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 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 findMathNode(self):
w = LatexWalker(self.equation)
(nodelist, pos, len_) = w.get_latex_nodes(pos=0)
return self._findMathNodeInList(nodelist)
def traverseLatex(self, text):
w = LatexWalker(text)
(nodelist, pos, len_) = w.get_latex_nodes(pos=0)
self._traverseLatex(nodelist)
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 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
# Ignore comment nodes
if isinstance(node, LatexCommentNode):
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)