Refactor MTex

This commit is contained in:
YishiMichael 2022-03-27 14:44:50 +08:00
parent e44a2fc8c6
commit 3b01ec48e6
No known key found for this signature in database
GPG key ID: EC615C0C5A86BC80

View file

@ -6,7 +6,7 @@ import itertools as it
from types import MethodType
from typing import Iterable, Union, Sequence
from manimlib.constants import WHITE
from manimlib.constants import BLACK, WHITE
from manimlib.mobject.svg.svg_mobject import SVGMobject
from manimlib.mobject.types.vectorized_mobject import VGroup
from manimlib.utils.color import color_to_int_rgb
@ -32,6 +32,9 @@ SCALE_FACTOR_PER_FONT_POINT = 0.001
class _TexSVG(SVGMobject):
CONFIG = {
"height": None,
"svg_default": {
"fill_color": BLACK,
},
"stroke_width": 0,
"stroke_color": WHITE,
"path_string_config": {
@ -80,25 +83,11 @@ class MTex(_TexSVG):
self.use_plain_tex
)
def get_file_path(self) -> str:
self.init_parser()
self.use_plain_file = any([
self.use_plain_tex,
self.color_cmd_repl_items,
self.base_color != WHITE
])
return self.get_file_path_(use_plain_file=self.use_plain_file)
def get_file_path_(self, use_plain_file: bool) -> str:
def get_file_path(self, use_plain_file: bool = False) -> str:
if use_plain_file:
content = "".join([
"{{",
self.get_color_command(self.color_to_int(self.base_color)),
self.string,
"}}"
])
content = self.plain_string
else:
content = self.get_labelled_string()
content = self.labelled_string
full_tex = self.get_tex_file_body(content)
with display_during_execution(f"Writing \"{self.string}\""):
@ -128,27 +117,26 @@ class MTex(_TexSVG):
def generate_mobject(self) -> None:
super().generate_mobject()
glyphs = self.submobjects
if not glyphs:
if not self.submobjects:
return
if self.use_plain_file:
file_path = self.get_file_path_(use_plain_file=False)
labelled_svg_glyphs = _TexSVG(file_path)
predefined_colors = [
labelled_glyph.get_fill_color()
for labelled_glyph in self.submobjects
]
else:
labelled_svg_glyphs = self
predefined_colors = [self.base_color] * len(glyphs)
glyph_labels = [
self.color_to_label(labelled_glyph.get_fill_color())
for labelled_glyph in labelled_svg_glyphs
self.color_to_label(glyph.get_fill_color())
for glyph in self.submobjects
]
for glyph, glyph_color in zip(glyphs, predefined_colors):
glyph.set_fill(glyph_color)
if any([
self.use_plain_tex,
self.color_cmd_repl_items,
self.base_color in (BLACK, WHITE)
]):
file_path = self.get_file_path(use_plain_file=True)
glyphs = _TexSVG(file_path).submobjects
for glyph, plain_glyph in zip(self.submobjects, glyphs):
glyph.set_fill(plain_glyph.get_fill_color())
else:
glyphs = self.submobjects
self.set_fill(self.base_color)
# Simply pack together adjacent mobjects with the same label.
submob_labels, glyphs_lists = self.group_neighbours(
@ -171,20 +159,18 @@ class MTex(_TexSVG):
## Static methods
@staticmethod
def color_to_int(color: ManimColor) -> int:
def color_to_label(color: ManimColor) -> int:
r, g, b = color_to_int_rgb(color)
rg = r * 256 + g
return rg * 256 + b
@staticmethod
def color_to_label(color: ManimColor) -> int:
result = MTex.color_to_int(color)
if result == 16777215: # white
rgb = rg * 256 + b
if rgb == 16777215: # white
return -1
return result
return rgb
@staticmethod
def get_color_command(label: int) -> str:
if label == -1:
label = 16777215 # white
rg, b = divmod(label, 256)
r, g = divmod(rg, 256)
return "".join([
@ -223,40 +209,83 @@ class MTex(_TexSVG):
groups.append(new_group)
return group_labels, groups
@staticmethod
def find_region_index(val: int, seq: list[int]) -> int:
# Returns an integer in `range(len(seq) + 1)` satisfying
# `seq[result - 1] <= val < seq[result]`
if not seq:
