mirror of
https://github.com/3b1b/manim.git
synced 2025-11-14 08:17:43 +00:00
Some refactors
- Split out `_TexParser` class - Replace `math_mode` parameter with `tex_environment` - Fix the bug that braces following even number of backslashes aren't matched
This commit is contained in:
Michael W
GitHub
parent
155839bde9
commit
0de303d5e0
1 changed files with 210 additions and 197 deletions
|
|
@ -18,6 +18,10 @@ SCALE_FACTOR_PER_FONT_POINT = 0.001
|
|||
tex_hash_to_mob_map = {}
|
||||
|
||||
|
||||
def _get_neighbouring_pairs(iterable):
|
||||
return list(adjacent_pairs(iterable))[:-1]
|
||||
|
||||
|
||||
class _LabelledTex(SVGMobject):
|
||||
CONFIG = {
|
||||
"height": None,
|
||||
|
|
@ -65,6 +69,185 @@ class _TexSpan(object):
|
|||
]) + ")"
|
||||
|
||||
|
||||
class _TexParser(object):
|
||||
def __init__(self, mtex):
|
||||
self.tex_string = mtex.tex_string
|
||||
strings_to_break_up = remove_list_redundancies([
|
||||
*mtex.isolate, *mtex.tex_to_color_map.keys(), mtex.tex_string
|
||||
])
|
||||
if "" in strings_to_break_up:
|
||||
strings_to_break_up.remove("")
|
||||
unbreakable_commands = mtex.unbreakable_commands
|
||||
|
||||
self.tex_spans_dict = {}
|
||||
self.current_label = 0
|
||||
self.break_up_by_braces()
|
||||
self.break_up_by_scripts()
|
||||
self.break_up_by_additional_strings(strings_to_break_up)
|
||||
self.merge_unbreakable_commands(unbreakable_commands)
|
||||
self.analyse_containing_labels()
|
||||
|
||||
@staticmethod
|
||||
def label_to_color_tuple(n):
|
||||
# Get a unique color different from black,
|
||||
# or the svg file will not include the color information.
|
||||
rgb = n + 1
|
||||
rg, b = divmod(rgb, 256)
|
||||
r, g = divmod(rg, 256)
|
||||
return r, g, b
|
||||
|
||||
@staticmethod
|
||||
def contains(span_0, span_1):
|
||||
return span_0[0] <= span_1[0] and span_1[1] <= span_0[1]
|
||||
|
||||
def add_tex_span(self, span_tuple, script_type=0, label=-1):
|
||||
if script_type == 0:
|
||||
# Should be additionally labelled.
|
||||
label = self.current_label
|
||||
self.current_label += 1
|
||||
|
||||
tex_span = _TexSpan(script_type, label)
|
||||
self.tex_spans_dict[span_tuple] = tex_span
|
||||
|
||||
def break_up_by_braces(self):
|
||||
tex_string = self.tex_string
|
||||
span_tuples = []
|
||||
left_brace_indices = []
|
||||
for match_obj in re.finditer(r"(\\*)(\{|\})", tex_string):
|
||||
# Braces following even numbers of backslashes are counted.
|
||||
if len(match_obj.group(1)) % 2 == 1:
|
||||
continue
|
||||
if match_obj.group(2) == "{":
|
||||
left_brace_index = match_obj.span(2)[0]
|
||||
left_brace_indices.append(left_brace_index)
|
||||
else:
|
||||
left_brace_index = left_brace_indices.pop()
|
||||
right_brace_index = match_obj.span(2)[1]
|
||||
span_tuples.append((left_brace_index, right_brace_index))
|
||||
if left_brace_indices:
|
||||
self.raise_tex_parsing_error()
|
||||
|
||||
self.paired_braces_tuples = span_tuples
|
||||
for span_tuple in span_tuples:
|
||||
self.add_tex_span(span_tuple)
|
||||
|
||||
def break_up_by_scripts(self):
|
||||
tex_string = self.tex_string
|
||||
brace_indices_dict = dict(self.tex_spans_dict.keys())
|
||||
for match_obj in re.finditer(r"((?<!\\)(_|\^)\s*)|(\s+(_|\^)\s*)", tex_string):
|
||||
script_type = 1 if "_" in match_obj.group() else 2
|
||||
token_begin, token_end = match_obj.span()
|
||||
if token_end in brace_indices_dict:
|
||||
content_span = (token_end, brace_indices_dict[token_end])
|
||||
else:
|
||||
content_match_obj = re.match(r"\w|\\[a-zA-Z]+", tex_string[token_end:])
|
||||
if not content_match_obj:
|
||||
self.raise_tex_parsing_error()
|
||||
content_span = tuple([
|
||||
index + token_end for index in content_match_obj.span()
|
||||
])
|
||||
self.add_tex_span(content_span)
|
||||
label = self.tex_spans_dict[content_span].label
|
||||
self.add_tex_span(
|
||||
(token_begin, content_span[1]),
|
||||
script_type=script_type,
|
||||
label=label
|
||||
)
|
||||
|
||||
def break_up_by_additional_strings(self, strings_to_break_up):
|
||||
tex_string = self.tex_string
|
||||
all_span_tuples = []
|
||||
for string in strings_to_break_up:
