Item Loaders provide a convenient mechanism for populating scraped Items. Even though Items can be populated using their owndictionary-like API, Item Loaders provide a much more convenient API forpopulating them from a scraping process, by automating some common tasks likeparsing the raw extracted data before assigning it.

In other words, Items provide the container ofscraped data, while Item Loaders provide the mechanism for populating thatcontainer.

Item Loaders are designed to provide a flexible, efficient and easy mechanismfor extending and overriding different field parsing rules, either by spider,or by source format (HTML, XML, etc) without becoming a nightmare to maintain.

## Using Item Loaders to populate items

To use an Item Loader, you must first instantiate it. You can eitherinstantiate it with a dict-like object (e.g. Item or dict) or without one, inwhich case an Item is automatically instantiated in the Item Loader init methodusing the Item class specified in the ItemLoader.default_item_classattribute.

Then, you start collecting values into the Item Loader, typically usingSelectors. You can add more than one value tothe same item field; the Item Loader will know how to “join” those values laterusing a proper processing function.

Note

Collected data is internally stored as lists,allowing to add several values to the same field.If an item argument is passed when creating a loader,each of the item’s values will be stored as-is if it’s alreadyan iterable, or wrapped with a list if it’s a single value.

Here is a typical Item Loader usage in a Spider, usingthe Product item declared in the Itemschapter:

from scrapy.loader import ItemLoaderfrom myproject.items import Product def parse(self, response):    l = ItemLoader(item=Product(), response=response)    l.add_xpath('name', '//div[@class="product_name"]')    l.add_xpath('name', '//div[@class="product_title"]')    l.add_xpath('price', '//p[@id="price"]')    l.add_css('stock', 'p#stock]')    l.add_value('last_updated', 'today') # you can also use literal values    return l.load_item()

By quickly looking at that code, we can see the name field is beingextracted from two different XPath locations in the page:

• //div[@class="product_name"]
• //div[@class="product_title"]In other words, data is being collected by extracting it from two XPathlocations, using the add_xpath() method. This is thedata that will be assigned to the name field later.

Afterwards, similar calls are used for price and stock fields(the latter using a CSS selector with the add_css() method),and finally the last_update field is populated directly with a literal value(today) using a different method: add_value().

Finally, when all data is collected, the ItemLoader.load_item() method iscalled which actually returns the item populated with the datapreviously extracted and collected with the add_xpath(),add_css(), and add_value() calls.

## Input and Output processors

An Item Loader contains one input processor and one output processor for each(item) field. The input processor processes the extracted data as soon as it’sreceived (through the add_xpath(), add_css() oradd_value() methods) and the result of the input processor iscollected and kept inside the ItemLoader. After collecting all data, theItemLoader.load_item() method is called to populate and get the populatedItem object. That’s when the output processor iscalled with the data previously collected (and processed using the inputprocessor). The result of the output processor is the final value that getsassigned to the item.

Let’s see an example to illustrate how the input and output processors arecalled for a particular field (the same applies for any other field):

l = ItemLoader(Product(), some_selector)l.add_xpath('name', xpath1) # (1)l.add_xpath('name', xpath2) # (2)l.add_css('name', css) # (3)l.add_value('name', 'test') # (4)return l.load_item() # (5)

So what happens is:

• Data from xpath1 is extracted, and passed through the input processor ofthe name field. The result of the input processor is collected and kept inthe Item Loader (but not yet assigned to the item).
• Data from xpath2 is extracted, and passed through the same inputprocessor used in (1). The result of the input processor is appended to thedata collected in (1) (if any).
• This case is similar to the previous ones, except that the data is extractedfrom the css CSS selector, and passed through the same inputprocessor used in (1) and (2). The result of the input processor is appended to thedata collected in (1) and (2) (if any).
• This case is also similar to the previous ones, except that the value to becollected is assigned directly, instead of being extracted from a XPathexpression or a CSS selector.However, the value is still passed through the input processors. In thiscase, since the value is not iterable it is converted to an iterable of asingle element before passing it to the input processor, because inputprocessor always receive iterables.
• The data collected in steps (1), (2), (3) and (4) is passed throughthe output processor of the name field.The result of the output processor is the value assigned to the namefield in the item.It’s worth noticing that processors are just callable objects, which are calledwith the data to be parsed, and return a parsed value. So you can use anyfunction as input or output processor. The only requirement is that they mustaccept one (and only one) positional argument, which will be an iterable.