return 0
if val >= seq[-1]:
return len(seq)
result = 0
while val >= seq[result]:
result += 1
return result
@staticmethod
def lstrip(index: int, skipped_spans: list[tuple[int, int]]) -> int:
index_seq = list(it.chain(*skipped_spans))
region_index = MTex.find_region_index(index, index_seq)
if region_index % 2 == 1:
return index_seq[region_index]
return index
@staticmethod
def rstrip(index: int, skipped_spans: list[tuple[int, int]]) -> int:
index_seq = list(it.chain(*skipped_spans))
region_index = MTex.find_region_index(index - 1, index_seq)
if region_index % 2 == 1:
return index_seq[region_index - 1]
return index
@staticmethod
def strip(
tex_span: tuple[int, int], skipped_spans: list[tuple[int, int]]
) -> tuple[int, int] | None:
result = (
MTex.lstrip(tex_span[0], skipped_spans),
MTex.rstrip(tex_span[1], skipped_spans)
)
if result[0] >= result[1]:
return None
return result
@staticmethod
def lslide(index: int, slid_spans: list[tuple[int, int]]) -> int:
slide_dict = dict(slid_spans)
while index in slide_dict.keys():
index = slide_dict[index]
return index
@staticmethod
def rslide(index: int, slid_spans: list[tuple[int, int]]) -> int:
slide_dict = dict([
slide_span[::-1] for slide_span in slid_spans
])
while index in slide_dict.keys():
index = slide_dict[index]
return index
@staticmethod
def slide(
tex_span: tuple[int, int], slid_spans: list[tuple[int, int]]
) -> tuple[int, int] | None:
result = (
MTex.lslide(tex_span[0], slid_spans),
MTex.rslide(tex_span[1], slid_spans)
)
if result[0] >= result[1]:
return None
return result
## Parser
def init_parser(self) -> None:
self.additional_substrings = self.get_additional_substrings()
self.full_span = self.get_full_span()
self.backslash_indices = self.get_backslash_indices()
self.left_brace_indices, self.right_brace_indices = \
self.get_left_and_right_indices()
self.script_char_spans = self.get_script_char_spans()
self.skipped_indices = self.get_skipped_indices()
self.script_spans = self.get_script_spans()
self.script_content_spans = self.get_script_content_spans()
self.double_braces_spans = self.get_double_braces_spans()
self.stripped_substrings = self.get_stripped_substrings()
self.specified_substrings = self.get_specified_substrings()
self.specified_spans, self.extended_specified_spans = \
self.get_specified_spans()
self.tex_span_list = self.get_tex_span_list()
self.extended_tex_span_list = self.get_extended_tex_span_list()
self.isolated_substrings = self.get_isolated_substrings()
self.containing_labels_dict = self.get_containing_labels_dict()
self.color_cmd_repl_items = self.get_color_cmd_repl_items()
self.span_repl_dict = self.get_span_repl_dict()
def get_additional_substrings(self) -> list[str]:
return list(it.chain(
self.tex_to_color_map.keys(),
self.isolate
))
def get_full_span(self) -> tuple[int, int]:
@property
def full_span(self) -> tuple[int, int]:
return (0, len(self.string))
def get_backslash_indices(self) -> list[int]:
@property
def backslash_indices(self) -> list[int]:
# Newlines (`\\`) don't count.
return [
match_obj.end() - 1
@ -264,7 +293,9 @@ class MTex(_TexSVG):
if len(match_obj.group()) % 2 == 1
]
def get_left_and_right_indices(self) -> list[tuple[int, int]]:
def get_left_and_right_brace_indices(
self
) -> tuple[list[tuple[int, int]], list[tuple[int, int]]]:
string = self.string
indices = list(filter(
lambda index: index - 1 not in self.backslash_indices,
@ -287,60 +318,69 @@ class MTex(_TexSVG):
right_brace_indices.append(index)
if left_brace_indices_stack:
raise ValueError("Missing '}' inserted")
