#!/usr/bin/env python2 import subprocess import sys import re import inspect import datetime import string from collections import defaultdict, Counter import nltk try: nltk.pos_tag('Just trying to see if the NLTK dataset is installed') except LookupError: nltk.download('maxent_treebank_pos_tagger') try: nltk.word_tokenize('test') except LookupError: nltk.download('punkt') try: from clint.textui import colored except: class Passthrough(object): def __getattr__(self, name): return lambda x: x colored = Passthrough() print "=== For colored output, install clint (via 'sudo pip install clint') ===" PREPOSITIONS = ["a", "abaft", "aboard", "about", "above", "absent", "across", "afore", "after", "against", "along", "alongside", "amid", "amidst", "among", "amongst", "an", "apropos", "around", "as", "aside", "astride", "at", "athwart", "atop", "barring", "before", "behind", "below", "beneath", "beside", "besides", "between", "betwixt", "beyond", "but", "by", "circa", "concerning", "despite", "down", "during", "except", "excluding", "failing", "following", "for", "from", "given", "in", "including", "inside", "into", "lest", "like", "mid", "midst", "minus", "modulo", "near", "next", "notwithstanding", "of", "off", "on", "onto", "opposite", "out", "outside", "over", "pace", "past", "per", "plus", "pro", "qua", "regarding", "round", "sans", "save", "since", "than", "through,", "throughout,", "till", "times", "to", "toward", "towards", "under", "underneath", "unlike", "until", "unto", "up", "upon", "versus", "via", "vice", "with", "within", "without", "worth", "through"] # Obtained with (avoiding the dependency): # from nltk.corpus import stopwords # stopwords.words("english") STOPWORDS = ['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', 'she', 'her', 'hers', 'herself', 'it', 'its', 'itself', 'they', 'them', 'their', 'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', 'these', 'those', 'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', 'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', 'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more', 'most', 'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don', 'should', 'now'] CONJUNCTIONS = ["and", "but", "or", "yet", "for", "nor", "so"] class Paper(object): ########################################################################## # INTERNAL STUFF (you can ignore these functions) ########################################################################## def __init__(self, filenames): self.__filenames = filenames self.errors = 0 self.__clear_caches() def __clear_caches(self): self.__text = {} self.__latex_text = None @staticmethod def __flatten_paragraphs(text): '''Given a text where paragraphs are separated by one or more empty lines, it puts every paragraph in a single, separate line. Example: I like sushi and pie. I ride horses. Becomes: I like sushi and pie. I ride horses ''' return '\n'.join(paragraph.replace('\n', ' ') for paragraph in re.split("\n(\s*\n)+", text) if paragraph.strip()) def __run_all_with_prefix(self, prefix): # Switch filename and clear caches for filename in self.__filenames: self.filename = filename self.__clear_caches() # Call all functions for name in filter(lambda n: n.startswith(prefix), dir(self)): attribute = getattr(self, name) if inspect.ismethod(attribute): attribute() def _run_all_checks(self): self.__run_all_with_prefix('check_') def _run_all_tests(self): self.__filenames = ["TEST"] self.__run_all_with_prefix('test_') def _format_re_match(self, m, text): start_of_sentence = max(text.rfind('\n', 0, m.start()) + 1, 0) end_of_sentence = text.find('\n', m.end()), len(text) if end_of_sentence == -1: end_of_sentence = len(text) a_string_start = max(start_of_sentence, m.start() - 10) a_string_end = min(end_of_sentence, m.end() + 10) a_string = text[a_string_start : m.start()] a_string += colored.yellow(text[m.start() : m.end()]) a_string += text[m.end() : a_string_end] to_return = a_string.split('\n', 1)[0] return to_return.replace('\r',' ').replace('\n',' ') ########################################################################## # FUCTIONS THAT ARE RELEVANT FOR CHECKS WRITERS ########################################################################## def get_latex_text(self): ''' Returns the complete paper, with each paragraph on a single line. No latex/tex command is stripped ''' if self.