selectolax 0.3.28__cp312-cp312-win32.whl → 0.3.34__cp312-cp312-win32.whl

selectolax/__init__.py

@@ -2,9 +2,7 @@
 
 
 __author__ = """Artem Golubin"""
-__email__ = 'me@rushter.com'
-__version__ = '0.3.28'
+__email__ = "me@rushter.com"
+__version__ = "0.3.34"
 
-from . import parser
-from . import lexbor
-from . import modest
+from . import lexbor, modest, parser

selectolax/lexbor/attrs.pxi

@@ -1,5 +1,6 @@
 cimport cython
 
+
 @cython.final
 cdef class LexborAttributes:
     """A dict-like object that represents attributes."""
@@ -23,16 +24,32 @@ cdef class LexborAttributes:
                 yield key.decode(_ENCODING)
             attr = attr.next
 
-    def __setitem__(self, str key, value):
-        value = str(value)
+    def __setitem__(self, str key, object value):
+        value = value
         bytes_key = key.encode(_ENCODING)
-        bytes_value = value.encode(_ENCODING)
-
-        lxb_dom_element_set_attribute(
-            <lxb_dom_element_t *> self.node,
-            <lxb_char_t *> bytes_key, len(bytes_key),
-            <lxb_char_t *> bytes_value, len(bytes_value),
-        )
+        bytes_value = value.encode(_ENCODING) if value else b""
+        cdef lxb_dom_attr_t *attr
+        cdef lxb_dom_document_t *doc
+
+        if value is None:
+            # N.B. This is suboptimal, but there is not API to set empty attributes
+            attr = lxb_dom_element_set_attribute(
+                <lxb_dom_element_t *> self.node,
+                <lxb_char_t *> bytes_key, len(bytes_key),
+                NULL, 0
+            )
+            doc = (<lxb_dom_node_t*>attr).owner_document
+            lexbor_str_destroy(attr.value, doc.text, 0)
+            attr.value = NULL
+
+        elif isinstance(value, str) or isinstance(value, unicode) :
+            lxb_dom_element_set_attribute(
+                <lxb_dom_element_t *> self.node,
+                <lxb_char_t *> bytes_key, len(bytes_key),
+                <lxb_char_t *> bytes_value, len(bytes_value),
+            )
+        else:
+            raise TypeError("Expected str or unicode, got %s" % type(value))
 
     def __delitem__(self, key):
         try:

selectolax/lexbor/node.pxi

@@ -1,4 +1,5 @@
 cimport cython
+from cpython.exc cimport PyErr_SetNone
 
 _TAG_TO_NAME = {
     0x0005: "- doctype",
@@ -6,26 +7,29 @@ _TAG_TO_NAME = {
     0x0004: "-comment",
 }
 ctypedef fused str_or_LexborNode:
-    basestring
+    str
     bytes
     LexborNode
 
 cdef inline bytes to_bytes(str_or_LexborNode value):
     cdef bytes bytes_val
-    if isinstance(value, (str, unicode)):
-        bytes_val = value.encode(_ENCODING)
+    if isinstance(value, unicode):
+        bytes_val = <bytes>value.encode("utf-8")
     elif isinstance(value, bytes):
-        bytes_val =  <char*> value
+        bytes_val = <bytes>value
     return bytes_val
 
+
 @cython.final
 cdef class LexborNode:
     """A class that represents HTML node (element)."""
 
-    cdef _cinit(self, lxb_dom_node_t *node, LexborHTMLParser parser):
-        self.parser = parser
-        self.node = node
-        return self
+    @staticmethod
+    cdef LexborNode new(lxb_dom_node_t *node, LexborHTMLParser parser):
+        cdef LexborNode lxbnode = LexborNode.__new__(LexborNode)
+        lxbnode.node = node
+        lxbnode.parser = parser
+        return lxbnode
 
     @property
     def mem_id(self):
@@ -41,8 +45,7 @@ cdef class LexborNode:
         """Return the first child node."""
         cdef LexborNode node
         if self.node.first_child:
-            node = LexborNode()
-            node._cinit(<lxb_dom_node_t *> self.node.first_child, self.parser)
+            node = LexborNode.new(<lxb_dom_node_t *> self.node.first_child, self.parser)
             return node
         return None
 
@@ -50,9 +53,8 @@ cdef class LexborNode:
     def parent(self):
         """Return the parent node."""
         cdef LexborNode node
-        if self.node.parent:
-            node = LexborNode()
-            node._cinit(<lxb_dom_node_t *> self.node.parent, self.parser)
+        if self.node.parent != NULL:
+            node = LexborNode.new(<lxb_dom_node_t *> self.node.parent, self.parser)
             return node
         return None
 
@@ -60,9 +62,8 @@ cdef class LexborNode:
     def next(self):
         """Return next node."""
         cdef LexborNode node
-        if self.node.next:
-            node = LexborNode()
-            node._cinit(<lxb_dom_node_t *> self.node.next, self.parser)
+        if self.node.next != NULL:
+            node = LexborNode.new(<lxb_dom_node_t *> self.node.next, self.parser)
             return node
         return None
 
@@ -70,9 +71,8 @@ cdef class LexborNode:
     def prev(self):
         """Return previous node."""
         cdef LexborNode node
-        if self.node.prev:
-            node = LexborNode()
-            node._cinit(<lxb_dom_node_t *> self.node.prev, self.parser)
+        if self.node.prev != NULL:
+            node = LexborNode.new(<lxb_dom_node_t *> self.node.prev, self.parser)
             return node
         return None
 
@@ -80,9 +80,8 @@ cdef class LexborNode:
     def last_child(self):
         """Return last child node."""
         cdef LexborNode node
-        if self.node.last_child:
-            node = LexborNode()
-            node._cinit(<lxb_dom_node_t *> self.node.last_child, self.parser)
+        if self.node.last_child != NULL:
+            node = LexborNode.new(<lxb_dom_node_t *> self.node.last_child, self.parser)
             return node
         return None
 
@@ -181,6 +180,12 @@ cdef class LexborNode:
         Matches pattern `query` against HTML tree.
         `CSS selectors reference <https://www.w3schools.com/cssref/css_selectors.asp>`_.
 
+        Special selectors:
+
+         - parser.css('p:lexbor-contains("awesome" i)') -- case-insensitive contains
+         - parser.css('p:lexbor-contains("awesome")') -- case-sensitive contains
+
+
         Parameters
         ----------
         query : str
@@ -256,7 +261,6 @@ cdef class LexborNode:
             text = c_text.decode(_ENCODING)
         return text
 
-
     def decompose(self, bool recursive=True):
         """Remove the current node from the tree.
 
@@ -273,6 +277,9 @@ cdef class LexborNode:
         >>>     tag.decompose()
 
         """
+        if self.node == <lxb_dom_node_t *> lxb_dom_document_root(&self.parser.document.dom_document):
+            raise SelectolaxError("Decomposing the root node is not allowed.")
+
         if recursive:
             lxb_dom_node_destroy_deep(<lxb_dom_node_t *> self.node)
         else:
@@ -298,11 +305,11 @@ cdef class LexborNode:
         '<html><body><div>Hello world!</div></body></html>'
 
         """
+        cdef LexborNode element
         for tag in tags:
             for element in self.css(tag):
                 element.decompose(recursive=recursive)
 
-
     @property
     def attributes(self):
         """Get all attributes that belong to the current node.
@@ -410,15 +417,18 @@ cdef class LexborNode:
                 node = node.next
                 continue
 
-            next_node = LexborNode()
-            next_node._cinit(<lxb_dom_node_t *> node, self.parser)
+            next_node = LexborNode.new(<lxb_dom_node_t *> node, self.parser)
             yield next_node
             node = node.next
 
-
-    def unwrap(self):
+    def unwrap(self, bint delete_empty=False):
         """Replace node with whatever is inside this node.
 
