html_to_vdom transform to remove html/body but preserve head content (#832)

Archmonger · rmorshea · web-flow · commit 9c3c300aa571 · 2022-12-01T11:24:38.000-08:00
* Insert head content into html_to_vdom * more efficient node appending * add changelog * nodes -> body node * parsed_document -> html_node * remove useless has_root_node variable * fix comment * del_html_body_transform * fix type hint errors * type hint fixes 2 * remove unusable head and body tags * uno mas * import future annotations * fix docs warnings * clean up last warning * fix docstrings * del_html_body_transform docstrings * re-add head API * fix changelog PR links * docstring cleanup * Better type hint * Revert "Better type hint" This reverts commit d78afdb. * more concise transform implementation * fix merge error * merge changelog better * annotate as VdomDict Co-authored-by: Ryan Morshead <ryan.morshead@gmail.com>
diff --git a/docs/source/about/changelog.rst b/docs/source/about/changelog.rst
@@ -23,21 +23,23 @@ more info, see the :ref:`Contributor Guide <Creating a Changelog Entry>`.
 Unreleased
 ----------
 
-**Removed**
-
-- :pull:`840` - Remove ``IDOM_FEATURE_INDEX_AS_DEFAULT_KEY`` option
-- :pull:`835` - ``serve_static_files`` option from backend configuration
-
 **Added**
 
 - :pull:`835` - Ability to customize the ``<head>`` element of IDOM's built-in client.
 - :pull:`835` - ``vdom_to_html`` utility function.
 - :pull:`843` - Ability to subscribe to changes that are made to mutable options.
+- :pull:`832` - ``del_html_head_body_transform`` to remove ``<html>``, ``<head>``, and ``<body>`` while preserving children.
 - :pull:`699` - Support for form element serialization
 
 **Fixed**
 
 - :issue:`582` - ``IDOM_DEBUG_MODE`` is now mutable and can be changed at runtime
+- :pull:`832` - Fix ``html_to_vdom`` improperly removing ``<html>``, ``<head>``, and ``<body>`` nodes.
+
+**Removed**
+- :pull:`832` - Removed ``idom.html.body`` as it is currently unusable due to technological limitations, and thus not needed.
+- :pull:`840` - remove ``IDOM_FEATURE_INDEX_AS_DEFAULT_KEY`` option
+- :pull:`835` - ``serve_static_files`` option from backend configuration
 
 
 v0.41.0
diff --git a/src/idom/html.py b/src/idom/html.py
@@ -14,7 +14,6 @@
 
 **Content sectioning**
 
-- :func:`body`
 - :func:`address`
 - :func:`article`
 - :func:`aside`
@@ -189,7 +188,6 @@ def _(*children: Any, key: Key | None = None) -> VdomDict:
 title = make_vdom_constructor("title")
 
 # Content sectioning
-body = make_vdom_constructor("body")
 address = make_vdom_constructor("address")
 article = make_vdom_constructor("article")
 aside = make_vdom_constructor("aside")
diff --git a/src/idom/utils.py b/src/idom/utils.py
@@ -5,7 +5,7 @@
 from typing import Any, Callable, Generic, Iterable, TypeVar, cast
 
 from lxml import etree
-from lxml.html import fragments_fromstring, tostring
+from lxml.html import fromstring, tostring
 
 import idom
 from idom.core.types import VdomDict
@@ -85,7 +85,7 @@ def html_to_vdom(
     using a ``key=...`` attribute within your HTML tag.
 
     Parameters:
-        source:
+        html:
             The raw HTML as a string
         transforms:
             Functions of the form ``transform(old) -> new`` where ``old`` is a VDOM
@@ -99,15 +99,15 @@ def html_to_vdom(
         raise TypeError(f"Expected html to be a string, not {type(html).__name__}")
 