Changed in version 2.0: Processors no longer need to be methods.

Note

Both input and output processors must receive an iterable as theirfirst argument. The output of those functions can be anything. The result ofinput processors will be appended to an internal list (in the Loader)containing the collected values (for that field). The result of the outputprocessors is the value that will be finally assigned to the item.

The other thing you need to keep in mind is that the values returned by inputprocessors are collected internally (in lists) and then passed to outputprocessors to populate the fields.

Last, but not least, Scrapy comes with some commonly used processors built-in for convenience.

Item Loaders are declared like Items, by using a class definition syntax. Hereis an example:

from scrapy.loader import ItemLoaderfrom scrapy.loader.processors import TakeFirst, MapCompose, Join class ProductLoader(ItemLoader):     default_output_processor = TakeFirst()     name_in = MapCompose(unicode.title)    name_out = Join()     price_in = MapCompose(unicode.strip)     # ...

As you can see, input processors are declared using the _in suffix whileoutput processors are declared using the _out suffix. And you can alsodeclare a default input/output processors using theItemLoader.default_input_processor andItemLoader.default_output_processor attributes.

## Declaring Input and Output Processors

As seen in the previous section, input and output processors can be declared inthe Item Loader definition, and it’s very common to declare input processorsthis way. However, there is one more place where you can specify the input andoutput processors to use: in the Item Fieldmetadata. Here is an example:

import scrapyfrom scrapy.loader.processors import Join, MapCompose, TakeFirstfrom w3lib.html import remove_tags def filter_price(value):    if value.isdigit():        return value class Product(scrapy.Item):    name = scrapy.Field(        input_processor=MapCompose(remove_tags),        output_processor=Join(),    )    price = scrapy.Field(        input_processor=MapCompose(remove_tags, filter_price),        output_processor=TakeFirst(),    )
>>> from scrapy.loader import ItemLoader>>> il = ItemLoader(item=Product())>>> il.add_value('name', [u'Welcome to my', u'<strong>website</strong>'])>>> il.add_value('price', [u'&euro;', u'<span>1000</span>'])>>> il.load_item(){'name': u'Welcome to my website', 'price': u'1000'}

The precedence order, for both input and output processors, is as follows:

The Item Loader Context is a dict of arbitrary key/values which is shared amongall input and output processors in the Item Loader. It can be passed whendeclaring, instantiating or using Item Loader. They are used to modify thebehaviour of the input/output processors.

For example, suppose you have a function parse_length which receives a textvalue and extracts a length from it:

def parse_length(text, loader_context):    unit = loader_context.get('unit', 'm')    # ... length parsing code goes here ...    return parsed_length

By accepting a loader_context argument the function is explicitly tellingthe Item Loader that it’s able to receive an Item Loader context, so the ItemLoader passes the currently active context when calling it, and the processorfunction (parse_length in this case) can thus use them.

There are several ways to modify Item Loader context values:

loader = ItemLoader(product)loader.context['unit'] = 'cm'
• On Item Loader instantiation (the keyword arguments of Item Loaderinit method are stored in the Item Loader context):
loader = ItemLoader(product, unit='cm')
• On Item Loader declaration, for those input/output processors that supportinstantiating them with an Item Loader context. MapCompose is one ofthem:
class ProductLoader(ItemLoader):    length_out = MapCompose(parse_length, unit='cm')

• class scrapy.loader.ItemLoader([item, selector, response, ]**kwargs)[source]
• Return a new Item Loader for populating the given Item. If no item isgiven, one is instantiated automatically using the class indefault_item_class.

