True, that element is retained.
The QueryableList types support all the operations of a list, and return the same QueryableList types so you can perform chaining.
Additionally, you can use ADD(+), SUB(-), AND(&), OR(|), and XOR(^) operators against other QueryableLists as another powerful means of filtering.
You specify the filter operations by passing arguments of $fieldName__$operation.
Example: e.x. results = objs.filter(name__ne=’Tim’) # get all objects where the ‘name’ field does not equal ‘Tim’
For all available operations, see the “Operations” section below.
QueryableList collections have several other methods to make them as closely api-compatible with server-side filtering ORMs as possible.
This allows you to use the same functions regardless of if you are server-side or client-side filtering.
You can build a reusable query, out of several chains of filters (either AND or OR) by using the QueryBuilder class.
The QueryBuilder class stores a “chain” of filters, which are applied in order. Each link in the chain contains a filter type (AND or OR), and the filters themselves (same as the filter methods on the QueryableList).
Use the addFilter(filterType, ..filters..) method to add a link to the chain.
To execute the query, call execute(lst) , where “lst” is your list of items. You can execute a query multiple times on any number of datasets.
Use the copy method to create a copy of the current set of filters.
If you know the type in advance, you can pass a QueryableListObjs or QueryableListDicts when calling execute to slightly speed up access times, otherwise a QueryableListMixed (supports both dict and object style access) will be used.
myQuery = QueryBuilder()
myQuery.addFilter(age__gt=21) # Age must be greater than 21
myQuery.addFilter(‘OR’, job__ieq=’Manager’, numSubordinates__gt=0) # Is a manager, or has more than 0 subordinates
managerPartyCompany1 = myQuery.execute(company1Persons) # Filter from all company1Persons those that meet above criteria
managerPartyCompany2 = myQuery.execute(company2Persons) # use same filter to apply same query to company2Persons
One of the powerful aspects of QueryableList is that it is easily extendable.
Generally, you won’t need to do this, as QueryableListDicts or QueryableListObjs will handle your needs.
But sometimes, you have more advanced requirements than can be satisfied by properties and matching on them.
For these cases, you can extend QueryableList.QueryableListBase to create your own QueryableList type.
You only need to implement a single method,
def _get_item_value(item, fieldName)
“item” will be an item in your collection, and “fieldName” is the field being queried.
For example, say you have a series of objects, “Job”, which contain some attributes and a “queue”.
You want to be able to filter on both the attributes on the object and various special attributes of it’s queue (like size, item ids, etc).
You can implement like this:
def _get_item_value(item, fieldName):
if fieldName == ‘queueSize’:
# queueSize is the number of items in the queue
elif fieldName == ‘queueItemIds’:
# queueItemIds is a list of the ids in the item queue,
# so a “contains” query can check if an id is in this item’s queue
return [qi.id for qi in item.queue]
elif hasattr(item, fieldName):
# Otherwise, if this is an attribute on the item, return it’s value
if fieldName == ‘queue’:raise KeyError(‘Cannot query queue directly. Try queueSize or queueItemIds.’)
return getattr(item, fieldName)
else:raise KeyError(‘Invalid attribute “%s” on %s’ %(fieldName, item.__class__.__name__))
The init method takes a list of items (and it contains all the methods a list has, like .append), so you can create it like:
myJob1 = MyQueue(…)
myJob2 = MyQueue(…)
myJobs = MyQueueCollection([myJob1, myJob2])
and use it like:
largeJobs = myJobs.filter(queueSize__gt=10)
So just by implementing that one method, you now have all the powerful filter capabilities that QueryableList provides!
Pydoc documentation can be found at: http://htmlpreview.github.io/?https://github.com/kata198/QueryableList/blob/master/doc/QueryableList.html?vers=4