     # If the user provided a string, convert it to a list of lxml.etree nodes
-    parser = etree.HTMLParser(
-        remove_comments=True,
-        remove_pis=True,
-        remove_blank_text=True,
-        recover=not strict,
-    )
     try:
-        nodes: list[etree._Element] = fragments_fromstring(
-            html, no_leading_text=True, parser=parser
+        root_node: etree._Element = fromstring(
+            html.strip(),
+            parser=etree.HTMLParser(
+                remove_comments=True,
+                remove_pis=True,
+                remove_blank_text=True,
+                recover=not strict,
+            ),
         )
     except etree.XMLSyntaxError as e:
         if not strict:
@@ -119,25 +119,8 @@ def html_to_vdom(
             "you can disable the strict parameter on html_to_vdom().\n"
             "Otherwise, repair your broken HTML and try again."
         ) from e
-    has_root_node = len(nodes) == 1
-
-    # Find or create a root node
-    if has_root_node:
-        root_node = nodes[0]
-    else:
-        # etree.Element requires a non-empty tag - we correct this below
-        root_node = etree.Element("TEMP", None, None)
-        for child in nodes:
-            root_node.append(child)
 
-    # Convert the lxml node to a VDOM dict
-    vdom = _etree_to_vdom(root_node, transforms)
-
-    # Change the artificially created root node to a React Fragment, instead of a div
-    if not has_root_node:
-        vdom["tagName"] = ""
-
-    return vdom
+    return _etree_to_vdom(root_node, transforms)
 
 
 class HTMLParseError(etree.LxmlSyntaxError):  # type: ignore[misc]
@@ -147,10 +130,10 @@ class HTMLParseError(etree.LxmlSyntaxError):  # type: ignore[misc]
 def _etree_to_vdom(
     node: etree._Element, transforms: Iterable[_ModelTransform]
 ) -> VdomDict:
-    """Recusively transform an lxml etree node into a DOM model
+    """Transform an lxml etree node into a DOM model
 
     Parameters:
-        source:
+        node:
             The ``lxml.etree._Element`` node
         transforms:
             Functions of the form ``transform(old) -> new`` where ``old`` is a VDOM
@@ -162,7 +145,7 @@ def _etree_to_vdom(
             f"Expected node to be a etree._Element, not {type(node).__name__}"
         )
 
-    # This will recursively call _etree_to_vdom() on all children
+    # Recursively call _etree_to_vdom() on all children
     children = _generate_vdom_children(node, transforms)
 
     # Convert the lxml node to a VDOM dict
@@ -289,6 +272,20 @@ def _hypen_to_camel_case(string: str) -> str:
     return first.lower() + remainder.title().replace("-", "")
 
 
+def del_html_head_body_transform(vdom: VdomDict) -> VdomDict:
+    """Transform intended for use with `html_to_vdom`.
+
+    Removes `<html>`, `<head>`, and `<body>` while preserving their children.
+
+    Parameters:
+        vdom:
+            The VDOM dictionary to transform.
+    """
+    if vdom["tagName"] in {"html", "body", "head"}:
+        return {"tagName": "", "children": vdom["children"]}
+    return vdom
+
+
 def _vdom_attr_to_html_str(key: str, value: Any) -> tuple[str, str]:
     if key == "style":
         if isinstance(value, dict):
diff --git a/tests/test_utils.py b/tests/test_utils.py
@@ -4,7 +4,12 @@
 
 import idom
 from idom import html
-from idom.utils import HTMLParseError, html_to_vdom, vdom_to_html
+from idom.utils import (
+    HTMLParseError,
+    del_html_head_body_transform,
+    html_to_vdom,
+    vdom_to_html,
+)
 
 
 def test_basic_ref_behavior():
@@ -144,7 +149,7 @@ def test_html_to_vdom_with_no_parent_node():
     source = "<p>Hello</p><div>World</div>"
 
     expected = {
-        "tagName": "",
+        "tagName": "div",
         "children": [
             {"tagName": "p", "children": ["Hello"]},
             {"tagName": "div", "children": ["World"]},
@@ -154,6 +159,37 @@ def test_html_to_vdom_with_no_parent_node():
     assert html_to_vdom(source) == expected
 
 
+def test_del_html_body_transform():
+    source = """
+    <!DOCTYPE html>
+    <html lang="en">
+
+    <head>
+    <title>My Title</title>
+    </head>
+
+    <body><h1>Hello World</h1></body>
+
+    </html>
+    """
+
+    expected = {
+        "tagName": "",
+        "children": [
+            {
+                "tagName": "",
+                "children": [{"tagName": "title", "children": ["My Title"]}],
+            },
+            {
+                "tagName": "",
+                "children": [{"tagName": "h1", "children": ["Hello World"]}],
+            },
+        ],
+    }
+
+    assert html_to_vdom(source, del_html_head_body_transform) == expected
+
+
 SOME_OBJECT = object()
 
 