+        Parameters
+        ----------
+        delete_empty : bool, default False
+            If True, removes empty tags.
+
         Examples
         --------
 
@@ -427,11 +437,14 @@ cdef class LexborNode:
         >>>  tree.html
         '<html><head></head><body><div>Hello world!</div></body></html>'
 
+        Note: by default, empty tags are ignored, use "delete_empty" to change this.
         """
         if self.node.first_child == NULL:
+            if delete_empty:
+                lxb_dom_node_destroy(<lxb_dom_node_t *> self.node)
             return
-        cdef lxb_dom_node_t* next_node;
-        cdef lxb_dom_node_t* current_node;
+        cdef lxb_dom_node_t* next_node
+        cdef lxb_dom_node_t* current_node
 
         if self.node.first_child.next != NULL:
             current_node = self.node.first_child
@@ -445,7 +458,7 @@ cdef class LexborNode:
             lxb_dom_node_insert_before(self.node, self.node.first_child)
         lxb_dom_node_destroy(<lxb_dom_node_t *> self.node)
 
-    def unwrap_tags(self, list tags):
+    def unwrap_tags(self, list tags, bint delete_empty = False):
         """Unwraps specified tags from the HTML tree.
 
         Works the same as the ``unwrap`` method, but applied to a list of tags.
@@ -454,6 +467,8 @@ cdef class LexborNode:
         ----------
         tags : list
             List of tags to remove.
+        delete_empty : bool, default False
+            If True, removes empty tags.
 
         Examples
         --------
@@ -462,12 +477,50 @@ cdef class LexborNode:
         >>> tree.body.unwrap_tags(['i','a'])
         >>> tree.body.html
         '<body><div>Hello world!</div></body>'
-        """
 
+        Note: by default, empty tags are ignored, use "delete_empty" to change this.
+        """
+        cdef LexborNode element
         for tag in tags:
             for element in self.css(tag):
-                element.unwrap()
+                element.unwrap(delete_empty)
+
+    def merge_text_nodes(self):
+        """Iterates over all text nodes and merges all text nodes that are close to each other.
 
+        This is useful for text extraction.
+        Use it when you need to strip HTML tags and merge "dangling" text.
+
+        Examples
+        --------
+
+        >>> tree = LexborHTMLParser("<div><p><strong>J</strong>ohn</p><p>Doe</p></div>")
+        >>> node = tree.css_first('div')
+        >>> tree.unwrap_tags(["strong"])
+        >>> tree.text(deep=True, separator=" ", strip=True)
+        "J ohn Doe" # Text extraction produces an extra space because the strong tag was removed.
+        >>> node.merge_text_nodes()
+        >>> tree.text(deep=True, separator=" ", strip=True)
+        "John Doe"
+        """
+        cdef lxb_dom_node_t *node = self.node.first_child
+        cdef lxb_dom_node_t *next_node
+        cdef lxb_char_t *left_text
+        cdef lxb_char_t *right_text
+        cdef size_t left_length, right_length
+
+        while node != NULL:
+            next_node = node.next
+            if node.type == LXB_DOM_NODE_TYPE_TEXT and node.prev and node.prev.type == LXB_DOM_NODE_TYPE_TEXT:
+                left_text = lxb_dom_node_text_content(node.prev, &left_length)
+                right_text = lxb_dom_node_text_content(node, &right_length)
+                if left_text and right_text:
+                    combined = (<bytes>left_text[:left_length]) + (<bytes>right_text[:right_length])
+                    lxb_dom_node_text_content_set(node, combined, len(combined))
+                    lxb_dom_node_remove(node.prev)
+            if node.first_child:
+                LexborNode.new(node, self.parser).merge_text_nodes()
+            node = next_node
 
     def traverse(self, include_text=False):
         """Iterate over all child and next nodes starting from the current level.
@@ -487,8 +540,7 @@ cdef class LexborNode:
 