|
||||
# Only matches non-crossing strings.
|
||||
for match_obj in re.finditer(re.escape(string), tex_string):
|
||||
all_span_tuples.append(match_obj.span())
|
||||
|
||||
script_spans_dict = dict([
|
||||
span_tuple[::-1]
|
||||
for span_tuple, tex_span in self.tex_spans_dict.items()
|
||||
if tex_span.script_type != 0
|
||||
])
|
||||
for span_begin, span_end in all_span_tuples:
|
||||
if span_end in script_spans_dict.values():
|
||||
# Deconstruct spans with subscripts & superscripts.
|
||||
while span_end in script_spans_dict:
|
||||
span_end = script_spans_dict[span_end]
|
||||
if span_begin >= span_end:
|
||||
continue
|
||||
span_tuple = (span_begin, span_end)
|
||||
if span_tuple not in self.tex_spans_dict:
|
||||
self.add_tex_span(span_tuple)
|
||||
|
||||
def merge_unbreakable_commands(self, unbreakable_commands):
|
||||
tex_string = self.tex_string
|
||||
command_merge_spans = []
|
||||
brace_indices_dict = dict(self.paired_braces_tuples)
|
||||
# Braces leading by `unbreakable_commands` shouldn't be marked.
|
||||
for command in unbreakable_commands:
|
||||
for match_obj in re.finditer(re.escape(command), tex_string):
|
||||
merge_begin_index = match_obj.span()[1]
|
||||
merge_end_index = merge_begin_index
|
||||
if merge_end_index not in brace_indices_dict:
|
||||
continue
|
||||
while merge_end_index in brace_indices_dict:
|
||||
merge_end_index = brace_indices_dict[merge_end_index]
|
||||
command_merge_spans.append((merge_begin_index, merge_end_index))
|
||||
|
||||
self.tex_spans_dict = {
|
||||
span_tuple: tex_span
|
||||
for span_tuple, tex_span in self.tex_spans_dict.items()
|
||||
if all([
|
||||
not _TexParser.contains(merge_span, span_tuple)
|
||||
for merge_span in command_merge_spans
|
||||
])
|
||||
}
|
||||
|
||||
def analyse_containing_labels(self):
|
||||
for span_0, tex_span_0 in self.tex_spans_dict.items():
|
||||
if tex_span_0.script_type != 0:
|
||||
continue
|
||||
for span_1, tex_span_1 in self.tex_spans_dict.items():
|
||||
if _TexParser.contains(span_1, span_0):
|
||||
tex_span_1.containing_labels.append(tex_span_0.label)
|
||||
|
||||
def get_labelled_expression(self):
|
||||
tex_string = self.tex_string
|
||||
if not self.tex_spans_dict:
|
||||
return tex_string
|
||||
|
||||
indices_with_labels = sorted([
|
||||
(span_tuple[i], i, span_tuple[1 - i], tex_span.label)
|
||||
for span_tuple, tex_span in self.tex_spans_dict.items()
|
||||
if tex_span.script_type == 0
|
||||
for i in range(2)
|
||||
], key=lambda t: (t[0], -t[1], -t[2]))
|
||||
# Add one more item to ensure all the substrings are joined.
|
||||
indices_with_labels.append((len(tex_string), 0, 0, 0))
|
||||
|
||||
result = tex_string[: indices_with_labels[0][0]]
|
||||
index_with_label_pairs = _get_neighbouring_pairs(indices_with_labels)
|
||||
for index_with_label, next_index_with_label in index_with_label_pairs:
|
||||
index, flag, _, label = index_with_label
|
||||
next_index, *_ = next_index_with_label
|
||||
# Adding one more pair of braces will help maintain the glyghs of tex file...
|
||||
if flag == 0:
|
||||
color_tuple = _TexParser.label_to_color_tuple(label)
|
||||
result += "".join([
|
||||
"{{",
|
||||
"\\color[RGB]",
|
||||
"{",
|
||||
",".join(map(str, color_tuple)),
|
||||
"}"
|
||||
])
|
||||
else:
|
||||
result += "}}"
|
||||
result += tex_string[index : next_index]
|
||||
return result
|
||||
|
||||
def raise_tex_parsing_error(self):
|
||||
raise ValueError(f"Failed to parse tex: \"{self.tex_string}\"")
|
||||
|
||||
|
||||
class MTex(VMobject):
|
||||
CONFIG = {
|
||||
"fill_opacity": 1.0,
|
||||
|
|
@ -74,7 +257,7 @@ class MTex(VMobject):
|
|||
"height": None,
|
||||
"organize_left_to_right": False,
|
||||
"alignment": "\\centering",
|
||||
"math_mode": True,
|
||||
"tex_environment": "align*",
|
||||
"isolate": [],
|
||||
"unbreakable_commands": ["\\begin", "\\end"],
|
||||
"tex_to_color_map": {},
|
||||
|
|
@ -82,14 +265,13 @@ class MTex(VMobject):
|
|||
|
||||
def __init__(self, tex_string, **kwargs):
|
||||
super().__init__(**kwargs)
|
||||
tex_string = tex_string.strip("\n")
|
||||
# Prevent from passing an empty string.
|
||||
if not tex_string:
|
||||
tex_string = "\\quad"
|
||||
self.tex_string = tex_string
|
||||
self.parse_tex()
|
||||
self.tex_string = MTex.modify_tex_string(tex_string)
|
||||
|
||||
full_tex = self.get_tex_file_body()
|
||||
tex_parser = _TexParser(self)
|
||||
self.tex_spans_dict = tex_parser.tex_spans_dict
|
||||
|
||||
new_tex = tex_parser.get_labelled_expression()
|
||||
full_tex = self.get_tex_file_body(new_tex)
|
||||
hash_val = hash(full_tex)
|
||||
if hash_val not in tex_hash_to_mob_map:
|
||||
with display_during_execution(f"Writing \"{tex_string}\""):
|
||||
|
|
@ -111,160 +293,22 @@ class MTex(VMobject):
|
|||
self.organize_submobjects_left_to_right()
|
||||
|
||||
@staticmethod
|
||||
def label_to_color_tuple(n):
|
||||
# Get a unique color different from black,
|
||||
# or the svg file will not include the color information.
|
||||
rgb = n + 1
|
||||
rg, b = divmod(rgb, 256)
|
||||
r, g = divmod(rg, 256)
|
||||
return r, g, b
|
||||
def modify_tex_string(tex_string):
|
||||
result = tex_string.strip("\n")
|
||||
# Prevent from passing an empty string.
|
||||
if not result:
|
||||
result = "\\quad"
|
||||
return result
|
||||
|
||||
@staticmethod
|
||||
def get_neighbouring_pairs(iterable):
|
||||
return list(adjacent_pairs(iterable))[:-1]
|
||||
|
||||
@staticmethod
|
||||
def contains(span_0, span_1):
|
||||
return span_0[0] <= span_1[0] and span_1[1] <= span_0[1]
|
||||
|
||||
def add_tex_span(self, span_tuple, script_type=0, label=-1):
|
||||
if script_type == 0:
|
||||
# Should be additionally labelled.
|
||||
label = self.current_label
|
||||
self.current_label += 1
|
||||
|
||||
tex_span = _TexSpan(script_type, label)
|
||||
self.tex_spans_dict[span_tuple] = tex_span
|
||||
|
||||
def parse_tex(self):
|
||||
self.tex_spans_dict = {}
|
||||
self.current_label = 0
|
||||
self.break_up_by_braces()
|
||||
self.break_up_by_scripts()
|
||||
self.break_up_by_additional_strings()
|
||||
self.merge_unbreakable_strings()
|
||||
self.analyse_containing_labels()
|
||||
|
||||
def break_up_by_braces(self):
|
||||
tex_string = self.tex_string
|
||||
span_tuples = []
|
||||
left_brace_indices = []
|
||||
for match_obj in re.finditer(r"(?<!\\)(\{|\})", tex_string):
|
||||
if match_obj.group() == "{":
|
||||
left_brace_index = match_obj.span()[0]
|
||||
left_brace_indices.append(left_brace_index)
|
||||
else:
|
||||
left_brace_index = left_brace_indices.pop()
|
||||
right_brace_index = match_obj.span()[1]
|
||||
span_tuples.append((left_brace_index, right_brace_index))
|
||||
if left_brace_indices:
|
||||
self.raise_tex_parsing_error()
|
||||
|
||||
self.paired_braces_tuples = span_tuples
|
||||
for span_tuple in span_tuples:
|
||||
self.add_tex_span(span_tuple)
|
||||
|
||||
def break_up_by_scripts(self):
|
||||
tex_string = self.tex_string
|
||||
brace_indices_dict = dict(self.tex_spans_dict.keys())
|
||||
script_spans = []
|
||||
for match_obj in re.finditer(r"((?<!\\)(_|\^)\s*)|(\s+(_|\^)\s*)", tex_string):
|
||||
script_type = 1 if "_" in match_obj.group() else 2
|
||||
token_begin, token_end = match_obj.span()
|
||||
if token_end in brace_indices_dict:
|
||||
content_span = (token_end, brace_indices_dict[token_end])
|
||||
else:
|
||||
content_match_obj = re.match(r"\w|\\[a-zA-Z]+", tex_string[token_end:])
|
||||
if not content_match_obj:
|
||||
self.raise_tex_parsing_error()
|
||||
content_span = tuple([
|
||||
index + token_end for index in content_match_obj.span()
|
||||
])
|
||||
self.add_tex_span(content_span)
|
||||
subscript_span = (token_begin, content_span[1])
|
||||
script_spans.append(subscript_span)
|
||||
label = self.tex_spans_dict[content_span].label
|
||||
self.add_tex_span(
|
||||
subscript_span,
|
||||
script_type=script_type,
|
||||
label=label
|
||||
)
|
||||
self.script_spans = script_spans
|
||||
|
||||
def break_up_by_additional_strings(self):
|
||||
additional_strings_to_break_up = remove_list_redundancies([
|
||||
*self.isolate, *self.tex_to_color_map.keys(), self.tex_string
|
||||
])
|
||||
if "" in additional_strings_to_break_up:
|
||||
additional_strings_to_break_up.remove("")
|
||||
if not additional_strings_to_break_up:
|
||||
return
|
||||
|
||||
tex_string = self.tex_string
|
||||
all_span_tuples = []
|
||||
for string in additional_strings_to_break_up:
|
||||
# Only matches non-crossing strings.
|
||||
for match_obj in re.finditer(re.escape(string), tex_string):
|
||||
all_span_tuples.append(match_obj.span())
|
||||
|
||||
script_spans_dict = dict([
|
||||
span_tuple[::-1]
|
||||
for span_tuple in self.script_spans
|
||||
])
|
||||
for span_begin, span_end in all_span_tuples:
|
||||
if span_end in script_spans_dict.values():
|
||||
# Deconstruct spans with subscripts & superscripts.
|
||||
while span_end in script_spans_dict:
|
||||
span_end = script_spans_dict[span_end]
|
||||
if span_begin >= span_end:
|
||||
continue
|
||||
span_tuple = (span_begin, span_end)
|
||||
if span_tuple not in self.tex_spans_dict:
|
||||
self.add_tex_span(span_tuple)
|
||||
|
||||
def merge_unbreakable_strings(self):
|
||||
tex_string = self.tex_string
|
||||
command_merge_spans = []
|
||||
brace_indices_dict = dict(self.paired_braces_tuples)