When instantiated with a selector or a response parametersthe ItemLoader class provides convenient mechanisms for extractingdata from web pages using selectors.

Parameters:

The item, selector, response and the remaining keyword arguments areassigned to the Loader context (accessible through the context attribute).

ItemLoader instances have the following methods:

• getvalue(_value, *processors, **kwargs)[source]
• Process the given value by the given processors and keywordarguments.

Available keyword arguments:

Parameters:re (str or __compiled regex) – a regular expression to use for extracting data from thegiven value using extract_regex() method,applied before processors

Examples:

>>> from scrapy.loader.processors import TakeFirst>>> loader.get_value(u'name: foo', TakeFirst(), unicode.upper, re='name: (.+)')'FOO
• addvalue(_field_name, value, *processors, **kwargs)[source]
• Process and then add the given value for the given field.

The value is first passed through get_value() by giving theprocessors and kwargs, and then passed through thefield input processor and its resultappended to the data collected for that field. If the field alreadycontains collected data, the new data is added.

The given field_name can be None, in which case values formultiple fields may be added. And the processed value should be a dictwith field_name mapped to values.

Examples:

loader.add_value('name', u'Color TV')loader.add_value('colours', [u'white', u'blue'])loader.add_value('length', u'100')loader.add_value('name', u'name: foo', TakeFirst(), re='name: (.+)')loader.add_value(None, {'name': u'foo', 'sex': u'male'})

Parameters:

- **xpath** ([_str_](https://docs.python.org/3/library/stdtypes.html#str)) – the XPath to extract data from- **re** ([_str_](https://docs.python.org/3/library/stdtypes.html#str)_ or __compiled regex_) – a regular expression to use for extracting data from theselected XPath region

Examples:

# HTML snippet: <p class="product-name">Color TV</p>loader.get_xpath('//p[@class="product-name"]')# HTML snippet: <p id="price">the price is $1200</p>loader.get_xpath('//p[@id="price"]', TakeFirst(), re='the price is (.*)') See get_xpath() for kwargs. Parameters:xpath (str) – the XPath to extract data from Examples: # HTML snippet: <p class="product-name">Color TV</p>loader.add_xpath('name', '//p[@class="product-name"]')# HTML snippet: <p id="price">the price is$1200</p>loader.add_xpath('price', '//p[@id="price"]', re='the price is (.*)')

Parameters:

- **css** ([_str_](https://docs.python.org/3/library/stdtypes.html#str)) – the CSS selector to extract data from- **re** ([_str_](https://docs.python.org/3/library/stdtypes.html#str)_ or __compiled regex_) – a regular expression to use for extracting data from theselected CSS region

Examples:

# HTML snippet: <p class="product-name">Color TV</p>loader.get_css('p.product-name')# HTML snippet: <p id="price">the price is $1200</p>loader.get_css('p#price', TakeFirst(), re='the price is (.*)') See get_css() for kwargs. Parameters:css (str) – the CSS selector to extract data from Examples: # HTML snippet: <p class="product-name">Color TV</p>loader.add_css('name', 'p.product-name')# HTML snippet: <p id="price">the price is$1200</p>loader.add_css('price', 'p#price', re='the price is (.*)')

ItemLoader instances have the following attributes:

• item
• The Item object being parsed by this Item Loader.This is mostly used as a property so when attempting to override thisvalue, you may want to check out default_item_class first.

• context

• The currently active Context of thisItem Loader.

• default_item_class

• An Item class (or factory), used to instantiate items when not given inthe init method.

• default_input_processor

• The default input processor to use for those fields which don’t specifyone.

• default_output_processor

• The default output processor to use for those fields which don’t specifyone.

• default_selector_class

• The class used to construct the selector of thisItemLoader, if only a response is given in the init method.If a selector is given in the init method this attribute is ignored.This attribute is sometimes overridden in subclasses.

• selector

• The Selector object to extract data from.It’s either the selector given in the init method or one created fromthe response given in the init method using thedefault_selector_class. This attribute is meant to beread-only.

