如何序列化作为 XML 导出器中项目列表的 Scrapy 字段

Question

我构建了复杂的项目，其中的字段可能是其他项目类型的列表。当我使用默认 XmlItemExporter 导出它时，子列表项以 <value> 标签为前缀。我正在寻找如何将子项标识符分配给那些值标签的示例。

文档的 Item Exporters 页面解释了这句话：

Unless overridden in the serialize_field() method, multi-valued fields are exported by serializing each value inside a <value> element. This is for convenience, as multi-valued fields are very common.

文档页面还提供了有关在现场声明序列化程序和覆盖Serialize_Field()方法的简单示例, 但两者都是针对单值字段的，没有关于如何为多值字段自定义它们的建议。

我在网上搜索了如何实现的示例，但没有找到。

这是我用于测试的示例项目树：

class Course(scrapy.Item):
    title = scrapy.Field()
    lessons = scrapy.Field()

class Lesson(scrapy.Item):
    session = scrapy.Field()
    topic = scrapy.Field()
    assignment = scrapy.Field()

class ReadingAssignment(scrapy.Item):
    textBook = scrapy.Field()
    pages = scrapy.Field()

course = Course()
course['title'] = 'Greatness'
course['lessons'] = []

lesson = Lesson()
lesson['session'] = 'Week 1'
lesson['topic'] = 'Think Great'
lesson['assignment'] = []

reading =  ReadingAssignment()
reading['textBook'] = 'Great Book 1'
reading['pages'] = '1-20'
lesson['assignment'].append(reading)
course['lessons'].append(lesson)

lesson = Lesson()
lesson['session'] = 'Week 2'
lesson['topic'] = 'Act Great'
lesson['assignment'] = []

reading =  ReadingAssignment()
reading['textBook'] = 'Great Book 2'
reading['pages'] = '21-40'
lesson['assignment'].append(reading)
course['lessons'].append(lesson)

lesson = Lesson()
lesson['session'] = 'Week 3'
lesson['topic'] = 'Look Great'
lesson['assignment'] = []

reading =  ReadingAssignment()
reading['textBook'] = 'Great Book 3'
reading['pages'] = '41-60'
lesson['assignment'].append(reading)
course['lessons'].append(lesson)

lesson = Lesson()
lesson['session'] = 'Week 4'
lesson['topic'] = 'Be Great'
lesson['assignment'] = []

reading =  ReadingAssignment()
reading['textBook'] = 'Great Book 4'
reading['pages'] = '61-80'
lesson['assignment'].append(reading)
course['lessons'].append(lesson)

输出：

>>> course
{'lessons': [{'assignment': [{'pages': '1-20', 'textBook': 'Great Book 1'}],
              'session': 'Week 1',
              'topic': 'Think Great'},
             {'assignment': [{'pages': '21-40', 'textBook': 'Great Book 2'}],
              'session': 'Week 2',
              'topic': 'Act Great'},
             {'assignment': [{'pages': '41-60', 'textBook': 'Great Book 3'}],
              'session': 'Week 3',
              'topic': 'Look Great'},
             {'assignment': [{'pages': '61-80', 'textBook': 'Great Book 4'}],
              'session': 'Week 4',
              'topic': 'Be Great'}],
 'title': 'Greatness'}

当我运行通过 XmlItemExporter 我得到：

<?xml version="1.0" encoding="utf-8"?>
<items>
  <course>
    <title>Greatness</title>
    <lessons>
      <value>
        <session>Week 1</session>
        <topic>Think Great</topic>
        <assignment>
          <value>
            <textBook>Great Book 1</textBook>
            <pages>1-20</pages>
          </value>
        </assignment>
      </value>
      <value>
        <session>Week 2</session>
        <topic>Act Great</topic>
        <assignment>
          <value>
            <textBook>Great Book 2</textBook>
            <pages>21-40</pages>
          </value>
        </assignment>
      </value>
      <value>
        <session>Week 3</session>
        <topic>Look Great</topic>
        <assignment>
          <value>
            <textBook>Great Book 3</textBook>
            <pages>41-60</pages>
          </value>
        </assignment>
      </value>
      <value>
        <session>Week 4</session>
        <topic>Be Great</topic>
        <assignment>
          <value>
            <textBook>Great Book 4</textBook>
            <pages>61-80</pages>
          </value>
        </assignment>
      </value>
    </lessons>
  </course>
</items>

我想做的是将那些 <value> 标签更改为附加到列表中的项目的名称。像这样：

<items>
  <course>
    <title>Greatness</title>
    <lessons>
      <lesson>
        <session>Week 1</session>
        <topic>Think Great</topic>
        <assignment>
          <reading>
            <textBook>Great Book 1</textBook>
            <pages>1-20</pages>
          </reading>
        </assignment>
      </lesson>
      <lesson>
        <session>Week 2</session>
        <topic>Act Great</topic>
        <assignment>
          <reading>
            <textBook>Great Book 2</textBook>
            <pages>21-40</pages>
          </reading>
        </assignment>
      </lesson>
      <lesson>
        <session>Week 3</session>
        <topic>Look Great</topic>
        <assignment>
          <reading>
            <textBook>Great Book 3</textBook>
            <pages>41-60</pages>
          </reading>
        </assignment>
      </lesson>
      <lesson>
        <session>Week 4</session>
        <topic>Be Great</topic>
        <assignment>
          <reading>
            <textBook>Great Book 4</textBook>
            <pages>61-80</pages>
          </reading>
        </assignment>
      </lesson>
    </lessons>
  </course>
</items>