|
||||
# Braces leading by `unbreakable_commands` shouldn't be marked.
|
||||
for command in self.unbreakable_commands:
|
||||
for match_obj in re.finditer(re.escape(command), tex_string):
|
||||
merge_begin_index = match_obj.span()[1]
|
||||
merge_end_index = merge_begin_index
|
||||
if merge_end_index not in brace_indices_dict:
|
||||
continue
|
||||
while merge_end_index in brace_indices_dict:
|
||||
merge_end_index = brace_indices_dict[merge_end_index]
|
||||
command_merge_spans.append((merge_begin_index, merge_end_index))
|
||||
|
||||
if not command_merge_spans:
|
||||
return
|
||||
self.tex_spans_dict = {
|
||||
span_tuple: tex_span
|
||||
for span_tuple, tex_span in self.tex_spans_dict.items()
|
||||
if all([
|
||||
not MTex.contains(merge_span, span_tuple)
|
||||
for merge_span in command_merge_spans
|
||||
def get_tex_file_body(self, new_tex):
|
||||
if self.tex_environment:
|
||||
new_tex = "\n".join([
|
||||
f"\\begin{{{self.tex_environment}}}",
|
||||
new_tex,
|
||||
f"\\end{{{self.tex_environment}}}"
|
||||
])
|
||||
}
|
||||
|
||||
def analyse_containing_labels(self):
|
||||
for span_0, tex_span_0 in self.tex_spans_dict.items():
|
||||
if tex_span_0.script_type != 0:
|
||||
continue
|
||||
for span_1, tex_span_1 in self.tex_spans_dict.items():
|
||||
if MTex.contains(span_1, span_0):
|
||||
tex_span_1.containing_labels.append(tex_span_0.label)
|
||||
|
||||
def raise_tex_parsing_error(self):
|
||||
raise ValueError(f"Failed to parse tex: \"{self.tex_string}\"")
|
||||
|
||||
def get_tex_file_body(self):
|
||||
new_tex = self.get_modified_expression()
|
||||
|
||||
if self.math_mode:
|
||||
new_tex = "\n".join(["\\begin{align*}", new_tex, "\\end{align*}"])
|
||||
new_tex = "\n".join([self.alignment, new_tex])
|
||||
if self.alignment:
|
||||
new_tex = "\n".join([self.alignment, new_tex])
|
||||
|
||||
tex_config = get_tex_config()
|
||||
return tex_config["tex_body"].replace(
|
||||
|
|
@ -272,40 +316,6 @@ class MTex(VMobject):
|
|||
new_tex
|
||||
)
|
||||
|
||||
def get_modified_expression(self):
|
||||
tex_string = self.tex_string
|
||||
if not self.tex_spans_dict:
|
||||
return tex_string
|
||||
|
||||
indices_with_labels = sorted([
|
||||
(span_tuple[i], i, span_tuple[1 - i], tex_span.label)
|
||||
for span_tuple, tex_span in self.tex_spans_dict.items()
|
||||
if tex_span.script_type == 0
|
||||
for i in range(2)
|
||||
], key=lambda t: (t[0], -t[1], -t[2]))
|
||||
# Add one more item to ensure all the substrings are joined.
|
||||
indices_with_labels.append((len(tex_string), 0, 0, 0))
|
||||
|
||||
result = tex_string[: indices_with_labels[0][0]]
|
||||
index_with_label_pairs = MTex.get_neighbouring_pairs(indices_with_labels)
|
||||
for index_with_label, next_index_with_label in index_with_label_pairs:
|
||||
index, flag, _, label = index_with_label
|
||||
next_index, *_ = next_index_with_label
|
||||
# Adding one more pair of braces will help maintain the glyghs of tex file...
|
||||
if flag == 0:
|
||||
color_tuple = MTex.label_to_color_tuple(label)
|
||||
result += "".join([
|
||||
"{{",
|
||||
"\\color[RGB]",
|
||||
"{",
|
||||
",".join(map(str, color_tuple)),
|
||||
"}"
|
||||
])
|
||||
else:
|
||||
result += "}}"
|
||||
result += tex_string[index : next_index]
|
||||
return result
|
||||
|
||||
def build_submobjects(self):
|
||||
if not self.submobjects:
|
||||
return
|
||||
|
|
@ -339,10 +349,11 @@ class MTex(VMobject):
|
|||
# This function sorts the submobjects of scripts in the order of tex given.
|
||||
index_and_span_list = sorted([
|
||||
(index, span_tuple)
|
||||
for span_tuple in self.script_spans
|
||||
for span_tuple, tex_span in self.tex_spans_dict.items()
|
||||
if tex_span.script_type != 0
|
||||
for index in span_tuple
|
||||
])
|
||||
index_and_span_pair = MTex.get_neighbouring_pairs(index_and_span_list)
|
||||
index_and_span_pair = _get_neighbouring_pairs(index_and_span_list)
|
||||
for index_and_span_0, index_and_span_1 in index_and_span_pair:
|
||||
index_0, span_tuple_0 = index_and_span_0
|
||||
index_1, span_tuple_1 = index_and_span_1
|
||||
|
|
@ -497,12 +508,14 @@ class MTex(VMobject):
|
|||
def print_tex_strings_of_submobjects(self):
|
||||
# For debugging
|
||||
# Work with `index_labels()`
|
||||
print("\n")
|
||||
print(f"Submobjects of \"{self.get_tex()}\":")
|
||||
for i, submob in enumerate(self.submobjects):
|
||||
print(f"{i}: \"{submob.get_tex()}\"")
|
||||
print("\n")
|
||||
|
||||
|
||||
class MTexText(MTex):
|
||||
CONFIG = {
|
||||
"math_mode": False,
|
||||
"tex_environment": None,
|
||||
}
|
||||
|
|
|
|||
Loading…
Add table
Reference in a new issue