def excise(args, parser):
logger = colorlog.getLogger('tacl')
tokenizer = utils.get_tokenizer(args)
corpus = tacl.Corpus(args.corpus, tokenizer)
with open(args.ngrams) as fh:
ngrams = [line.strip() for line in fh.readlines()]
# It is no issue if the output directory already exists; it is a
# reasonable use case to create an excised corpus from multiple
# excise operations.
try:
os.mkdir(args.output)
except FileExistsError:
pass
for work in args.works:
# It is worth warning about writing in existing work
# directories, however, since that might be unintended. Do not
# prevent this, however, since it is a reasonable use case.
try:
os.mkdir(os.path.join(args.output, work))
except FileExistsError:
logger.warn(constants.EXCISE_OVERWRITE_WORK_WARNING.format(work))
for witness in corpus.get_witnesses(work):
path = os.path.join(args.output, witness.get_filename())
content = witness.excise(ngrams, args.replacement)
with open(path, 'w') as fh:
fh.write(content)
def highlight_text(args, parser):
"""Outputs the result of highlighting a text."""
tokenizer = utils.get_tokenizer(args)
corpus = utils.get_corpus(args)
output_dir = os.path.abspath(args.output)
if os.path.exists(output_dir):
parser.exit(status=3,
message='Output directory already exists, '
'aborting.\n')
os.makedirs(output_dir, exist_ok=True)
if args.ngrams:
if args.label is None or len(args.label) != len(args.ngrams):
parser.error('There must be as many labels as there are files '
'of n-grams')
report = tacl.NgramHighlightReport(corpus, tokenizer)
ngrams = []
for ngram_file in args.ngrams:
ngrams.append(utils.get_ngrams(ngram_file))
minus_ngrams = []
if args.minus_ngrams:
minus_ngrams = utils.get_ngrams(args.minus_ngrams)
report.generate(args.output, args.base_name, ngrams, args.label,
minus_ngrams)
else:
report = tacl.ResultsHighlightReport(corpus, tokenizer)
report.generate(args.output, args.base_name, args.results)
def results(args, parser):
if args.results == '-':
results_fh = io.TextIOWrapper(sys.stdin.buffer,
encoding='utf-8',
newline='')
else:
results_fh = open(args.results, 'r', encoding='utf-8', newline='')
tokenizer = utils.get_tokenizer(args)
results = tacl.Results(results_fh, tokenizer)
if args.extend:
corpus = tacl.Corpus(args.extend, tokenizer)
results.extend(corpus)
if args.bifurcated_extend:
if not args.bifurcated_extend_size:
parser.error('The bifurcated extend option requires that the '
'--max-be-count option also be supplied')
corpus = tacl.Corpus(args.bifurcated_extend, tokenizer)
results.bifurcated_extend(corpus, args.bifurcated_extend_size)
if args.reduce:
results.reduce()
if args.reciprocal:
results.reciprocal_remove()
if args.excise:
results.excise(args.excise)
if args.zero_fill:
corpus = tacl.Corpus(args.zero_fill, tokenizer)
results.zero_fill(corpus)
if args.ngrams:
with open(args.ngrams, encoding='utf-8') as fh:
ngrams = fh.read().split()
results.prune_by_ngram(ngrams)
if args.min_works or args.max_works:
results.prune_by_work_count(args.min_works, args.max_works)
if args.min_size or args.max_size:
results.prune_by_ngram_size(args.min_size, args.max_size)
if args.min_count or args.max_count:
results.prune_by_ngram_count(args.min_count, args.max_count)
if args.min_count_work or args.max_count_work:
results.prune_by_ngram_count_per_work(args.min_count_work,
args.max_count_work)
if args.remove:
results.remove_label(args.remove)
if args.sort:
results.sort()