         while node != NULL:
             if not (not include_text and node.type == LXB_DOM_NODE_TYPE_TEXT):
-                lxb_node = LexborNode()
-                lxb_node._cinit(<lxb_dom_node_t *> node, self.parser)
+                lxb_node = LexborNode.new(<lxb_dom_node_t *> node, self.parser)
                 yield lxb_node
 
             if node.first_child != NULL:
@@ -552,7 +604,6 @@ cdef class LexborNode:
         else:
             raise SelectolaxError("Expected a string or LexborNode instance, but %s found" % type(value).__name__)
 
-
     def insert_before(self, str_or_LexborNode value):
         """
         Insert a node before the current Node.
@@ -727,7 +778,7 @@ cdef class LexborNode:
         >>> selector.child.raw_value
         b'&#x3C;test&#x3E;'
         """
-        raise SelectolaxError("This features is not supported by the lexbor backend. Please use Modest backend.")
+        raise NotImplementedError("This features is not supported by the lexbor backend. Please use Modest backend.")
 
     def scripts_contain(self, str query):
         """Returns True if any of the script tags contain specified text.
@@ -740,6 +791,7 @@ cdef class LexborNode:
             The query to check.
 
         """
+        cdef LexborNode node
         if self.parser.cached_script_texts is None:
             nodes = self.parser.selector.find('script', self)
             text_nodes = []
@@ -764,6 +816,7 @@ cdef class LexborNode:
         queries : tuple of str
 
         """
+        cdef LexborNode node
         if self.parser.cached_script_srcs is None:
             nodes = self.parser.selector.find('script', self)
             src_nodes = []
@@ -819,31 +872,44 @@ cdef class LexborNode:
         """
         cdef unsigned char * text
         cdef lxb_dom_node_t* node = <lxb_dom_node_t*> self.node.first_child
-
-        container = TextContainer()
+        cdef TextContainer container
         if self.node == NULL or self.node.type != LXB_DOM_NODE_TYPE_TEXT:
             return None
+
         text = <unsigned char *> lexbor_str_data_noi(&(<lxb_dom_character_data_t *> self.node).data)
         if text != NULL:
+            container = TextContainer.new_with_defaults()
             py_text = text.decode(_ENCODING)
             container.append(py_text)
             return container.text
+
+
+@cython.internal
 @cython.final
 cdef class TextContainer:
     cdef str _text
-    cdef public str separator
-    cdef public bool strip
+    cdef str separator
+    cdef bint strip
+
+    @staticmethod
+    cdef TextContainer new_with_defaults():
+        cdef TextContainer cls = TextContainer.__new__(TextContainer)
+        cls._text = ''
+        cls.separator = ''
+        cls.strip = False
+        return cls
 
     def __init__(self, str separator = '', bool strip = False):
         self._text = ""
         self.separator = separator
         self.strip = strip
 
-    def append(self, node_text):
+    def append(self, str node_text):
         if self.strip:
             self._text += node_text.strip() + self.separator
         else:
             self._text += node_text + self.separator
+
     @property
     def text(self):
         if self.separator and self._text and self._text.endswith(self.separator):
@@ -852,7 +918,7 @@ cdef class TextContainer:
 
 
 cdef lexbor_action_t text_callback(lxb_dom_node_t *node, void *ctx):
-    cdef unsigned char *text;
+    cdef unsigned char *text
     cdef lxb_tag_id_t tag_id = lxb_dom_node_tag_id_noi(node)
     if tag_id != LXB_TAG__TEXT:
         return LEXBOR_ACTION_OK
@@ -860,8 +926,15 @@ cdef lexbor_action_t text_callback(lxb_dom_node_t *node, void *ctx):
     text = <unsigned char*> lexbor_str_data_noi(&(<lxb_dom_text_t *> node).char_data.data)
     if not text:
         return LEXBOR_ACTION_OK
-    py_str = text.decode(_ENCODING)
-    cdef object cls
-    cls = <object> ctx
+
+    try:
+        py_str = text.decode(_ENCODING)
+
+    except Exception as e:
+        PyErr_SetNone(e)
+        return LEXBOR_ACTION_STOP
+
+    cdef TextContainer cls
+    cls = <TextContainer> ctx
     cls.append(py_str)
     return LEXBOR_ACTION_OK

