Instantiate parsers only once

Author: Gallaecio

parsel/parser/__init__.py

@@ -0,0 +1,36 @@
+from lxml import etree
+from lxml.etree import XMLParser as _UnsafeXMLParser
+from lxml.html import HTMLParser as _HTMLParser
+
+
+class _LXMLBaseParser(object):
+
+    def __init__(self, parser_cls):
+        self._parser = parser_cls(recover=True, encoding='utf8')
+
+    def parse(self, text, base_url):
+        body = text.strip().replace('\x00', '').encode('utf8') or b'<html/>'
+        root = etree.fromstring(body, parser=self._parser, base_url=base_url)
+        if root is None:
+            root = etree.fromstring(b'<html/>', parser=self._parser,
+                                    base_url=base_url)
+        return root
+
+
+class HTMLParser(_LXMLBaseParser):
+
+    def __init__(self):
+        super(HTMLParser, self).__init__(_HTMLParser)
+
+
+class _XMLParser(_UnsafeXMLParser):
+
+    def __init__(self, *args, **kwargs):
+        kwargs.setdefault('resolve_entities', False)
+        super(_XMLParser, self).__init__(*args, **kwargs)
+
+
+class XMLParser(_LXMLBaseParser):
+
+    def __init__(self):
+        super(XMLParser, self).__init__(_XMLParser)

parsel/parser/html.py

@@ -0,0 +1,3 @@
+from parsel.parser import HTMLParser
+
+HTML_PARSER = HTMLParser()

parsel/parser/xml.py

@@ -0,0 +1,3 @@
+from parsel.parser import XMLParser
+
+XML_PARSER = XMLParser()

parsel/selector.py

@@ -3,14 +3,39 @@
 """
 
 import sys
+from importlib import import_module
+from warnings import warn
 
 import six
-from lxml import etree, html
+from lxml import etree
 
 from .utils import flatten, iflatten, extract_regex, shorten
 from .csstranslator import HTMLTranslator, GenericTranslator
 
 
+def _load_object(path):
+    """Load an object given its absolute object path, and return it.
+
+    `path` can point to a class, function, variable or a class instance. For
+    example: ``'parsel.parser.html.HTML_PARSER'``.
+    """
+
+    try:
+        dot = path.rindex('.')
+    except ValueError:
+        raise ValueError("Error loading object '%s': not a full path" % path)
+
+    module, name = path[:dot], path[dot+1:]
+    mod = import_module(module)
+
+    try:
+        obj = getattr(mod, name)
+    except AttributeError:
+        raise NameError("Module '%s' doesn't define any object named '%s'" % (module, name))
+
+    return obj
+
+
 class CannotRemoveElementWithoutRoot(Exception):
     pass
 
@@ -21,14 +46,16 @@ class CannotRemoveElementWithoutParent(Exception):
 
 class SafeXMLParser(etree.XMLParser):
     def __init__(self, *args, **kwargs):
+        warn('parsel.selector.SafeXMLParser is deprecated',
+             DeprecationWarning, stacklevel=2)
         kwargs.setdefault('resolve_entities', False)
         super(SafeXMLParser, self).__init__(*args, **kwargs)
 
 _ctgroup = {
-    'html': {'_parser': html.HTMLParser,
+    'html': {'_parser': 'parsel.parser.html.HTML_PARSER',
              '_csstranslator': HTMLTranslator(),
              '_tostring_method': 'html'},
-    'xml': {'_parser': SafeXMLParser,
+    'xml': {'_parser': 'parsel.parser.xml.XML_PARSER',
             '_csstranslator': GenericTranslator(),
             '_tostring_method': 'xml'},
 }
@@ -46,6 +73,8 @@ def _st(st):
 def create_root_node(text, parser_cls, base_url=None):
     """Create root node for text using given parser class.
     """
+    warn('parsel.selector.create_root_node is deprecated',
+         DeprecationWarning, stacklevel=2)
     body = text.strip().replace('\x00', '').encode('utf8') or b'<html/>'
     parser = parser_cls(recover=True, encoding='utf8')
     root = etree.fromstring(body, parser=parser, base_url=base_url)
@@ -198,7 +227,7 @@ class Selector(object):
     def __init__(self, text=None, type=None, namespaces=None, root=None,
                  base_url=None, _expr=None):
         self.type = st = _st(type or self._default_type)
-        self._parser = _ctgroup[st]['_parser']
+        self._parser = _load_object(_ctgroup[st]['_parser'])
         self._csstranslator = _ctgroup[st]['_csstranslator']
         self._tostring_method = _ctgroup[st]['_tostring_method']
 
@@ -219,7 +248,7 @@ def __getstate__(self):
         raise TypeError("can't pickle Selector objects")
 
     def _get_root(self, text, base_url=None):
-        return create_root_node(text, self._parser, base_url=base_url)
+        return self._parser.parse(text=text, base_url=base_url)
 
     def xpath(self, query, namespaces=None, **kwargs):
         """

tests/test_deprecations.py

@@ -0,0 +1,21 @@
+# -*- coding:utf-8 -*-
+
+
+from unittest import TestCase
+from warnings import catch_warnings
+
+from parsel.selector import create_root_node, SafeXMLParser
+from lxml.html import HTMLParser
+
+
+class TestDeprecations(TestCase):
+
+    def test_create_root_node(self):
+        with catch_warnings(record=True) as warnings:
+            create_root_node(u'…', HTMLParser)
+            self.assertEqual(len(warnings), 1)
+
+    def test_SafeXMLParser(self):
+        with catch_warnings(record=True) as warnings:
+            parser = SafeXMLParser()
+            self.assertEqual(len(warnings), 1)

tests/test_selector.py

@@ -7,6 +7,7 @@
 
 from parsel import Selector
 from parsel.selector import (
+    _load_object,
     CannotRemoveElementWithoutRoot,
     CannotRemoveElementWithoutParent,
 )
@@ -913,3 +914,24 @@ def test_set(self):
                                //div[@itemtype="http://schema.org/Event"]
                                     //*[@itemscope]/*/@itemprop)''').extract(),
                          [u'url', u'name', u'startDate', u'location', u'offers'])
+
+
+try:
+    ModuleNotFoundError
+except NameError:
+    ModuleNotFoundError = ImportError
+
+
+class LoadObjectTestCase(unittest.TestCase):
+
+    def test_incomplete_path(self):
+        with self.assertRaises(ValueError):
+            object = _load_object('parsel')
+
+    def test_inexistent_module(self):
+        with self.assertRaises(ModuleNotFoundError):
+            object = _load_object('parsel.inexistent.inexistent')
+
+    def test_inexistent_object(self):
+        with self.assertRaises(NameError):
+            object = _load_object('parsel.parser.inexistent')