def printLatexNodes(self, text):
print("print " + text)
w = LatexWalker(text)
(nodelist, pos, len_) = w.get_latex_nodes(pos=0)
print("len_ " + str(len_))
self._printLatexNodes(nodelist)
def test_input(self):
latex = r'''ABCDEF fdksanfkld safnkd anfklsa
\input{test_input_1.tex}
MORENKFDNSN'''
correct_text = r'''ABCDEF fdksanfkld safnkd anfklsa
hi there! This is an equation:
x + y i = 0
where i is the imaginary unit.
MORENKFDNSN'''
testdir = os.path.realpath(os.path.abspath(os.path.dirname(__file__)))
l2t = LatexNodes2Text()
l2t.set_tex_input_directory(testdir)
output = l2t.nodelist_to_text(LatexWalker(latex).get_latex_nodes()[0])
self.assertEqualUpToWhitespace(output, correct_text)
latex = r'''ABCDEF fdksanfkld safnkd anfklsa
\input{test_input_1}
MORENKFDNSN'''
self.assertEqualUpToWhitespace(
l2t.nodelist_to_text(LatexWalker(latex).get_latex_nodes()[0]),
correct_text)
latex = r'''ABCDEF fdksanfkld safnkd anfklsa
\input{../test_input_1}
MORENKFDNSN'''
correct_text_unsafe = correct_text # as before
correct_text_safe = r'''ABCDEF fdksanfkld safnkd anfklsa
MORENKFDNSN'''
# make sure that the \input{} directive failed to include the file.
l2t = LatexNodes2Text()
l2t.set_tex_input_directory(os.path.join(testdir, 'dummy'))
self.assertEqualUpToWhitespace(
l2t.nodelist_to_text(LatexWalker(latex).get_latex_nodes()[0]),
correct_text_safe)
# but without the strict_input flag, it can access it.
l2t.set_tex_input_directory(os.path.join(testdir, 'dummy'),
strict_input=False)
self.assertEqualUpToWhitespace(
l2t.nodelist_to_text(LatexWalker(latex).get_latex_nodes()[0]),
correct_text_unsafe)
def test_accents(self):
self.assertEqual(
LatexNodes2Text().nodelist_to_text(LatexWalker(r"Fran\c cais").get_latex_nodes()[0]),
'''Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais'''
)
self.assertEqual(
LatexNodes2Text().nodelist_to_text(LatexWalker(r"Fr\'en{\'{e}}tique").get_latex_nodes()[0]),
'''Fr\N{LATIN SMALL LETTER E WITH ACUTE}n\N{LATIN SMALL LETTER E WITH ACUTE}tique'''
)
def test_accents(self):
self.assertEqual(
LatexNodes2Text().nodelist_to_text(
LatexWalker(r"Fran\c cais").get_latex_nodes()[0]),
'''Fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais''')
self.assertEqual(
LatexNodes2Text().nodelist_to_text(
LatexWalker(r"Fr\'en{\'{e}}tique").get_latex_nodes()[0]),
'''Fr\N{LATIN SMALL LETTER E WITH ACUTE}n\N{LATIN SMALL LETTER E WITH ACUTE}tique'''
)
self.assertEqual(
LatexNodes2Text(math_mode='with-delimiters').nodelist_to_text(
LatexWalker(r"$1 \not= 2$").get_latex_nodes()[0]),
'''$1 {} 2$'''.format(
unicodedata.normalize('NFC',
"=\N{COMBINING LONG SOLIDUS OVERLAY}")))
def test_keep_braced_groups(self):
self.assertEqual(
LatexNodes2Text(keep_braced_groups=True)
.nodelist_to_text(LatexWalker(r"\textit{Voil\`a du texte}. Il est \'{e}crit {en fran{\c{c}}ais}")
.get_latex_nodes()[0]),
'''Voil\N{LATIN SMALL LETTER A WITH GRAVE} du texte. Il est \N{LATIN SMALL LETTER E WITH ACUTE}crit {en fran\N{LATIN SMALL LETTER C WITH CEDILLA}ais}'''
)
self.assertEqual(
LatexNodes2Text(keep_braced_groups=True, keep_braced_groups_minlen=4)
.nodelist_to_text(LatexWalker(r"A{XYZ}{ABCD}").get_latex_nodes()[0]),
'''AXYZ{ABCD}'''
)
self.assertEqual(
LatexNodes2Text(keep_braced_groups=True, keep_braced_groups_minlen=0)
.nodelist_to_text(LatexWalker(r"{A}{XYZ}{ABCD}").get_latex_nodes()[0]),
'''{A}{XYZ}{ABCD}'''
)
def test_get_latex_environment(self):
latextext = r'''Text and \`accent and \textbf{bold text} and $\vec b$ more stuff for Fran\c cois
\begin{enumerate}[(i)]
\item Hi there! % here goes a comment
\item[a] Hello! @@@
\end{enumerate}
Indeed thanks to \cite[Lemma 3]{Author}, we know that...
Also: {\itshape some italic text}.
'''
lw = LatexWalker(latextext, tolerant_parsing=False)
p = latextext.find(r'\begin{enumerate}')
self.assertEqual(
lw.get_latex_environment(pos=p, environmentname='enumerate'), (
LatexEnvironmentNode('enumerate', [
LatexCharsNode('\n'),
LatexMacroNode('item', None, [], macro_post_space=' '),
LatexCharsNode('Hi there! '),
LatexCommentNode(' here goes a comment'),
LatexMacroNode('item', LatexGroupNode(
[LatexCharsNode('a')]), []),
LatexCharsNode(' Hello! @@@\n ')
], [LatexGroupNode([LatexCharsNode('(i)')])], []),
p,
latextext.find(r'\end{enumerate}') + len(r'\end{enumerate}') -
p,
))
self.assertEqual(lw.get_latex_environment(pos=p), (
LatexEnvironmentNode('enumerate', [
LatexCharsNode('\n'),
LatexMacroNode('item', None, [], macro_post_space=' '),
LatexCharsNode('Hi there! '),
LatexCommentNode(' here goes a comment'),
LatexMacroNode('item', LatexGroupNode([LatexCharsNode('a')]),
[]),
LatexCharsNode(' Hello! @@@\n ')
], [LatexGroupNode([LatexCharsNode('(i)')])], []),
p,
latextext.find(r'\end{enumerate}') + len(r'\end{enumerate}') - p,
))
with self.assertRaises(LatexWalkerParseError):
dummy = lw.get_latex_environment(pos=p,
environmentname='XYZNFKLD-WRONG')
def test_get_latex_maybe_optional_arg(self):
latextext = r'''Text and \`accent and \textbf{bold text} and $\vec b$ more stuff for Fran\c cois
\begin{enumerate}[(i)]
\item Hi there! % here goes a comment
\item[a] Hello! @@@
\end{enumerate}
Indeed thanks to \cite[Lemma 3]{Author}, we know that...
'''
lw = LatexWalker(latextext, tolerant_parsing=False)
p = latextext.find(r'\textbf') + len(r'\textbf')
self.assertEqual(lw.get_latex_maybe_optional_arg(pos=p), None)
p = latextext.find(r'\cite') + len(r'\cite')
self.assertEqual(lw.get_latex_maybe_optional_arg(pos=p), (
LatexGroupNode([LatexCharsNode('Lemma 3')]),
p,
9,
))
def test_get_latex_braced_group(self):
latextext = r'''Text and \`accent and \textbf{bold text} and $\vec b$ more stuff for Fran\c cois
\begin{enumerate}[(i)]
\item Hi there! % here goes a comment
\item[a] Hello! @@@
\end{enumerate}
Indeed thanks to \cite[Lemma 3]{Author}, we know that...
Also: {\itshape some italic text}.