# `right_brace_indices` is already sorted.
return left_brace_indices, right_brace_indices
def get_script_char_spans(self) -> list[tuple[int, int]]:
@property
def left_brace_indices(self) -> list[tuple[int, int]]:
return self.get_left_and_right_brace_indices()[0]
@property
def right_brace_indices(self) -> list[tuple[int, int]]:
return self.get_left_and_right_brace_indices()[1]
@property
def skipped_spans(self) -> list[tuple[int, int]]:
return [
match_obj.span()
for match_obj in re.finditer(r"(\s*)[_^]\s*", self.string)
if match_obj.group(1)
or match_obj.start() - 1 not in self.backslash_indices
for match_obj in re.finditer(r"\s*([_^])\s*|(\s+)", self.string)
if match_obj.group(2) is not None
or match_obj.start(1) - 1 not in self.backslash_indices
]
def get_skipped_indices(self) -> list[int]:
return sorted(remove_list_redundancies([
match_obj.start()
for match_obj in re.finditer(r"\s", self.string)
] + list(it.chain(*[
range(*script_char_span)
for script_char_span in self.script_char_spans
]))))
def lstrip_span(self, index: int) -> int:
return self.lstrip(index, self.skipped_spans)
def get_script_spans(self) -> list[tuple[int, int]]:
string = self.string
def rstrip_span(self, index: int) -> int:
return self.rstrip(index, self.skipped_spans)
def strip_span(self, index: int) -> int:
return self.strip(index, self.skipped_spans)
@property
def script_char_spans(self) -> list[tuple[int, int]]:
return list(filter(
lambda tex_span: self.string[slice(*tex_span)].strip(),
self.skipped_spans
))
@property
def script_content_spans(self) -> list[tuple[int, int]]:
result = []
brace_indices_dict = dict(zip(
self.left_brace_indices, self.right_brace_indices
))
for char_begin, span_begin in self.script_char_spans:
for _, span_begin in self.script_char_spans:
if span_begin in brace_indices_dict.keys():
span_end = brace_indices_dict[span_begin] + 1
else:
pattern = re.compile(r"[a-zA-Z0-9]|\\[a-zA-Z]+")
match_obj = pattern.match(string, pos=span_begin)
match_obj = pattern.match(self.string, pos=span_begin)
if not match_obj:
script_name = {
"_": "subscript",
"^": "superscript"
}[script_char]
log.warning(
raise ValueError(
f"Unclear {script_name} detected while parsing. "
"Please use braces to clarify"
)
continue
span_end = match_obj.end()
result.append((char_begin, span_end))
result.append((span_begin, span_end))
return result
def get_script_content_spans(self) -> list[tuple[int, int]]:
return [
(script_char_span[1], script_span[1])
for script_char_span, script_span in zip(
self.script_char_spans, self.script_spans
)
]
def get_double_braces_spans(self) -> list[tuple[int, int]]:
@property
def double_braces_spans(self) -> list[tuple[int, int]]:
# Match paired double braces (`{{...}}`).
result = []
reversed_brace_indices_dict = dict(zip(
@ -348,7 +388,7 @@ class MTex(_TexSVG):
))
skip = False
for prev_right_index, right_index in self.get_neighbouring_pairs(
sorted(reversed_brace_indices_dict.keys())
self.right_brace_indices
):
if skip:
skip = False
@ -363,74 +403,64 @@ class MTex(_TexSVG):
skip = True
return result
def get_stripped_substrings(self) -> list[str]:
result = remove_list_redundancies([
substr.strip()
for substr in self.additional_substrings
])
@property
def additional_substrings(self) -> list[str]:
result = remove_list_redundancies(list(it.chain(
self.tex_to_color_map.keys(),
self.isolate
)))
if "" in result:
result.remove("")
return result
def get_specified_substrings(self) -> list[str]:
return remove_list_redundancies([
self.string[slice(*double_braces_span)]
for double_braces_span in self.double_braces_spans
] + list(filter(
lambda s: s in self.string,
self.stripped_substrings
)))
def get_specified_spans(
def get_tex_span_lists(
self
) -> tuple[list[tuple[int, int]], list[tuple[int, int]]]:
tex_spans = sorted(remove_list_redundancies([
self.full_span,
*self.double_braces_spans,
*[
match_obj.span()
for substr in self.stripped_substrings
for match_obj in re.finditer(re.escape(substr), self.string)
]
]), key=lambda t: (t[0], -t[1]))
result = []
extended_result = []
script_spans_dict = dict(self.script_spans)
reversed_script_spans_dict = dict([
script_span[::-1] for script_span in self.script_spans
script_content_spans = self.script_content_spans
script_spans = [
(script_char_span[0], script_content_span[1])
for script_char_span, script_content_span in zip(