# Run format-changing operations last.
if args.add_label_count:
results.add_label_count()
if args.add_label_work_count:
results.add_label_work_count()
if args.group_by_ngram:
catalogue = tacl.Catalogue()
catalogue.load(args.group_by_ngram)
results.group_by_ngram(catalogue.ordered_labels)
if args.group_by_witness:
results.group_by_witness()
if args.collapse_witnesses:
results.collapse_witnesses()
results.csv(sys.stdout)
def align_results(args, parser):
if args.results == "-":
results = io.TextIOWrapper(sys.stdin.buffer, encoding="utf-8", newline="")
else:
results = open(args.results, "r", encoding="utf-8", newline="")
tokenizer = utils.get_tokenizer(args)
corpus = tacl.Corpus(args.corpus, tokenizer)
report = tacl.SequenceReport(corpus, tokenizer, results)
report.generate(args.output, args.minimum)
def ngram_diff(args, parser):
"""Outputs the results of performing a diff query."""
store = utils.get_data_store(args)
corpus = utils.get_corpus(args)
catalogue = utils.get_catalogue(args.catalogue)
tokenizer = utils.get_tokenizer(args)
store.validate(corpus, catalogue)
if args.asymmetric:
store.diff_asymmetric(catalogue, args.asymmetric, tokenizer, sys.stdout)
else:
store.diff(catalogue, tokenizer, sys.stdout)
def align_results(args, parser):
if args.results == '-':
results = io.TextIOWrapper(sys.stdin.buffer,
encoding='utf-8',
newline='')
else:
results = open(args.results, 'r', encoding='utf-8', newline='')
tokenizer = utils.get_tokenizer(args)
corpus = tacl.Corpus(args.corpus, tokenizer)
report = tacl.SequenceReport(corpus, tokenizer, results)
report.generate(args.output, args.minimum)
def ngram_diff(args, parser):
"""Outputs the results of performing a diff query."""
store = utils.get_data_store(args)
corpus = utils.get_corpus(args)
catalogue = utils.get_catalogue(args)
tokenizer = utils.get_tokenizer(args)
store.validate(corpus, catalogue)
if args.asymmetric:
store.diff_asymmetric(catalogue, args.asymmetric, tokenizer,
sys.stdout)
else:
store.diff(catalogue, tokenizer, sys.stdout)
def main():
parser = generate_parser()
args = parser.parse_args()
if hasattr(args, 'verbose'):
utils.configure_logging(args.verbose, logger)
store = utils.get_data_store(args)
corpus = utils.get_corpus(args)
catalogue = utils.get_catalogue(args)
tokenizer = utils.get_tokenizer(args)
check_catalogue(catalogue, args.label)
store.validate(corpus, catalogue)
output_dir = os.path.abspath(args.output)
if os.path.exists(output_dir):
logger.warning('Output directory already exists; any results therein '
'will be reused rather than regenerated.')
os.makedirs(output_dir, exist_ok=True)
report = tacl.JitCReport(store, corpus, tokenizer)
report.generate(output_dir, catalogue, args.label)
def main():
parser = generate_parser()
args = parser.parse_args()
if hasattr(args, 'verbose'):
utils.configure_logging(args.verbose, logger)
store = utils.get_data_store(args)
corpus = utils.get_corpus(args)
catalogue = utils.get_catalogue(args.catalogue)
tokenizer = utils.get_tokenizer(args)
check_catalogue(catalogue, args.label)
store.validate(corpus, catalogue)
output_dir = os.path.abspath(args.output)
if os.path.exists(output_dir):
logger.warning('Output directory already exists; any results therein '
'will be reused rather than regenerated.')