selectolax/lexbor/selection.pxi

@@ -1,4 +1,7 @@
 cimport cython
+from cpython.exc cimport PyErr_SetObject
+from cpython.list cimport PyList_GET_SIZE
+
 
 @cython.final
 cdef class LexborCSSSelector:
@@ -8,21 +11,22 @@ cdef class LexborCSSSelector:
         self.results = []
         self.current_node = None
 
-    cdef _create_css_parser(self):
+    cdef int _create_css_parser(self) except -1:
         cdef lxb_status_t status
 
-
         self.parser = lxb_css_parser_create()
         status = lxb_css_parser_init(self.parser, NULL)
 
         if status != LXB_STATUS_OK:
-            raise SelectolaxError("Can't initialize CSS parser.")
+            PyErr_SetObject(SelectolaxError, "Can't initialize CSS parser.")
+            return -1
 
         self.css_selectors = lxb_css_selectors_create()
         status = lxb_css_selectors_init(self.css_selectors)
 
         if status != LXB_STATUS_OK:
-            raise SelectolaxError("Can't initialize CSS selector.")
+            PyErr_SetObject(SelectolaxError, "Can't initialize CSS selector.")
+            return -1
 
         lxb_css_parser_selectors_set(self.parser, self.css_selectors)
 
@@ -30,14 +34,18 @@ cdef class LexborCSSSelector:
         status = lxb_selectors_init(self.selectors)
         lxb_selectors_opt_set(self.selectors, LXB_SELECTORS_OPT_MATCH_ROOT)
         if status != LXB_STATUS_OK:
-            raise SelectolaxError("Can't initialize CSS selector.")
-
+            PyErr_SetObject(SelectolaxError, "Can't initialize CSS selector.")
+            return -1
+        return 0
 
-    cpdef find(self, str query, LexborNode node):
+    cpdef list find(self, str query, LexborNode node):
         cdef lxb_css_selector_list_t* selectors
         cdef lxb_char_t* c_selector
         cdef lxb_css_selector_list_t * selectors_list
 
+        if not isinstance(query, str):
+            raise TypeError("Query must be a string.")
+
         bytes_query = query.encode(_ENCODING)
         selectors_list = lxb_css_selectors_parse(self.parser, <lxb_char_t *> bytes_query, <size_t>len(query))
 
@@ -54,28 +62,32 @@ cdef class LexborCSSSelector:
         lxb_css_selector_list_destroy_memory(selectors_list)
         return results
 
-    cpdef any_matches(self, str query, LexborNode node):
+    cpdef int any_matches(self, str query, LexborNode node) except -1:
         cdef lxb_css_selector_list_t * selectors
         cdef lxb_char_t * c_selector
         cdef lxb_css_selector_list_t * selectors_list
+        cdef int result
+
+        if not isinstance(query, str):
+            raise TypeError("Query must be a string.")
 
         bytes_query = query.encode(_ENCODING)
         selectors_list = lxb_css_selectors_parse(self.parser, <lxb_char_t *> bytes_query, <size_t> len(query))
 
         if selectors_list == NULL:
-            raise SelectolaxError("Can't parse CSS selector.")
+            PyErr_SetObject(SelectolaxError, "Can't parse CSS selector.")
 