'''
lw = LatexWalker(latextext, tolerant_parsing=False)
p = latextext.find(r'Also: {') + len(
'Also:') # points on space after 'Also:'
self.assertEqual(lw.get_latex_braced_group(pos=p, brace_type='{'), (
LatexGroupNode([
LatexMacroNode('itshape', None, [], macro_post_space=' '),
LatexCharsNode('some italic text')
]),
p + 1,
len('{\itshape some italic text}'),
))
self.assertEqual(lw.get_latex_braced_group(
pos=p + 1, brace_type='{'), (
LatexGroupNode([
LatexMacroNode('itshape', None, [], macro_post_space=' '),
LatexCharsNode('some italic text')
]),
p + 1,
len('{\itshape some italic text}'),
))
p = latextext.find(r'[(i)]')
self.assertEqual(lw.get_latex_braced_group(pos=p, brace_type='['), (
LatexGroupNode([LatexCharsNode('(i)')]),
p,
5,
))
def test_basic(self):
self.assertEqual(
LatexNodes2Text().nodelist_to_text(LatexWalker(r'\textbf{A}').get_latex_nodes()[0]),
'A'
)
latex = r'''\textit{hi there!} This is {\em an equation}:
\begin{equation}
x + y i = 0
\end{equation}
where $i$ is the imaginary unit.
'''
self.assertEqualUpToWhitespace(
LatexNodes2Text().nodelist_to_text(LatexWalker(latex).get_latex_nodes()[0]),
r'''hi there! This is an equation:
x + y i = 0
where i is the imaginary unit.
'''
)
self.assertEqualUpToWhitespace(
LatexNodes2Text(keep_inline_math=True).nodelist_to_text(LatexWalker(latex).get_latex_nodes()[0]),
r'''hi there! This is an equation:
x + y i = 0
where $i$ is the imaginary unit.
'''
)
self.assertEqual(
LatexNodes2Text().nodelist_to_text(LatexWalker(latex).get_latex_nodes()[0]),
LatexNodes2Text().latex_to_text(latex)
)
def test_basic(self):
self.assertEqual(
LatexNodes2Text().nodelist_to_text(
LatexWalker(r'\textbf{A}').get_latex_nodes()[0]), 'A')
latex = r'''\textit{hi there!} This is {\em an equation}:
\begin{equation}
x + y i = 0
\end{equation}
where $i$ is the ``imaginary unit.''
'''
self.assertEqualUpToWhitespace(
LatexNodes2Text().nodelist_to_text(
LatexWalker(latex).get_latex_nodes()[0]),
u'''hi there! This is an equation:
x + y i = 0
where i is the “imaginary unit.”
''')
self.assertEqualUpToWhitespace(
LatexNodes2Text(math_mode='with-delimiters').nodelist_to_text(
LatexWalker(latex).get_latex_nodes()[0]),
u'''hi there! This is an equation:
\\begin{equation}
x + y i = 0
\\end{equation}
where $i$ is the “imaginary unit.”
''')
self.assertEqual(
LatexNodes2Text().nodelist_to_text(
LatexWalker(latex).get_latex_nodes()[0]),
LatexNodes2Text().latex_to_text(latex))
def test_get_latex_nodes(self):
latextext = r'''Text and \`accent and \textbf{bold text} and $\vec b$ more stuff for Fran\c cois
\begin{enumerate}[(i)]
\item Hi there! % here goes a comment
\item[a] Hello! @@@
\end{enumerate}
Indeed thanks to \cite[Lemma 3]{Author}, we know that...
Also: {\itshape some italic text}.
'''
lw = LatexWalker(latextext, tolerant_parsing=False)
#lw.get_latex_nodes(pos=0,stop_upon_closing_brace=None,stop_upon_end_environment=None,
# stop_upon_closing_mathmode=None)
p = latextext.find('Also: {')
self.assertEqual(lw.get_latex_nodes(pos=p), ([
LatexCharsNode('Also: '),
LatexGroupNode([
LatexMacroNode('itshape', None, [], macro_post_space=' '),
LatexCharsNode('some italic text')
]),
LatexCharsNode('.')
], p, len(latextext) - p - 1)) # trailing '\n' is not included
p = latextext.find('Also: {') + len(
'Also: {') # points inside right after open brace
self.assertEqual(
lw.get_latex_nodes(pos=p, stop_upon_closing_brace='}'), ([
LatexMacroNode('itshape', None, [], macro_post_space=' '),
LatexCharsNode('some italic text')
], p, len('\itshape some italic text}')))