os.makedirs(output_dir, exist_ok=True)
report = tacl.JitCReport(store, corpus, tokenizer)
report.generate(output_dir, catalogue, args.label)
def results(args, parser):
if args.results == "-":
results_fh = io.TextIOWrapper(sys.stdin.buffer, encoding="utf-8", newline="")
else:
results_fh = open(args.results, "r", encoding="utf-8", newline="")
tokenizer = utils.get_tokenizer(args)
results = tacl.Results(results_fh, tokenizer)
if args.extend:
corpus = tacl.Corpus(args.extend, tokenizer)
results.extend(corpus)
if args.bifurcated_extend:
if not args.bifurcated_extend_size:
parser.error("The bifurcated extend option requires that the " "--max-be-count option also be supplied")
corpus = tacl.Corpus(args.bifurcated_extend, tokenizer)
results.bifurcated_extend(corpus, args.bifurcated_extend_size)
if args.reduce:
results.reduce()
if args.reciprocal:
results.reciprocal_remove()
if args.zero_fill:
if not args.catalogue:
parser.error("The zero-fill option requires that the -c option " "also be supplied.")
corpus = tacl.Corpus(args.zero_fill, tokenizer)
catalogue = utils.get_catalogue(args.catalogue)
results.zero_fill(corpus, catalogue)
if args.ngrams:
with open(args.ngrams, encoding="utf-8") as fh:
ngrams = fh.read().split()
results.prune_by_ngram(ngrams)
if args.min_works or args.max_works:
results.prune_by_work_count(args.min_works, args.max_works)
if args.min_size or args.max_size:
results.prune_by_ngram_size(args.min_size, args.max_size)
if args.min_count or args.max_count:
results.prune_by_ngram_count(args.min_count, args.max_count)
if args.min_count_work or args.max_count_work:
results.prune_by_ngram_count_per_work(args.min_count_work, args.max_count_work)
if args.remove:
results.remove_label(args.remove)
if args.sort:
results.sort()
results.csv(sys.stdout)
def highlight_text(args, parser):
"""Outputs the result of highlighting a text."""
tokenizer = utils.get_tokenizer(args)
corpus = utils.get_corpus(args)
output_dir = os.path.abspath(args.output)
if os.path.exists(output_dir):
parser.exit(status=3, message="Output directory already exists, " "aborting.\n")
os.makedirs(output_dir, exist_ok=True)
if args.ngrams:
if args.label is None or len(args.label) != len(args.ngrams):
parser.error("There must be as many labels as there are files " "of n-grams")
report = tacl.NgramHighlightReport(corpus, tokenizer)
ngrams = []
for ngram_file in args.ngrams:
ngrams.append(utils.get_ngrams(ngram_file))
minus_ngrams = []
if args.minus_ngrams:
minus_ngrams = utils.get_ngrams(args.minus_ngrams)
report.generate(args.output, args.base_name, ngrams, args.label, minus_ngrams)
else:
report = tacl.ResultsHighlightReport(corpus, tokenizer)
report.generate(args.output, args.base_name, args.results)
def collapse_witnesses(args):
results = tacl.Results(args.results, utils.get_tokenizer(args))
results.collapse_witnesses()
results.csv(sys.stdout)
def label_count(args):
results = tacl.Results(args.results, utils.get_tokenizer(args))
results.add_label_count()
results.csv(sys.stdout)
def supplied_diff(args, parser):
labels = args.labels
results = args.supplied
store = utils.get_data_store(args)
tokenizer = utils.get_tokenizer(args)
store.diff_supplied(results, labels, tokenizer, sys.stdout)
def generate_statistics(args, parser):
corpus = utils.get_corpus(args)
tokenizer = utils.get_tokenizer(args)
report = tacl.StatisticsReport(corpus, tokenizer, args.results)
report.generate_statistics()
report.csv(sys.stdout)
def generate_statistics(args, parser):
corpus = utils.get_corpus(args)
tokenizer = utils.get_tokenizer(args)
report = tacl.StatisticsReport(corpus, tokenizer, args.results)
report.generate_statistics()
report.csv(sys.stdout)
def supplied_diff(args, parser):
labels = args.labels
results = args.supplied
store = utils.get_data_store(args)
tokenizer = utils.get_tokenizer(args)
store.diff_supplied(results, labels, tokenizer, sys.stdout)
