def test_handle_unclosed_environments():
extractor = EquationExtractor()
equations = list(extractor.parse("main.tex", "$x + \\hbox{\\begin{equation}y}$"))
assert len(equations) == 1
equation = equations[0]
assert equation.start == 0
assert equation.end == 30
def test_extract_equation_environment_with_argument():
extractor = EquationExtractor()
equations = list(extractor.parse("main.tex", "\\begin{array}{c}x\\end{array}"))
assert len(equations) == 1
equation = equations[0]
assert equation.content_start == 16
def annotate_symbols_and_equations_for_file(
tex: str, tex_path: RelativePath, symbols: SymbolDict,
characters: CharacterDict) -> Tuple[str, Set[str]]:
# Extract all equations
equation_extractor = EquationExtractor()
equations = list(equation_extractor.parse(tex_path, tex))
# Group symbols by equation ID
symbols_by_equation_id = _group_by_equation(symbols)
# Create a list of annotations
annotations: List[Annotation] = []
symbol_tex: Set[str] = set()
for equation in equations:
equation_id = EquationId(tex_path, equation.i)
equation_symbols = symbols_by_equation_id.get(equation_id, [])
equation_annotations = _create_annotations_for_equation(
tex, equation, equation_id, equation_symbols, characters)
annotations.extend(equation_annotations.annotations)
symbol_tex.update(equation_annotations.symbol_tex)
# Annotate the TeX
annotated_tex = tex
annotations_reverse_order = sorted(annotations,
key=lambda a: a.position,
reverse=True)
for annotation in annotations_reverse_order:
position = annotation.position
annotated_tex = (annotated_tex[:position] + annotation.text +
annotated_tex[position:])
return annotated_tex, symbol_tex
def adjust_color_positions(entity: SerializableEntity) -> CharacterRange:
"""
Color commands sometimes introduce unwanted space when added right before or after an equation.
One solution is to put color commands right inside the equation.
"""
term = cast(Term, entity)
equation_extractor = EquationExtractor()
equations = list(equation_extractor.parse(entity.tex_path, term.tex))
if len(equations) == 0:
return CharacterRange(term.start, term.end)
# If the term starts with an equation, move the coloring command inside the equation.
adjusted_start = term.start
first_equation = min(equations, key=lambda e: e.start)
first_nonspace = re.search("\S", term.tex)
if first_nonspace is not None:
if first_nonspace.start(0) == first_equation.start:
adjusted_start = term.start + first_equation.content_start
# If the term ends with an equation, move the coloring command inside the equation.
adjusted_end = term.end
last_equation = max(equations, key=lambda e: e.end)
last_nonspace = re.search("\S(?=\s*$)", term.tex)
if last_nonspace is not None:
if last_nonspace.end(0) == last_equation.end:
adjusted_end = term.start + last_equation.content_end
return CharacterRange(adjusted_start, adjusted_end)
def test_extract_equation_from_double_dollar_signs():
extractor = EquationExtractor()
equations = list(extractor.parse("main.tex", "$$x$$"))
assert len(equations) == 1
equation = equations[0]
assert equation.start == 0
assert equation.end == 5
def test_extract_equation_from_star_environment():
extractor = EquationExtractor()
equations = list(extractor.parse("main.tex", "\\begin{equation*}x\\end{equation*}"))
assert len(equations) == 1
equation = equations[0]
assert equation.start == 0
assert equation.end == 33
def test_extract_equation_from_brackets():
extractor = EquationExtractor()
equations = list(extractor.parse("main.tex", "\\[x + y\\]"))
assert len(equations) == 1
equation = equations[0]
assert equation.start == 0
assert equation.content_start == 2
assert equation.end == 9
def test_extract_nested_equations():
extractor = EquationExtractor()
equations = list(
extractor.parse("main.tex",
"$x + \\hbox{\\begin{equation}y\\end{equation}}$"))
assert len(equations) == 2
outer = next(filter(lambda e: e.start == 0, equations))
assert outer.end == 44
inner = next(filter(lambda e: e.start == 11, equations))
assert inner.end == 42
def test_extract_equation_from_dollar_sign():
extractor = EquationExtractor()
equations = list(extractor.parse("main.tex", "$x + y$"))
assert len(equations) == 1
equation = equations[0]
assert equation.start == 0
assert equation.content_start == 1
assert equation.end == 7
assert equation.content_tex == "x + y"
assert equation.tex == "$x + y$"
def test_extract_equation_from_equation_environment():
extractor = EquationExtractor()
equations = list(extractor.parse("main.tex", "\\begin{equation}x\\end{equation}"))
assert len(equations) == 1
equation = equations[0]
assert equation.start == 0
assert equation.content_start == 16
assert equation.end == 31
assert equation.content_tex == "x"
assert equation.tex == "\\begin{equation}x\\end{equation}"
def test_ignore_escaped_dollar_sign():
extractor = EquationExtractor()
equations = list(extractor.parse("main.tex", "\\$\\$"))
assert len(equations) == 0
def test_dont_extract_equation_from_command_argument_brackets():
extractor = EquationExtractor()
equations = list(
extractor.parse("main.tex", "\\documentclass[11pt]{article}"))
assert len(equations) == 0
def test_extract_equation_from_dollar_sign_after_newline():
# This pattern was observed in arXiv paper 1703.00102.
extractor = EquationExtractor()
equations = list(extractor.parse("main.tex", r"\\$x$"))
assert len(equations) == 1