Merge pull request 3b1b#1557 from naveen521kk/add-markuptext

Add MarkupText

manimlib/mobject/svg/text_mobject.py

@@ -2,8 +2,12 @@
 import hashlib
 import os
 import re
+import io
 import typing
 import warnings
+import xml.etree.ElementTree as ET
+import functools
+
 from contextlib import contextmanager
 from pathlib import Path
 
@@ -15,8 +19,7 @@
 from manimlib.utils.config_ops import digest_config
 from manimlib.utils.customization import get_customization
 from manimlib.utils.directories import get_downloads_dir, get_text_dir
-from manimpango import PangoUtils
-from manimpango import TextSetting
+from manimpango import PangoUtils, TextSetting, MarkupUtils
 
 TEXT_MOB_SCALE_FACTOR = 0.0076
 DEFAULT_LINE_SPACING_SCALE = 0.3
@@ -238,6 +241,237 @@ def text2svg(self):
         )
 
 
+class MarkupText(SVGMobject):
+    CONFIG = {
+        # Mobject
+        "color": WHITE,
+        "height": None,
+        # Text
+        "font": '',
+        "font_size": 48,
+        "lsh": None,
+        "justify": False,
+        "slant": NORMAL,
+        "weight": NORMAL,
+        "tab_width": 4,
+        "gradient": None,
+        "disable_ligatures": True,
+    }
+    def __init__(self, text, **config):
+        digest_config(self, config)
+        self.text = f'<span>{text}</span>'
+        self.original_text = self.text
+        self.text_for_parsing = self.text
+        text_without_tabs = text
+        if "\t" in text:
+            text_without_tabs = text.replace("\t", " " * self.tab_width)
+        try:
+            colormap = self.extract_color_tags()
+            gradientmap = self.extract_gradient_tags()
+        except ET.ParseError:
+            # let pango handle that error
+            pass
+        validate_error = MarkupUtils.validate(self.text)
+        if validate_error:
+            raise ValueError(validate_error)
+        file_name = self.text2svg()
+        PangoUtils.remove_last_M(file_name)
+        super().__init__(
+            file_name,
+            **config,
+        )
+        self.chars = self.get_group_class()(*self.submobjects)
+        self.text = text_without_tabs.replace(" ", "").replace("\n", "")
+        if self.gradient:
+            self.set_color_by_gradient(*self.gradient)
+        for col in colormap:
+            self.chars[
+                col["start"]
+                - col["start_offset"] : col["end"]
+                - col["start_offset"]
+                - col["end_offset"]
+            ].set_color(self._parse_color(col["color"]))
+        for grad in gradientmap:
+            self.chars[
+                grad["start"]
+                - grad["start_offset"] : grad["end"]
+                - grad["start_offset"]
+                - grad["end_offset"]
+            ].set_color_by_gradient(
+                *(self._parse_color(grad["from"]), self._parse_color(grad["to"]))
+            )
+        # anti-aliasing
+        if self.height is None:
+            self.scale(TEXT_MOB_SCALE_FACTOR)
+    def text2hash(self):
+        """Generates ``sha256`` hash for file name."""
+        settings = (
+            "MARKUPPANGO" + self.font + self.slant + self.weight + self.color
+        )  # to differentiate from classical Pango Text
+        settings += str(self.lsh) + str(self.font_size)
+        settings += str(self.disable_ligatures)
+        settings += str(self.justify)
+        id_str = self.text + settings
+        hasher = hashlib.sha256()
+        hasher.update(id_str.encode())
+        return hasher.hexdigest()[:16]
+
+    def text2svg(self):
+        """Convert the text to SVG using Pango."""
+        size = self.font_size
+        dir_name = get_text_dir()
+        disable_liga = self.disable_ligatures
+        if not os.path.exists(dir_name):
+            os.makedirs(dir_name)
+        hash_name = self.text2hash()
+        file_name = os.path.join(dir_name, hash_name) + ".svg"
+        if os.path.exists(file_name):
+            return file_name
+
+        extra_kwargs = {}
+        extra_kwargs['justify'] = self.justify
+        extra_kwargs['pango_width'] = DEFAULT_PIXEL_WIDTH - 100
+        if self.lsh:
+            extra_kwargs['line_spacing']=self.lsh
+        return MarkupUtils.text2svg(
+            f'<span foreground="{self.color}">{self.text}</span>',
+            self.font,
+            self.slant,
+            self.weight,
+            size,
+            0, # empty parameter
+            disable_liga,
+            file_name,
+            START_X,
+            START_Y,
+            DEFAULT_PIXEL_WIDTH,  # width
+            DEFAULT_PIXEL_HEIGHT,  # height
+            **extra_kwargs
+        )
+
+
+    def _parse_color(self, col):
+        """Parse color given in ``<color>`` or ``<gradient>`` tags."""
+        if re.match("#[0-9a-f]{6}", col):
+            return col
+        else:
+            return globals()[col.upper()] # this is hacky
+
+    @functools.lru_cache(10)
+    def get_text_from_markup(self, element=None):
+        if not element:
+            element = ET.fromstring(self.text_for_parsing)
+        final_text = ''
+        for i in element.itertext():
+            final_text += i