         self.results = []
         status = lxb_selectors_find(self.selectors, node.node, selectors_list,
                                     <lxb_selectors_cb_f> css_matcher_callback, <void *> self)
         if status != LXB_STATUS_OK:
-            raise SelectolaxError("Can't parse CSS selector.")
-        result = bool(self.results)
+            lxb_css_selector_list_destroy_memory(selectors_list)
+            PyErr_SetObject(SelectolaxError, "Can't parse CSS selector.")
+        result = PyList_GET_SIZE(self.results) > 0
         self.results = []
         lxb_css_selector_list_destroy_memory(selectors_list)
         return result
 
-
     def __dealloc__(self):
         if self.selectors != NULL:
             lxb_selectors_destroy(self.selectors, True)
@@ -85,7 +97,6 @@ cdef class LexborCSSSelector:
             lxb_css_selectors_destroy(self.css_selectors, True)
 
 
-
 cdef class LexborSelector:
     """An advanced CSS selector that supports additional operations.
 
@@ -100,10 +111,9 @@ cdef class LexborSelector:
         self.node = node
         self.nodes = self.node.parser.selector.find(query, self.node) if query else [node, ]
 
-
     cpdef css(self, str query):
         """Evaluate CSS selector against current scope."""
-        raise SelectolaxError("This features is not supported by the lexbor backend. Please use Modest backend.")
+        raise NotImplementedError("This features is not supported by the lexbor backend. Please use Modest backend.")
 
     @property
     def matches(self) -> list:
@@ -117,7 +127,7 @@ cdef class LexborSelector:
 
     def text_contains(self, str text, bool deep=True, str separator='', bool strip=False) -> LexborSelector:
         """Filter all current matches given text."""
-        nodes = []
+        cdef list nodes = []
         for node in self.nodes:
             node_text = node.text(deep=deep, separator=separator, strip=strip)
             if node_text and text in node_text:
@@ -127,7 +137,7 @@ cdef class LexborSelector:
 
     def any_text_contains(self, str text, bool deep=True, str separator='', bool strip=False) -> bool:
         """Returns True if any node in the current search scope contains specified text"""
-        nodes = []
+        cdef LexborNode node
         for node in self.nodes:
             node_text = node.text(deep=deep, separator=separator, strip=strip)
             if node_text and text in node_text:
@@ -139,7 +149,7 @@ cdef class LexborSelector:
 
         Similar to `string-length` in XPath.
         """
-        nodes = []
+        cdef list nodes = []
         for node in self.nodes:
             attr = node.attributes.get(attribute)
             if attr and start and start in attr:
@@ -154,7 +164,7 @@ cdef class LexborSelector:
 
         Similar to `string-length` in XPath.
         """
-        nodes = []
+        cdef LexborNode node
         for node in self.nodes:
             attr = node.attributes.get(attribute)
             if attr and start and start in attr:
@@ -169,16 +179,15 @@ cdef class LexborSelector:
 
 cdef lxb_status_t css_finder_callback(lxb_dom_node_t *node, lxb_css_selector_specificity_t *spec, void *ctx):
     cdef LexborNode lxb_node
-    cdef object cls
-    cls = <object> ctx
-    lxb_node = LexborNode()
-    lxb_node._cinit(<lxb_dom_node_t *> node, cls.current_node.parser)
+    cdef LexborCSSSelector cls
+    cls = <LexborCSSSelector> ctx
+    lxb_node = LexborNode.new(<lxb_dom_node_t *> node, cls.current_node.parser)
     cls.results.append(lxb_node)
     return LXB_STATUS_OK
 
 cdef lxb_status_t css_matcher_callback(lxb_dom_node_t *node, lxb_css_selector_specificity_t *spec, void *ctx):
     cdef LexborNode lxb_node
-    cdef object cls
-    cls = <object> ctx
+    cdef LexborCSSSelector cls
+    cls = <LexborCSSSelector> ctx
     cls.results.append(True)
     return LXB_STATUS_STOP

selectolax/lexbor/util.pxi

@@ -1,5 +1,6 @@
 include "../utils.pxi"
 
+
 def create_tag(tag: str):
     """
     Given an HTML tag name, e.g. `"div"`, create a single empty node for that tag,