__latex_text != None: return self.__latex_text else: with open(self.filename) as f: text = f.read() self.__latex_text = self.__flatten_paragraphs(text) return self.__latex_text _IGNORED_ENVIRONMENTS = ("array", "eqnarray", "equation", "figure", "mathmatica", "picture", "table", "verbatim", "lstlisting") def get_text(self, ignored_environments=None): ''' Returns the textual content of the tex files, with latex/tex enviroments stripped. You can control the enviroments to strip via the 'ignored_environments' argument: if you don't, the default ones will be stripped''' if ignored_environments == None: ignored_environments = Paper._IGNORED_ENVIRONMENTS try: return self.__text[ignored_environments] except: # Cleanup annoying things text = self.get_latex_text() text = re.sub(r'\\cite{[^}]*}', '', text) text = re.sub(r'\\-', '', text) # Run it through detex p = subprocess.Popen(["detex", "-l", "-n", "-e", ','.join(ignored_environments)], stdin=subprocess.PIPE, stdout=subprocess.PIPE) p.stdin.write(text) text = p.communicate()[0] p.stdin.close() self.__text[ignored_environments] = self.__flatten_paragraphs( text) return self.__text[ignored_environments] def perform_test(self, function, expected_errors_num): self.errors = 0 function() assert expected_errors_num == self.errors def print_issue(self, message, match=None, text=None): if text == None: text = self.get_text() message = colored.red(message) if match and text: example = self._format_re_match(match, text) else: example = '' print "%30s - %s: %s" % (colored.green(self.filename), colored.red(message), example) ########################################################################## # CHECKS ########################################################################## def check_exempli_gratia_without_comma(self): for m in re.finditer(r'e\.g\.[^,]', self.get_text(), re.MULTILINE): self.print_issue("E.G. without comma", m) self.errors += 1 def test__check_exempli_gratia_without_comma(self): self.get_text = lambda: "e.g. a pony \n e.g. what?, e.g., cool!" self.perform_test(self.check_exempli_gratia_without_comma, 2) ########################################################################## def check_id_est_without_comma(self): for m in re.finditer(r'i\.e\.[^,]', self.get_text(), re.MULTILINE): self.print_issue("I.E. without comma", m) self.errors += 1 def test__check_id_est_without_comma(self): self.get_text = lambda: "i.e. a pony \n i.e. what?, i.e., cool!" self.perform_test(self.check_id_est_without_comma, 2) ########################################################################## def check_quotes(self): for m in re.finditer(r'"', self.get_text(), re.MULTILINE): self.print_issue('"hello" should be ``hello\'\'', m) self.errors += 1 def test__check_quotes(self): self.get_text = lambda: '"this is not ok" ``but this is\'\'' self.perform_test(self.check_quotes, 2) ########################################################################## def check_citing_with_unbreakeable_spaces(self): for keyword in ["cite", "ref"]: for m in re.finditer(r'(\W?)(^|\s)+\\' + keyword + r'\s*{', self.get_latex_text(), re.MULTILINE): if m.group(1) in [',', '&']: continue self.print_issue('use hello~\%s{ instead of hello \%s{' % (keyword, keyword), m, self.get_latex_text()) self.errors += 1 def test__check_citing_with_unbreakeable_spaces(self): self.get_latex_text = lambda: r'citing: wrong \cite{ciao} - right~\cite{ciao}' self.perform_test(self.check_citing_with_unbreakeable_spaces, 1) self.get_latex_text = lambda: r'refs done wrong \ref{ciao} - right~\ref{ciao}' self.perform_test(self.check_citing_with_unbreakeable_spaces, 1) self.get_latex_text = lambda: r'hello& \cite{ciao}' self.perform_test(self.check_citing_with_unbreakeable_spaces, 