# test our own macro lists etc.
pindeed = latextext.find('Indeed thanks to')
lineindeed = latextext[pindeed:latextext.find('\n', pindeed)]
lw2 = LatexWalker(lineindeed,
tolerant_parsing=False,
macro_dict={'cite': MacrosDef('cite', False, 4)})
self.assertEqual(lw2.get_latex_nodes(pos=0), ([
LatexCharsNode('Indeed thanks to '),
LatexMacroNode('cite', None, [
LatexCharsNode('['),
LatexCharsNode('L'),
LatexCharsNode('e'),
LatexCharsNode('m'),
]),
LatexCharsNode('ma 3]'),
LatexGroupNode([LatexCharsNode('Author')]),
LatexCharsNode(', we know that...'),
], 0, len(lineindeed)))
def test_errors(self):
latextext = get_test_latex_data_with_possible_inconsistencies()
lw = LatexWalker(latextext, tolerant_parsing=False)
with self.assertRaises(LatexWalkerParseError):
dummy = lw.get_latex_nodes()
lwOk = LatexWalker(latextext, tolerant_parsing=True)
# make sure that it goes through without raising:
try:
lwOk.get_latex_nodes()
except LatexWalkerParseError as e:
# should not raise this.
self.fail(
u"get_latex_nodes() raised LatexWalkerParseError, but it shouldn't have in "
u"tolerant parsing mode!\n" + unicode(e))
def test_get_token(self):
latextext = r'''Text \`accent and \textbf{bold text} and $\vec b$ vector \& also Fran\c cois
\begin{enumerate}[(i)]
\item Hi there! % here goes a comment
\item[a] Hello! @@@
\end{enumerate}
'''
lw = LatexWalker(latextext)
self.assertEqual(
lw.get_token(pos=0),
LatexToken(tok='char', arg='T', pos=0, len=1, pre_space=''))
self.assertEqual(
lw.get_token(pos=1),
LatexToken(tok='char', arg='e', pos=1, len=1, pre_space=''))
p = latextext.find(r'\`')
self.assertEqual(
lw.get_token(pos=p),
LatexToken(tok='macro', arg='`', pos=p, len=2, pre_space=''))
p = latextext.find(r'\textbf') - 1 # pre space
self.assertEqual(
lw.get_token(pos=p),
LatexToken(tok='macro',
arg='textbf',
pos=p + 1,
len=7,
pre_space=' '))
p = latextext.find(r'\vec') # post-space
self.assertEqual(
lw.get_token(pos=p),
LatexToken(tok='macro',
arg='vec',
pos=p,
len=5,
pre_space='',
post_space=' '))
p = latextext.find(r'\&') - 1 # pre-space and *no* post-space
self.assertEqual(
lw.get_token(pos=p),
LatexToken(tok='macro',
arg='&',
pos=p + 1,
len=2,
pre_space=' ',
post_space=''))
p = latextext.find(r'\begin')
self.assertEqual(
lw.get_token(pos=p, environments=False),
LatexToken(tok='macro', arg='begin', pos=p, len=6, pre_space=''))
p = latextext.find(r'\begin')
self.assertEqual(
lw.get_token(pos=p),
LatexToken(tok='begin_environment',
arg='enumerate',
pos=p,
len=len(r'\begin{enumerate}'),
pre_space=''))
p = latextext.find(r'@@@') + 3 # pre space to \end
self.assertEqual(
lw.get_token(pos=p),
LatexToken(tok='end_environment',
arg='enumerate',
pos=p + 6,
len=len(r'\end{enumerate}'),
pre_space='\n '))
p = latextext.find(r'%') - 1
self.assertEqual(
lw.get_token(pos=p),
LatexToken(tok='comment',
arg=' here goes a comment',
pos=p + 1,
len=len('% here goes a comment\n'),
pre_space=' ',
post_space='\n'))
p = latextext.find(r'{')
self.assertEqual(
lw.get_token(pos=p),
LatexToken(tok='brace_open', arg='{', pos=p, len=1, pre_space=''))
p = latextext.find(r'}')
self.assertEqual(
lw.get_token(pos=p),
LatexToken(tok='brace_close', arg='}', pos=p, len=1, pre_space=''))
p = latextext.find(r'[')
self.assertEqual(
lw.get_token(pos=p, brackets_are_chars=False),
LatexToken(tok='brace_open', arg='[', pos=p, len=1, pre_space=''))
p = latextext.find(r']')
self.assertEqual(
lw.get_token(pos=p, brackets_are_chars=False),