self.script_char_spans, script_content_spans
)
]
tex_spans = remove_list_redundancies([
self.full_span,
*self.double_braces_spans,
*filter(lambda stripped_span: stripped_span is not None, [
self.strip_span(match_obj.span())
for substr in self.additional_substrings
for match_obj in re.finditer(re.escape(substr), self.string)
]),
*script_content_spans
])
for tex_span in tex_spans:
if tex_span in self.script_content_spans:
if tex_span in script_content_spans:
result.append(tex_span)
extended_result.append(tex_span)
continue
span_begin, span_end = tex_span
extended_span_end = span_end
while span_end in reversed_script_spans_dict.keys():
span_end = reversed_script_spans_dict[span_end]
while extended_span_end in script_spans_dict.keys():
extended_span_end = script_spans_dict[extended_span_end]
specified_span = (span_begin, span_end)
extended_specified_span = (span_begin, extended_span_end)
if span_begin >= span_end:
shrinked_span = (span_begin, self.rslide(span_end, script_spans))
extended_span = (span_begin, self.lslide(span_end, script_spans))
if shrinked_span[0] >= shrinked_span[1]:
continue
if extended_specified_span in result:
if shrinked_span in result:
continue
result.append(specified_span)
extended_result.append(extended_specified_span)
result.append(shrinked_span)
extended_result.append(extended_span)
return result, extended_result
def get_tex_span_list(self) -> list[tuple[int, int]]:
return self.specified_spans + self.script_content_spans
@property
def tex_span_list(self) -> list[tuple[int, int]]:
return self.get_tex_span_lists()[0]
def get_extended_tex_span_list(self) -> list[tuple[int, int]]:
return self.extended_specified_spans + self.script_content_spans
@property
def extended_tex_span_list(self) -> list[tuple[int, int]]:
return self.get_tex_span_lists()[1]
def get_isolated_substrings(self) -> list[str]:
return remove_list_redundancies([
self.string[slice(*tex_span)]
for tex_span in self.tex_span_list
])
def get_containing_labels_dict(self) -> dict[tuple[int, int], list[int]]:
@property
def containing_labels_dict(self) -> dict[tuple[int, int], list[int]]:
tex_span_list = self.tex_span_list
result = {
tex_span: []
@ -449,40 +479,46 @@ class MTex(_TexSVG):
"Partially overlapping substrings detected: "
f"'{string_0}' and '{string_1}'"
)
result[self.full_span] = list(range(len(tex_span_list)))
result[self.full_span] = list(range(-1, len(tex_span_list)))
return result
def get_color_cmd_repl_items(self) -> list[tuple[tuple[int, int], str]]:
color_related_command_items = [
("color", 1, ""),
("textcolor", 1, ""),
("pagecolor", 1, "\\pagecolor{white}"),
("colorbox", 1, "\\colorbox{white}"),
("fcolorbox", 2, "\\fcolorbox{white}{white}"),
]
@property
def color_cmd_repl_items(self) -> list[tuple[tuple[int, int], str]]:
color_related_command_dict = {
"color": (1, False),
"textcolor": (1, False),
"pagecolor": (1, True),
"colorbox": (1, True),
"fcolorbox": (2, True),
}
result = []
string = self.string
backslash_indices = self.backslash_indices
left_indices = self.left_brace_indices
brace_indices_dict = dict(zip(
self.left_brace_indices, self.right_brace_indices
))
for cmd_name, n_braces, repl_str in color_related_command_items:
pattern = cmd_name + r"(?![a-zA-Z])"
for match_obj in re.finditer(pattern, string):
span_begin, span_end = match_obj.span()
span_begin -= 1
if span_begin not in backslash_indices:
continue
for _ in range(n_braces):
left_index = min(filter(
lambda index: index >= span_end, left_indices
))
span_end = brace_indices_dict[left_index] + 1
result.append(((span_begin, span_end), repl_str))
right_brace_indices = self.right_brace_indices
pattern = "".join([
r"\\",
"(",
"|".join(color_related_command_dict.keys()),
")",
r"(?![a-zA-Z])"
])
for match_obj in re.finditer(pattern, self.string):
span_begin, cmd_end = match_obj.span()
if span_begin not in backslash_indices:
continue
cmd_name = match_obj.group(1)
n_braces, substitute_cmd = color_related_command_dict[cmd_name]
span_end = right_brace_indices[self.find_region_index(
cmd_end, right_brace_indices
) + n_braces - 1] + 1
if substitute_cmd:
repl_str = "\\" + cmd_name + n_braces * "{white}"
else:
repl_str = ""
result.append(((span_begin, span_end), repl_str))
return result
def get_span_repl_dict(self) -> dict[tuple[int, int], str]:
@property
def span_repl_dict(self) -> dict[tuple[int, int], str]:
indices, _, _, cmd_strings = zip(*sorted([