+        return final_text
+
+    def extract_color_tags(self, text=None, colormap = None):
+        """Used to determine which parts (if any) of the string should be formatted
+        with a custom color.
+        Removes the ``<color>`` tag, as it is not part of Pango's markup and would cause an error.
+        Note: Using the ``<color>`` tags is deprecated. As soon as the legacy syntax is gone, this function
+        will be removed.
+        """
+        if not text:
+            text = self.text_for_parsing
+        if not colormap:
+            colormap = list()
+        elements = ET.fromstring(text)
+        text_from_markup = self.get_text_from_markup()
+        final_xml = ET.fromstring(f'<span>{elements.text if elements.text else ""}</span>')
+        def get_color_map(elements):
+            for element in elements:
+                if element.tag == 'color':
+                    element_text = self.get_text_from_markup(element)
+                    start = text_from_markup.find(element_text)
+                    end = start + len(element_text)
+                    offsets = element.get('offset').split(",") if element.get('offset') else [0]
+                    start_offset = int(offsets[0]) if offsets[0] else 0
+                    end_offset = int(offsets[1]) if len(offsets) == 2 and offsets[1] else 0
+                    colormap.append(
+                        {
+                            "start": start,
+                            "end": end,
+                            "color": element.get('col'),
+                            "start_offset": start_offset,
+                            "end_offset": end_offset,
+                        }
+                    )
+
+                    _elements_list = list(element.iter())
+                    if len(_elements_list) <= 1:
+                        final_xml.append(ET.fromstring(f'<span>{element.text if element.text else ""}</span>'))
+                    else:
+                        final_xml.append(_elements_list[-1])
+                else:
+                    if len(list(element.iter())) == 1:
+                        final_xml.append(element)
+                    else:
+                        get_color_map(element)
+        get_color_map(elements)
+        with io.BytesIO() as f:
+            tree = ET.ElementTree()  
+            tree._setroot(final_xml)
+            tree.write(f)
+            self.text = f.getvalue().decode()
+        self.text_for_parsing = self.text # gradients will use it
+        return colormap
+
+    def extract_gradient_tags(self, text=None,gradientmap=None):
+        """Used to determine which parts (if any) of the string should be formatted
+        with a gradient.
+        Removes the ``<gradient>`` tag, as it is not part of Pango's markup and would cause an error.
+        """
+        if not text:
+            text = self.text_for_parsing
+        if not gradientmap:
+            gradientmap = list()
+
+        elements = ET.fromstring(text)
+        text_from_markup = self.get_text_from_markup()
+        final_xml = ET.fromstring(f'<span>{elements.text if elements.text else ""}</span>')
+        def get_gradient_map(elements):
+            for element in elements:
+                if element.tag == 'gradient':
+                    element_text = self.get_text_from_markup(element)
+                    start = text_from_markup.find(element_text)
+                    end = start + len(element_text)
+                    offsets = element.get('offset').split(",") if element.get('offset') else [0]
+                    start_offset = int(offsets[0]) if offsets[0] else 0
+                    end_offset = int(offsets[1]) if len(offsets) == 2 and offsets[1] else 0
+                    gradientmap.append(
+                        {
+                            "start": start,
+                            "end": end,
+                            "from": element.get('from'),
+                            "to": element.get('to'),
+                            "start_offset": start_offset,
+                            "end_offset": end_offset,
+                        }
+                    )
+                    _elements_list = list(element.iter())
+                    if len(_elements_list) == 1:
+                        final_xml.append(ET.fromstring(f'<span>{element.text if element.text else ""}</span>'))
+                    else:
+                        final_xml.append(_elements_list[-1])
+                else:
+                    if len(list(element.iter())) == 1:
+                        final_xml.append(element)
+                    else:
+                        get_gradient_map(element)
+        get_gradient_map(elements)
+        with io.BytesIO() as f:
+            tree = ET.ElementTree()  
+            tree._setroot(final_xml)
+            tree.write(f)
+            self.text = f.getvalue().decode()
+
+        return gradientmap
+
+    def __repr__(self):
+        return f"MarkupText({repr(self.original_text)})"
+
 @contextmanager
 def register_font(font_file: typing.Union[str, Path]):
     """Temporarily add a font file to Pango's search path.