When parsing related values from a subsection of a document, it can beuseful to create nested loaders. Imagine you’re extracting details froma footer of a page that looks something like:

Example:

<footer>    <a class="social" href="https://facebook.com/whatever">Like Us</a>    <a class="social" href="https://twitter.com/whatever">Follow Us</a>    <a class="email" href="mailto:[email protected]">Email Us</a></footer>

Without nested loaders, you need to specify the full xpath (or css) for each valuethat you wish to extract.

Example:

loader = ItemLoader(item=Item())# load stuff not in the footerloader.add_xpath('social', '//footer/a[@class = "social"]/@href')loader.add_xpath('email', '//footer/a[@class = "email"]/@href')loader.load_item()

Instead, you can create a nested loader with the footer selector and add valuesrelative to the footer. The functionality is the same but you avoid repeatingthe footer selector.

Example:

loader = ItemLoader(item=Item())# load stuff not in the footerfooter_loader = loader.nested_xpath('//footer')footer_loader.add_xpath('social', 'a[@class = "social"]/@href')footer_loader.add_xpath('email', 'a[@class = "email"]/@href')# no need to call footer_loader.load_item()loader.load_item()

You can nest loaders arbitrarily and they work with either xpath or css selectors.As a general guideline, use nested loaders when they make your code simpler but donot go overboard with nesting or your parser can become difficult to read.

## Reusing and extending Item Loaders

As your project grows bigger and acquires more and more spiders, maintenancebecomes a fundamental problem, especially when you have to deal with manydifferent parsing rules for each spider, having a lot of exceptions, but alsowanting to reuse the common processors.

Item Loaders are designed to ease the maintenance burden of parsing rules,without losing flexibility and, at the same time, providing a convenientmechanism for extending and overriding them. For this reason Item Loaderssupport traditional Python class inheritance for dealing with differences ofspecific spiders (or groups of spiders).

Suppose, for example, that some particular site encloses their product names inthree dashes (e.g. —-Plasma TV—-) and you don’t want to end up scrapingthose dashes in the final product names.

Here’s how you can remove those dashes by reusing and extending the defaultProduct Item Loader (ProductLoader):

from scrapy.loader.processors import MapComposefrom myproject.ItemLoaders import ProductLoader def strip_dashes(x):    return x.strip('-') class SiteSpecificLoader(ProductLoader):    name_in = MapCompose(strip_dashes, ProductLoader.name_in)

Another case where extending Item Loaders can be very helpful is when you havemultiple source formats, for example XML and HTML. In the XML version you maywant to remove CDATA occurrences. Here’s an example of how to do it:

from scrapy.loader.processors import MapComposefrom myproject.ItemLoaders import ProductLoaderfrom myproject.utils.xml import remove_cdata class XmlProductLoader(ProductLoader):    name_in = MapCompose(remove_cdata, ProductLoader.name_in)

And that’s how you typically extend input processors.

As for output processors, it is more common to declare them in the field metadata,as they usually depend only on the field and not on each specific site parsingrule (as input processors do). See also:Declaring Input and Output Processors.

There are many other possible ways to extend, inherit and override your ItemLoaders, and different Item Loaders hierarchies may fit better for differentprojects. Scrapy only provides the mechanism; it doesn’t impose any specificorganization of your Loaders collection - that’s up to you and your project’sneeds.

## Available built-in processors

Even though you can use any callable function as input and output processors,Scrapy provides some commonly used processors, which are described below. Someof them, like the MapCompose (which is typically used as inputprocessor) compose the output of several functions executed in order, toproduce the final parsed value.

Here is a list of all built-in processors:

• class scrapy.loader.processors.Identity[source]
• The simplest processor, which doesn’t do anything. It returns the originalvalues unchanged. It doesn’t receive any init method arguments, nor does itaccept Loader contexts.

Example:

>>> from scrapy.loader.processors import Identity>>> proc = Identity()>>> proc(['one', 'two', 'three'])['one', 'two', 'three']
• class scrapy.loader.processors.TakeFirst[source]
• Returns the first non-null/non-empty value from the values received,so it’s typically used as an output processor to single-valued fields.It doesn’t receive any init method arguments, nor does it accept Loader contexts.