(
tex_span[flag],
@ -490,7 +526,7 @@ class MTex(_TexSVG):
-tex_span[1 - flag],
("{{" + self.get_color_command(label), "}}")[flag]
)
for label, tex_span in enumerate(self.tex_span_list)
for label, tex_span in enumerate(self.extended_tex_span_list)
for flag in range(2)
]))
result = {
@ -502,7 +538,17 @@ class MTex(_TexSVG):
result.update(self.color_cmd_repl_items)
return result
def get_labelled_string(self) -> str:
@property
def plain_string(self) -> str:
return "".join([
"{{",
self.get_color_command(self.color_to_label(self.base_color)),
self.string,
"}}"
])
@property
def labelled_string(self) -> str:
if not self.span_repl_dict:
return self.string
@ -545,7 +591,6 @@ class MTex(_TexSVG):
ordered_containing_labels[1:], ordered_span_begins[1:]
)
]
string_span_begins.insert(0, ordered_span_begins[0])
string_span_ends = [
next_begin if next_label in containing_labels else curr_end
for next_begin, next_label, containing_labels, curr_end in zip(
@ -553,49 +598,72 @@ class MTex(_TexSVG):
ordered_containing_labels[:-1], ordered_span_ends[:-1]
)
]
string_span_ends.append(ordered_span_ends[-1])
string_spans = list(zip(
(ordered_span_begins[0], *string_span_begins),
(*string_span_ends, ordered_span_ends[-1])
))
string = self.string
left_indices = self.left_brace_indices
right_indices = self.right_brace_indices
skipped_indices = sorted(it.chain(
self.skipped_indices,
left_indices,
right_indices
))
left_brace_indices = self.left_brace_indices
right_brace_indices = self.right_brace_indices
slid_spans = self.skipped_spans + [
(index, index + 1)
for index in left_brace_indices + right_brace_indices
]
result = []
for span_begin, span_end in zip(string_span_begins, string_span_ends):
while span_begin in skipped_indices:
span_begin += 1
if span_begin >= span_end:
result.append("")
for str_span in string_spans:
str_span = self.strip_span(str_span)
if str_span is None:
continue
while span_end - 1 in skipped_indices:
span_end -= 1
unclosed_left_brace = 0
unclosed_right_brace = 0
for index in range(span_begin, span_end):
if index in left_indices:
unclosed_left_brace += 1
elif index in right_indices:
if unclosed_left_brace == 0:
unclosed_right_brace += 1
str_span = self.slide(str_span, slid_spans)
if str_span is None:
continue
unclosed_left_braces = 0
unclosed_right_braces = 0
for index in range(*str_span):
if index in left_brace_indices:
unclosed_left_braces += 1
elif index in right_brace_indices:
if unclosed_left_braces == 0:
unclosed_right_braces += 1
else:
unclosed_left_brace -= 1
unclosed_left_braces -= 1
result.append("".join([
unclosed_right_brace * "{",
string[span_begin:span_end],
unclosed_left_brace * "}"
unclosed_right_braces * "{",
string[slice(*str_span)],
unclosed_left_braces * "}"
]))
return result
@property
def specified_substrings(self) -> list[str]:
return remove_list_redundancies([
self.string[slice(*double_braces_span)]
for double_braces_span in self.double_braces_spans
] + list(filter(
lambda s: s in self.string,
self.additional_substrings
)))
def get_specified_substrings(self) -> list[str]:
return self.specified_substrings
@property
def isolated_substrings(self) -> list[str]:
return remove_list_redundancies([
self.string[slice(*tex_span)]
for tex_span in self.tex_span_list
])
def get_isolated_substrings(self) -> list[str]:
return self.isolated_substrings
## Selector
def find_span_components_of_custom_span(
self,
custom_span: tuple[int, int]
) -> list[tuple[int, int]] | None:
skipped_indices = self.skipped_indices
tex_span_choices = sorted(filter(
lambda tex_span: all([
tex_span[0] >= custom_span[0],
@ -606,13 +674,13 @@ class MTex(_TexSVG):
# Choose spans that reach the farthest.
tex_span_choices_dict = dict(tex_span_choices)
span_begin, span_end = custom_span
result = []
span_begin, span_end = custom_span
span_begin = self.rstrip_span(span_begin)
span_end = self.rstrip_span(span_end)
while span_begin != span_end:
span_begin = self.lstrip_span(span_begin)
if span_begin not in tex_span_choices_dict.keys():
if span_begin in skipped_indices:
span_begin += 1
continue
return None
next_begin = tex_span_choices_dict[span_begin]
result.append((span_begin, next_begin))
@ -640,7 +708,7 @@ class MTex(_TexSVG):
return VGroup(*[
self.get_part_by_custom_span(match_obj.span())
for match_obj in re.finditer(
re.escape(tex.strip()), self.string
re.escape(tex), self.string
)
])