LatexToken(tok='brace_close', arg=']', pos=p, len=1, pre_space=''))
p = latextext.find(r'$')
self.assertEqual(
lw.get_token(pos=p),
LatexToken(tok='char', arg='$', pos=p, len=1, pre_space=''))
lw2 = LatexWalker(latextext, keep_inline_math=True)
p = latextext.find(r'$')
self.assertEqual(
lw2.get_token(pos=p),
LatexToken(tok='mathmode_inline',
arg='$',
pos=p,
len=1,
pre_space=''))
re_single_quote = re.compile(r"`|'")
def is_symbol(node, LatexMacroNode):
# TODO: improve this function
return len(node.macroname) == 1
if __name__ == "__main__":
# Read latex file
with open(input_tex, "r") as fp:
text = fp.read()
# Parse latex
walker = LatexWalker(text)
# Parse nodes
nodes, pos, length = walker.get_latex_nodes()
# Write output to a file
with open(output_txt, "w") as fp:
# Number of titles
num_titles = 0
# If previous node is a citation
no_new_paragraph = False
# Loop over all nodes
for node in nodes:
# TODO: Implement more rules and process environments recursively
def test2(self):
w = LatexWalker(r"""\[\int_{a}^{b} x^2 \,dx \]""")
(nodelist, pos, len_) = w.get_latex_nodes(pos=0)
for i in range(len(nodelist[0].nodelist)):
print(nodelist[0].nodelist[i])
def traverseLatex(self, text):
w = LatexWalker(text)
(nodelist, pos, len_) = w.get_latex_nodes(pos=0)
self._traverseLatex(nodelist)
def findMathNode(self):
w = LatexWalker(self.equation)
(nodelist, pos, len_) = w.get_latex_nodes(pos=0)
return self._findMathNodeInList(nodelist)
def test_get_latex_expression(self):
latextext = r'''Text and \`accent and \textbf{bold text} and $\vec b$ more stuff for Fran\c cois
\begin{enumerate}[(i)]
\item Hi there! % here goes a comment
\item[a] Hello! @@@
\end{enumerate}
'''
lw = LatexWalker(latextext, tolerant_parsing=True)
self.assertEqual(lw.get_latex_expression(pos=0), (
LatexCharsNode('T'),
0,
1,
))
p = latextext.find(r'\`')
self.assertEqual(lw.get_latex_expression(pos=p), (
LatexMacroNode('`', None, []),
p,
2,
))
p = latextext.find(r'{')
self.assertEqual(lw.get_latex_expression(pos=p), (
LatexGroupNode([LatexCharsNode('bold text')]),
p,
11,
))
p = latextext.find(r'%') # check: correctly skips comments
self.assertEqual(lw.get_latex_expression(pos=p), (
LatexMacroNode('item', None, []),
p + len('% here goes a comment\n'),
5,
))
p = latextext.find(r'%') # check: correctly skips comments
self.assertEqual(lw.get_latex_expression(pos=p), (
LatexMacroNode('item', None, []),
p + len('% here goes a comment\n'),
5,
))
# check correct behavior if directly on brace close
p = latextext.find(r'}')
self.assertEqual(lw.get_latex_expression(pos=p, strict_braces=True), (
LatexCharsNode(''),
p,
0,
))
lw2 = LatexWalker(latextext, tolerant_parsing=False)
self.assertEqual(lw2.get_latex_expression(pos=p, strict_braces=False),
(
LatexCharsNode(''),
p,
0,
))
with self.assertRaises(LatexWalkerParseError):
dummy = lw2.get_latex_expression(pos=p, strict_braces=True)
from pylatexenc.latexwalker import LatexWalker, LatexEnvironmentNode
import pylatexenc
import sys
fname = sys.argv[1]
with open(fname, "r") as f:
tex = f.read()
print(fname)
walker = LatexWalker(tex)
nodelist, pos, leen = walker.get_latex_nodes()
charslist = []
for node in nodelist:
if node.isNodeType(pylatexenc.latexwalker.LatexEnvironmentNode):
for subnode in node.nodelist:
if subnode.isNodeType(pylatexenc.latexwalker.LatexCharsNode):
charslist.append(str(subnode.chars).strip())
fullst = " ".join(charslist).replace("\n"," ")
print(fullst)