Answer 1

这确实没有很好的记录，我们将不得不求助于阅读 XmlItemExporter source code, where it turns out that the <value> tag choice has been hard-coded in the XmlItemExporter._export_xml_field() method:

elif is_listlike(serialized_value):
    self._beautify_newline()
    for value in serialized_value:
        self._export_xml_field('value', value, depth=depth+1)
    self._beautify_indent(depth=depth)

幸运的是，还有出路，在之前的几行中：

if hasattr(serialized_value, 'items'):
    self._beautify_newline()
    for subname, value in serialized_value.items():
        self._export_xml_field(subname, value, depth=depth+1)
    self._beautify_indent(depth=depth)

这意味着要处理字典，但实际上它会接受任何具有 .items() 方法的任何东西，即 returns 字符串和项目的元组！

但是，导出器中缺少一个重要步骤：递归。您基本上只能在 top-level 项目字段上设置 serializer 标志，Item subclass 上的任何 Field() 元素超出 top-level 项目是当前的 Scrapy 实现完全忽略了。每个导出器在如何驱动内部 BaseItemExporter._get_serialized_fields() method 方面都有自己的特点，因此我们不能像每个特定的导出器（JSON、XML 等）一样预先处理递归。他们需要序列化字段的方式不同。我们可以使用 XmlItemExporter class 的子 class 来解决这个问题，详情见下文。

所以这里的第一个技巧是创建一个具有 .items() 方法并为您提供 <container> 标签的专用对象。请注意，您必须自己处理序列化的递归！ Scrapy 序列化器本身不处理嵌套结构的递归：

class CustomXMLValuesSerializer:
    @classmethod
    def serialize_as(cls, name):
        def serializer(items, serialize):
            return cls(name, items, serialize)
        return serializer

    def __init__(self, name, items, serialize=None):
        self._name = name
        self._items = items
        self._serialize = serialize if serialise is not None else lambda x: x

    def items(self):
        for item in self._items:
            yield (self._name, self._serialize(item))

然后使用 CustomXMLValuesSerializer.serialize_as() class 方法为您的列表字段创建自定义序列化程序：

class Course(scrapy.Item):
    title = scrapy.Field()
    lessons = scrapy.Field(
        serializer=CustomXMLValuesSerializer.serialize_as("lesson")
    )

class Lesson(scrapy.Item):
    session = scrapy.Field()
    topic = scrapy.Field()
    assignment = scrapy.Field(
        serializer=CustomXMLValuesSerializer.serialize_as("reading")
    )

class ReadingAssignment(scrapy.Item):
    textBook = scrapy.Field()
    pages = scrapy.Field()

最后，我们需要一个稍微自定义的导出器，它实际上可以让我们递归地处理嵌套项：

from functools import partial

class RecursingXmlItemExporter(XmlItemExporter):
    def _recursive_serialized_fields(self, item):
        if isinstance(item, scrapy.Item):
            return dict(self._get_serialized_fields(item, default_value=''))
        return item

    def serialize_field(self, field, name, value):
        serializer = field.get('serializer', lambda x: x)
        try:
            return serializer(value, self._recursive_serialized_fields)
        except TypeError:
            return serializer(value)

请注意，这会传入 default_value=''，因为 that's what the base XmlItemExporter.export_item() implementation uses。

确保使用此自定义导出器，因为它传入所需的上下文以序列化嵌套项：

exporter = RecursingXmlItemExporter(some_file, indent=2, item_element='course')
exporter.start_exporting()
exporter.export_item(course)
exporter.finish_exporting()

现在容器实际上是使用 name 字符串作为容器元素导出的：

<?xml version="1.0" encoding="utf-8"?>
<items>
  <course>
    <title>Greatness</title>
    <lessons>
      <lesson>
        <session>Week 1</session>
        <topic>Think Great</topic>
        <assignment>
          <reading>
            <textBook>Great Book 1</textBook>
            <pages>1-20</pages>
          </reading>
        </assignment>
      </lesson>
      <lesson>
        <session>Week 2</session>
        <topic>Act Great</topic>
        <assignment>
          <reading>
            <textBook>Great Book 2</textBook>
            <pages>21-40</pages>
          </reading>
        </assignment>
      </lesson>
      <lesson>
        <session>Week 3</session>
        <topic>Look Great</topic>
        <assignment>
          <reading>
            <textBook>Great Book 3</textBook>
            <pages>41-60</pages>
          </reading>
        </assignment>
      </lesson>
      <lesson>
        <session>Week 4</session>
        <topic>Be Great</topic>
        <assignment>
          <reading>
            <textBook>Great Book 4</textBook>
            <pages>61-80</pages>
          </reading>
        </assignment>
      </lesson>
    </lessons>
  </course>
</items>

我用 Scrapy 字段 issue #3888 看看项目是否有兴趣更好地支持嵌套 Item 结构。

另一种方法是通过单独调用 XmlItemExporter.export_item() 方法来导出嵌套项，但这要求导出器可以在与序列化器相同的命名空间中作为全局访问，或者您子class 导出器并...将导出器传递给序列化程序。然后你必须满足于 XmlItemExporter.export_item() hard-codes 缩进的事实。

如何序列化作为 XML 导出器中项目列表的 Scrapy 字段

How to Serialize Scrapy Fields that are Lists of Items in XML Exporter

python

xml

serialization

exporter

scrapy