Example:

>>> from scrapy.loader.processors import TakeFirst>>> proc = TakeFirst()>>> proc(['', 'one', 'two', 'three'])'one'
• class scrapy.loader.processors.Join(separator=u' ')[source]
• Returns the values joined with the separator given in the init method, whichdefaults to u' '. It doesn’t accept Loader contexts.

When using the default separator, this processor is equivalent to thefunction: u' '.join

Examples:

>>> from scrapy.loader.processors import Join>>> proc = Join()>>> proc(['one', 'two', 'three'])'one two three'>>> proc = Join('<br>')>>> proc(['one', 'two', 'three'])'one<br>two<br>three'
• class scrapy.loader.processors.Compose(*functions, **default_loader_context)[source]
• A processor which is constructed from the composition of the givenfunctions. This means that each input value of this processor is passed tothe first function, and the result of that function is passed to the secondfunction, and so on, until the last function returns the output value ofthis processor.

By default, stop process on None value. This behaviour can be changed bypassing keyword argument stop_on_none=False.

Example:

>>> from scrapy.loader.processors import Compose>>> proc = Compose(lambda v: v[0], str.upper)>>> proc(['hello', 'world'])'HELLO'

Each function can optionally receive a loader_context parameter. Forthose which do, this processor will pass the currently active Loadercontext through that parameter.

The keyword arguments passed in the init method are used as the defaultLoader context values passed to each function call. However, the finalLoader context values passed to functions are overridden with the currentlyactive Loader context accessible through the ItemLoader.context()attribute.

• class scrapy.loader.processors.MapCompose(*functions, **default_loader_context)[source]
• A processor which is constructed from the composition of the givenfunctions, similar to the Compose processor. The difference withthis processor is the way internal results are passed among functions,which is as follows:

The input value of this processor is iterated and the first function isapplied to each element. The results of these function calls (one for each element)are concatenated to construct a new iterable, which is then used to apply thesecond function, and so on, until the last function is applied to eachvalue of the list of values collected so far. The output values of the lastfunction are concatenated together to produce the output of this processor.

Each particular function can return a value or a list of values, which isflattened with the list of values returned by the same function applied tothe other input values. The functions can also return None in whichcase the output of that function is ignored for further processing over thechain.

This processor provides a convenient way to compose functions that onlywork with single values (instead of iterables). For this reason theMapCompose processor is typically used as input processor, sincedata is often extracted using theextract() method of selectors, which returns a list of unicode strings.

The example below should clarify how it works:

>>> def filter_world(x):...     return None if x == 'world' else x...>>> from scrapy.loader.processors import MapCompose>>> proc = MapCompose(filter_world, str.upper)>>> proc(['hello', 'world', 'this', 'is', 'scrapy'])['HELLO, 'THIS', 'IS', 'SCRAPY']

As with the Compose processor, functions can receive Loader contexts, andinit method keyword arguments are used as default context values. SeeCompose processor for more info.

• class scrapy.loader.processors.SelectJmes(json_path)[source]
• Queries the value using the json path provided to the init method and returns the output.Requires jmespath (https://github.com/jmespath/jmespath.py) to run.This processor takes only one input at a time.

Example:

>>> from scrapy.loader.processors import SelectJmes, Compose, MapCompose>>> proc = SelectJmes("foo") #for direct use on lists and dictionaries>>> proc({'foo': 'bar'})'bar'>>> proc({'foo': {'bar': 'baz'}}){'bar': 'baz'}

Working with Json:

>>> import json>>> proc_single_json_str = Compose(json.loads, SelectJmes("foo"))>>> proc_single_json_str('{"foo": "bar"}')'bar'>>> proc_json_list = Compose(json.loads, MapCompose(SelectJmes('foo')))>>> proc_json_list('[{"foo":"bar"}, {"baz":"tar"}]')['bar']`