0) self.get_latex_text = lambda: r', \ref{ciao}' self.perform_test(self.check_citing_with_unbreakeable_spaces, 0) ########################################################################## def check_variations_of_word_spellings(self): words = defaultdict(Counter) for word in re.findall(r'\b\S+\b', self.get_text(), re.MULTILINE): word_alphanum = re.sub("[^a-zA-Z0-9_']+", '', word).lower() words[word_alphanum].update([word]) for _, spellings_counter in words.iteritems(): variations = len(spellings_counter.keys()) total_appereances = sum(spellings_counter.values()) if variations > 1: if len(set(w[1:] for w in spellings_counter.keys())) == 1: # FIXME: for now, if it's just a case mismatch on the first # letter, skip continue normalized_word = spellings_counter.keys()[0].lower() if normalized_word in STOPWORDS + PREPOSITIONS + CONJUNCTIONS: # Ignore common words continue # Ignore numbers try: float(normalized_word) continue except ValueError: pass self.print_issue('This word has multiple spellings: %s' % ( dict(spellings_counter)), None) self.errors += 1 def test__check_variations_of_word_spellings(self): self.get_text = lambda: (r'I has a cybercriminal. I had a cyber-criminal. ' r'or was it a CyberCriminal?') self.perform_test(self.check_variations_of_word_spellings, 1) self.get_text = lambda: (r'no strange words here, however put. ' r'However, is that true?') self.perform_test(self.check_variations_of_word_spellings, 0) self.get_text = lambda: (r'It matters the it factor ') self.perform_test(self.check_variations_of_word_spellings, 0) self.get_text = lambda: (r'1.6 16 and other fancy numbers.') self.perform_test(self.check_variations_of_word_spellings, 0) ########################################################################## def check_commas_in_numbers(self): # We also check in tables text = self.get_text(ignored_environments=tuple( set(Paper._IGNORED_ENVIRONMENTS) - set('table'))) for m in re.finditer('(^|[^\w\-])\d{4}', text, re.MULTILINE): if text[m.start():m.start() + 1] in string.punctuation: continue try: number = int(text[m.start():m.end()]) except: number = 0 if number not in range(1990, datetime.date.today().year + 2): self.errors += 1 self.print_issue('Put commas in numbers over 1,000', m) # This is the correct rule, but nobody follows it # for m in re.finditer('(^|[^\w\-])\d{5}', text, re.MULTILINE): # self.print_issue('Put commas in numbers over 10,000', m) # self.errors += 1 # for m in re.finditer('[^\d]\d,\d{3}[^,]', text, re.MULTILINE): # self.print_issue("Don't put commas in numbers under 10,000", m) # self.errors += 1 def test__check_commas_in_numbers(self): def get_text(*args, **kwargs): return text self.get_text = get_text text = r'10000 cats eat 10,000 mice' self.perform_test(self.check_commas_in_numbers, 1) text = r'9999 cats eat 9,999 mice' self.perform_test(self.check_commas_in_numbers, 1) text = r'1000 cats eat 999,999 mice' self.perform_test(self.check_commas_in_numbers, 1) text = r'project N10000, grant CNS-20000' self.perform_test(self.check_commas_in_numbers, 0) text = r'In 2001, we ate spaghetti' self.perform_test(self.check_commas_in_numbers, 0) ########################################################################## def check_commas_after_quotes(self): for m in re.finditer("''\s*,", self.get_text(), re.MULTILINE): self.print_issue("Convert ``hello'', => ``hello,''", m) self.errors += 1 def test__check_commas_after_quotes(self): self.get_text = lambda: r"``flower'', should be ``flower,''" self.perform_test(self.check_commas_after_quotes, 1) ########################################################################## def check_always_capitalize(self): for reg in ["internet", "javascript"]: for m in re.finditer(r"\b{0}".format(reg), self.get_text(), re.MULTILINE): self.print_issue("Always capitalize", m) self.errors += 1 def test__check_always_capitalize(self): self.get_text = lambda: r"internet" self.perform_test(self.check_always_capitalize, 1) self.get_text = lambda: r"testinternet" self.perform_test(self.check_always_capitalize, 0) ########################################################################## # # def check_comma_before_that(self): # for m in re.finditer(",\s+that", # self.get_text(), # re.MULTILINE): # phrase_start = max([ # self.get_text().rfind(c, 0, m.start()) # for c in ['\n', '.', ':', ';'] # ] + [0]) # phrase = self.get_text()[phrase_start:m.start() + 1] # if len([c for c in phrase if c == ',']) % 2 == 0: # # An even number of commas found, skipping # continue # self.print_issue("Do not put a comma before 'that'", m) # self.errors += 1 # # def test__check_comma_before_that(self): # self.get_text = lambda: r"I like cats, that eat mice" # self.perform_test(self.check_comma_before_that, 1) # self.get_text = lambda: r"I like cats that eat mice" # self.perform_test(self.check_comma_before_that, 0) # # ########################################################################## # # def check_comma_before_which(self): # for m in re.finditer("[^,'*\s]\s+which", # self.get_text(), # re.MULTILINE): # word_before_start = self.get_text().rfind(' ', 0, m.start()) # word_before = re.search("\w+", self.get_text()[ # word_before_start + 1: # m.start() + 1]).group() # if word_before in PREPOSITIONS + CONJUNCTIONS: # continue # if word_before.endswith('ing') or word_before.endswith('ly'): # continue # # More expensive analysis: is the word before a verb? # phrase_start = max([ # self.get_text().rfind(c, 0, m.start()) # for c in ['\n', '.', ':', ';'] # ] + [0]) # phrase = self.get_text()[phrase_start:m.start() + 1] # word_before_kind = filter(lambda x: x[0] == word_before, nltk.pos_tag( # nltk.word_tokenize(phrase)))[0][1] # if word_before_kind.startswith('VB'): # continue # self.print_issue("Put a comma before 'which'", m) # self.errors += 1 # # def test__check_comma_before_which(self): # self.get_text = lambda: r"I like that cat, which eat mice" # self.perform_test(self.check_comma_before_which, 0) # self.get_text = lambda: r"I like that cat which eat mice" # self.perform_test(self.check_comma_before_which, 1) # self.get_text = lambda: r"I know which cat eat mice" # self.perform_test(self.check_comma_before_which, 0) ########################################################################## def check_word_before_ref_is_capitalized(self): text = self.get_latex_text() for m in re.finditer('\\\\ref', text, re.MULTILINE): word_before_start = max(text.rfind(' ', 0, m.start() - 2), text.rfind('~', 0, m.start() - 2)) word_before = re.findall("\w+", text[word_before_start + 1: m.start() + 1])[-1] if not word_before in ["and"] and not word_before[0].isupper(): self.print_issue(r'Capitalize the word before \ref', m, text) self.errors += 1 def test__check_word_before_ref_is_capitalized(self): self.get_latex_text = lambda: "in Section \\ref{sec}, see Figure \\ref{fig}" self.perform_test(self.check_word_before_ref_is_capitalized, 0) self.get_latex_text = lambda: "in section \\ref{sec}, see figure \\ref{fig}" self.perform_test(self.check_word_before_ref_is_capitalized, 2) self.get_latex_text = lambda: "section \\ref{sec}" self.perform_test(self.check_word_before_ref_is_capitalized, 1) ########################################################################## def check_british_spelling(self): british_spellings = {"acknowledgement": "acknowledgment", "afterwards": "afterward", "arse": "ass", "backwards": "backward", "cancelling": "canceling", "catalogue": "catalog", "centre": "center", "cheque": "check", "colour": "color", "dialogue": "dialog", "favour": "favor", "flavour": "flavor", "forwards": "forward", "grey": "gray", "judgement": "judgment", "labour": "labor", "lustre": "luster", "modelled": "modeled", "revelled": "raveled", "shovelled": "shoveled", "snivelled": "sniveled", "theatre": "theater", "towards": "toward", "travelling": "traveling", "yodelling": "yodeling"} for british, american in british_spellings.iteritems(): for m in re.finditer("[^\w]+" + british + "[^\w]+", self.get_text()): self.print_issue("Don't spell like a bugger (that's british english) -" \ " it's ' "+ american + " ' , not", m) self.errors += 1 def test__check_british_spelling(self): self.get_text = lambda: r"Go to the (centre) of town to pick up the best flavour colour." self.perform_test(self.check_british_spelling, 3) self.get_text = lambda: r"I am an American, therefore I am" self.perform_test(self.check_british_spelling, 0) self.get_text = lambda: r"This cheque, right here, is unacceptable. I'll have to cheque with my manager." self.perform_test(self.check_british_spelling, 2) self.get_text = lambda: r"It is hard to parse this sentence." self.perform_test(self.check_british_spelling, 0) ########################################################################## def check_slang_and_gergal_terms(self): gergal = ['basically'] for w in gergal: for m in re.finditer(w, self.get_text(), re.IGNORECASE): self.print_issue( "This word doesn't sound like it should be in a paper: " + w, m) self.errors += 1 def test__check_slang_and_gergal_terms(self): self.get_text = lambda: r"Basically, this is wat we do" self.perform_test(self.check_slang_and_gergal_terms, 1) ########################################################################## def check_misplelled_standard_phrases(self): mispellings = {"in more details": "in more detail"} for wrong, right in mispellings.iteritems(): for m in re.finditer("[^\w]+" + wrong + "[^\w]+", self.get_text()): self.print_issue("Mispelled standard phrase - ' %s ' " "should be ' %s' in" % (wrong, right), m) self.errors += 1 def test__check_misplelled_standard_phrases(self): self.get_text = lambda: r"I'll discuss this in more details in section" self.perform_test(self.check_misplelled_standard_phrases, 1) ########################################################################## def check_banned_words(self): banned_words = ["is[\s]+comprised[\s]+of", "doesn't", "beside", "won't", "can't"] for banned_word in banned_words: for m in re.finditer("([^\w]+|^)" + banned_word + "[^\w]+", self.get_text(), flags=re.IGNORECASE): self.print_issue("Don't use %s" % banned_word, m) self.errors += 1 def test__check_banned_words(self): self.get_text = lambda: r"Adam is comprised of a brain and a stomach." self.perform_test(self.check_banned_words, 1) self.get_text = lambda: r"Adam comprises a brain and a stomach." self.perform_test(self.check_banned_words, 0) self.get_text = lambda: r"You don't know what that is. Comprised. Of." self.perform_test(self.check_banned_words, 0) self.get_text = lambda: r"Is comprised of blah and bloop." self.perform_test(self.check_banned_words, 1) self.get_text = lambda: r"Adam is awesome. Is comprised of blah and bloop." self.perform_test(self.check_banned_words, 1) self.get_text = lambda: r"Don't do this. I won't, tell anybody." self.perform_test(self.check_banned_words, 1) ########################################################################## def check_repeated_words(self): for m in re.finditer(r"\b(\w+)\W+\1\b", self.get_text(), flags=re.IGNORECASE): if m.group(1).isdigit(): continue self.print_issue("Repeated word '%s'" % m.group(1), m) self.errors += 1 def test__check_repeated_words(self): self.get_text = lambda: r"We use this this and that." self.perform_test(self.check_repeated_words, 1) self.get_text = lambda: r"We use this and that, and this and that too." self.perform_test(self.check_repeated_words, 0) self.get_text = lambda: r"This. This is a sentence Sentence." self.perform_test(self.check_repeated_words, 2) self.get_text = lambda: r"Version 4.4." self.perform_test(self.check_repeated_words, 0) if __name__ == '__main__': if len(sys.argv) < 2: print """Usage: - chrisper *.tex - chrisper test Runs the test suite. """ sys.exit(0) if sys.argv[1] == "test": Paper(sys.argv[2:])._run_all_tests() print colored.green("\n=== ALL TESTS PASSED ===") else: paper = Paper(sys.argv[1:]) paper._run_all_checks() if paper.errors == 0: print colored.green('=== IT LOOKS GOOD TO ME. CONGRATS! ===') else: print colored.yellow("\n=== I'VE FOUND %d ISSUES ===" % paper.errors) sys.exit(1)