Merge branch 'main' into to-dict-from-arraylike

doc/user_guide/customization.rst

@@ -823,14 +823,6 @@ The preferred format of numbers, dates, and currencies varies by language and lo
 Vega-Altair takes advantage of `D3's localization support`_ to make it easy to configure
 the locale for your chart using the global ``alt.renderers.set_embed_options`` function.
 
-.. altair-plot::
-   :output: none
-
-   import altair as alt
-   alt.renderers.set_embed_options(
-       format_locale=format_locale, time_format_locale=time_format_locale
-   )
-
 Here ``format_locale`` and ``time_format_locale`` may either be D3 format dictionaries,
 or strings with the names of pre-defined locales. For example, here we use the
 Italian locale (named ``it-IT``) for both currencies and dates:

pyproject.toml

@@ -68,12 +68,12 @@ dev = [
     "hatch",
     "ruff>=0.5.3",
     "ibis-framework[polars]",
-    "ipython",
+    "ipython[kernel]",
     "pandas>=0.25.3",
     "pytest",
     "pytest-cov",
     "pytest-xdist[psutil]~=3.5",
-    "m2r",
+    "mistune",
     "mypy",
     "pandas-stubs",
     "types-jsonschema",

tools/generate_schema_wrapper.py

@@ -12,7 +12,6 @@
 from typing import Final, Iterable, Literal, Iterator
 from itertools import chain
 from urllib import request
-import m2r
 
 sys.path.insert(0, str(Path.cwd()))
 from tools.schemapi import codegen, CodeSnippet, SchemaInfo
@@ -23,6 +22,7 @@
     rst_syntax_for_class,
     indent_docstring,
     ruff_write_lint_format_str,
+    rst_parse,
 )
 
 
@@ -271,26 +271,20 @@ def _process_description(description: str) -> str:
 
 
 def process_description(description: str) -> str:
+    # remove formatting from links
     description = "".join(
         [
             reSpecial.sub("", d) if i % 2 else d
             for i, d in enumerate(reLink.split(description))
         ]
-    )  # remove formatting from links
-    description = m2r.convert(description)
-    description = description.replace(m2r.prolog, "")
-    description = description.replace(":raw-html-m2r:", ":raw-html:")
-    description = description.replace(r"\ ,", ",")
-    description = description.replace(r"\ ", " ")
-    # turn explicit references into anonymous references
-    description = description.replace(">`_", ">`__")
+    )
+    description = rst_parse(description)
     # Some entries in the Vega-Lite schema miss the second occurence of '__'
     description = description.replace("__Default value: ", "__Default value:__ ")
     # Fixing ambiguous unicode, RUF001 produces RUF002 in docs
     description = description.replace("’", "'")  # noqa: RUF001 [RIGHT SINGLE QUOTATION MARK]
     description = description.replace("–", "-")  # noqa: RUF001 [EN DASH]
     description = description.replace(" ", " ")  # noqa: RUF001 [NO-BREAK SPACE]
-    description += "\n"
     return description.strip()
 
 

tools/schemapi/utils.py

@@ -9,9 +9,15 @@
 import urllib
 from typing import Any, Final, Iterable, TYPE_CHECKING, Iterator, Sequence
 from operator import itemgetter
+from html import unescape
+
+import mistune
+from mistune.renderers.rst import RSTRenderer as _RSTRenderer
+
 from tools.schemapi.schemapi import _resolve_references as resolve_references
 
 if TYPE_CHECKING:
+    from mistune import BlockState
     from typing_extensions import LiteralString
     from pathlib import Path
 
@@ -652,6 +658,33 @@ def is_union_enum(self) -> bool:
         return self.is_union() and all(el.is_enum() for el in self.anyOf)
 
 
+class RSTRenderer(_RSTRenderer):
+    def __init__(self) -> None:
+        super().__init__()
+
+    def inline_html(self, token: dict[str, Any], state: BlockState) -> str:
+        html = token["raw"]
+        return rf"\ :raw-html:`{html}`\ "
+
+
+class RSTParse(mistune.Markdown):
+    def __init__(
+        self,
+        renderer: mistune.BaseRenderer,
+        block: mistune.BlockParser | None = None,
+        inline: mistune.InlineParser | None = None,
+        plugins=None,
+    ) -> None:
+        super().__init__(renderer, block, inline, plugins)
+
+    def __call__(self, s: str) -> str:
+        s = super().__call__(s)
+        return unescape(s).replace(r"\ ,", ",").replace(r"\ ", " ")
+
+
+rst_parse: RSTParse = RSTParse(RSTRenderer())
+
+
 def indent_docstring(
     lines: list[str], indent_level: int, width: int = 100, lstrip=True
 ) -> str: