def copy_derge_layout(derge_layout):
for f in os.listdir('output'):
no_layout = True
if f in [a.replace('_raw_page_reinserted.txt', '') for a in os.listdir(derge_layout)]:
content = open_file('{}/{}_raw_page_reinserted.txt'.format(derge_layout, f))
reformated = re.sub(r'\n-+', '', content).replace('\\', '')
write_file('output/{}/{}'.format(f, f+'_derge_layout.txt'), reformated)
no_layout = False
if no_layout and f in [a.replace('_with_a.txt', '') for a in os.listdir('../4-a-final_formatting/output/2-0-with_a')]:
content = open_file('../4-a-final_formatting/output/2-0-with_a/{}_with_a.txt'.format(f))
reformated = content.replace('\n', ' ').replace('a', '\n')
write_file('output/{}/{}'.format(f, f + '_derge_lines.txt'), reformated)
def open_ngrams():
ngrams = {}
for i in range(1, 13):
lines = open_file('./resources/kangyur_ngrams/{}-grams_raw.txt'.format(
i)).strip().split('\n')
for line in lines:
parts = line.split(' ')
text = ''.join(parts[:-1]).strip().strip('་')
ngrams[text] = int(parts[-1])
return ngrams
def raw_folder_content(path, num=10):
files_content = {}
for f in os.listdir(path):
full_path = '{}/{}'.format(path, f)
# open file
raw = open_file(full_path)
if num != 0:
raw = first_parts(raw, num)
files_content[full_path] = raw
return files_content
def generate_outputs(text_name,
notes_name,
context,
in_dir=inDir,
out_dir=outDir):
# extract text and reinsert notes
editions = reinsert_notes(open_file(in_dir / text_name),
open_file(in_dir / notes_name).replace(';', ','))
work_name = text_name.split('.')[0].replace(' ', '_')
print(work_name)
generate_editions(editions, out_dir, work_name)
export_unified_structure(editions, work_name)
generate_context_versions(editions,
work_name,
out_dir,
left=context,
right=context)
def reconstruct_version_texts(in_path):
for f in os.listdir(in_path):
text_name = f.strip('_updated_structure.txt')
# create a folder for the layered files if missing
if text_name in os.listdir('output'):
current_out_folder = 'output/' + text_name
# open structure file
from_structure = yaml.load(open_file('{}/{}'.format(in_path, f)))
# reconstruct the editions
editions = reconstruct_edition_versions(from_structure)
# write them in the corresponding folder
for ed, version in editions.items():
version = version.replace('_', ' ') # reconstruct spaces
write_file('{}/{}_{}_layer.txt'.format(current_out_folder, text_name, ed), version)
def process(in_path, file_origin, name_end, out_path):
for f in os.listdir(in_path):
work_name = f.replace(name_end, '')
# raw_content = open_file(file_origin.format(work_name.replace('_', ' ')))
try:
raw_content = open_file(file_origin.format(work_name))
except FileNotFoundError:
continue
content = re.sub(r'\n?[0-9]+\.\s+', '', raw_content)
content = re.sub(r' ', '\n', content)
write_file(out_path[0].format(work_name), content)
content = content.replace('a', '')
write_file(out_path[1].format(work_name), content)
def find_string(in_path, string):
"""wrapper function """
found = []
for f in os.listdir(in_path):
raw = open_file('{}/{}'.format(in_path, f))
raw = clean_string(raw, strip=True, del_returns=True)
# unsegment the files
raw = clean_string(raw, del_dashes=True, del_spaces=True)
if string in raw:
found.append(f)
if found == []:
return 'Nothing found.\n'
else:
return 'The following files were found:\n{}\n'.format('\n'.join(found))
def reinsert_raw(in_path, out_path, patterns):
print('raw reinsertion')
for f in os.listdir(in_path):
work_name = f.replace('_with_a.txt', '')
print(work_name)
if work_name in patterns:
print('\t', work_name)
content = open_file('{}/{}'.format(in_path, f))
lines = deque(content.replace('\n', ' ').split('a'))
pages = []
text_pattern = patterns[work_name][2:]
counter = patterns[work_name][0][1]
side = patterns[work_name][0][2]
# beginning pages
for num in text_pattern[0]:
pages.append(create_page(lines, num, counter, side))
counter, side = increment_counter(counter, side)
# body of the text
while len(lines) > 0:
if len(lines) >= text_pattern[1]:
pages.append(
create_page(lines, text_pattern[1], counter, side))
counter, side = increment_counter(counter, side)
elif text_pattern[2] == len(lines):
pages.append(create_page(lines, len(lines), counter, side))
counter, side = increment_counter(counter, side)
else:
pages.append(create_page(lines, len(lines), counter, side))
counter, side = increment_counter(counter, side)
output = '\n{}\n'.format('-' * 100).join(pages)
write_file(
'{}/{}_raw_page_reinserted.txt'.format(out_path, work_name),
output)
print('2-2!')
else:
out.append(t.content)
else:
out.append(t.content.replace(' ', '_'))
return ' '.join(out)
in_path = 'out'
out_path = 'segmented'
# populate total with the mistakes of all files in in_path
total = defaultdict(list)
for f in sorted(os.listdir(in_path)):
if f.endswith('txt'):
work_name = f.replace('.txt', '')
print(work_name)
content = open_file('{}/{}'.format(in_path, f))
content = rawify(content)
# segmented = PyTib.Segment().segment(content)
pybo_segmented = pybo_segment(content)
mistakes = mistake_conc(pybo_segmented, work_name)
for k, v in mistakes.items():
if not contains_sskrt(k):
total[k].extend(v)
# write individual files for each text, presenting the mistakes in total frequency order
len_ordered_mistakes = sorted(total, key=lambda x: len(total[x]), reverse=True)
for f in os.listdir(in_path):
if f.endswith('txt'):
current_text = f.replace('.txt', '')
# filter mistakes of the current file
output = []
reviewed_path = '../3-b-reviewed_texts'
structure_path = '../3-a-revision_format/output/updated_structure'
limit = False
for f in sorted(os.listdir(reviewed_path)):
print(f)
if f: # == '1-1_ཆོས་ཀྱི་དབྱིངས་སུ་བསྟོད་པ།_DUCKed.csv':
work_name = f.replace('_DUCKed.csv', '')
if work_name == 'N3118':
limit = True
if not limit:
continue
note_choice = parse_decisions(
open_file('{}/{}'.format(reviewed_path, f)))
# parse the file to keep only the decision and the note number
try:
updated_structure = yaml.load(
open_file('{}/{}_updated_structure.txt'.format(
structure_path, work_name)))
except FileNotFoundError:
continue
try:
unified_structure = yaml.load(
open_file(
'../1-a-reinsert_notes/output/unified_structure/{}_unified_structure.yaml'
.format(work_name)))
except FileNotFoundError:
continue
import os
import re
from PyTib.common import open_file, write_file
missing_space = r'^([^ -])'
missing_space_rpl = r' \1'
missing_tsek = r'(ཅོ|སྣར|སྡེ|པེ)(?=):'
missing_tsek_rpl = r'\1་:'
files = [a for a in os.listdir('.') if a != 'conc_sanity_check.py']
for f in files:
print(f)
raw = open_file('./' + f)
raw = re.sub(missing_space, missing_space_rpl, raw)
raw = re.sub(missing_tsek, missing_tsek_rpl, raw)
raw = raw.replace('-1-,,,,,,,,,,,,,,,', '-1-,,,,,,,,,,,,,,,')
write_file('./' + f, raw)
lines = raw.split('\n')
for num, line in enumerate(lines):
toprint = False
if line.startswith('-'):
pass
elif line.startswith(r' ཅོ་:'):
pass
elif line.startswith(r' སྣར་:'):
pass
elif line.startswith(' སྡེ་:'):
pass
import sys, os
from PyTib.common import open_file
import os
bad = []
for f in os.listdir('./input'):
if f.endswith('Csv'):
content = open_file('./input/' + f)
if content.startswith(
'མཚན་བྱང་།,པར་གྲངས།,ཤོག་གྲངས།,1,མཚན་གྲངས།,པར་མ།,མཆན།'
) or content.startswith(
'མཚན་བྱང་།,པར་གྲངས།,ཤོག་གྲངས།,1,མཆན་གྲངས།,པར་མ།,མཆན།'):
bad.append(f)
for b in sorted(bad):
print(b)
print(f'No. of Bad files: {len(bad)}')
def reinsert(in_path, out_path1, out_path2, patterns):
print('reinsertion with notes')
for f in os.listdir(in_path):
work_name = f.replace('_a_reinserted.txt', '')
if work_name in patterns:
print('\t', work_name)
content = open_file('{}/{}'.format(in_path, f))
if not re.findall(r'\n\[\^[0-9A-Z]+\]\:', content):
text = content
notes = ''
else:
text, notes = [
a for a in re.split(
r'((?:\n?\[\^[0-9A-Z]+\]\:[^\n]+\n?)+)', content)
if a != ''
]
lines = deque(text.replace('\n', ' ').split('a'))
pages = []
text_pattern = patterns[work_name][2:]
counter = patterns[work_name][0][1]
side = patterns[work_name][0][2]
# beginning pages
for num in text_pattern[0]:
pages.append(create_page(lines, num, counter, side))
counter, side = increment_counter(counter, side)
# body of the text
while len(lines) > 0:
if len(lines) >= text_pattern[1]:
pages.append(
create_page(lines, text_pattern[1], counter, side))
counter, side = increment_counter(counter, side)
elif text_pattern[2] == len(lines):
pages.append(create_page(lines, len(lines), counter, side))
counter, side = increment_counter(counter, side)
else:
print(
'There is a line number issue: only {} lines were left for the last page.'
.format(len(lines)))
pages.append(create_page(lines, len(lines), counter, side))
counter, side = increment_counter(counter, side)
output = '\n{}\n'.format('-' * 100).join(pages) + '\n\n' + notes
write_file(
'{}/{}_page_reinserted.txt'.format(out_path1, work_name),
output)
# write to the file to 3-2-compared if it is not yet there
existing = [
g.replace('_compared.txt', '') for g in os.listdir(out_path2)
if g.endswith('.txt')
]
#if work_name not in existing:
write_file('{}/{}_compared.txt'.format(out_path2, work_name),
output)
text_path = '{}/extra_copies/{}'.format(out_path2, work_name)
if not os.path.exists(text_path):
os.makedirs(text_path)
in_path = './output/2-1-a_reinserted'
out_path1 = './output/3-1-page_reinserted'
out_path2 = './output/3-2-compared'
raw_in_path = './output/2-0-with_a'
raw_out_path = './output/2-2-raw_page_reinserted'
# '': [
# ('start', 0, ''), # page start + front/back
# ('end', 0, ''), # idem
# [0], # list of lines per page for the beginning of the text
# 0, # general number of lines per page
# 0 # number of lines pertaining to the current text on the last page
# ]
patterns_raw = open_file(
'C:/Users/trinley/github/canon_notes/4-a-final_formatting/resources/དཀར་ཆག་ཀུན་གསལ་མེ་ལོང་།.csv'
).strip().split('\n')
patterns = {}
for line in patterns_raw[1:]:
parts = line.split('\t')
names = patterns_raw[0].split('\t')
# 'ཆོས་ཚན།'
title = re.sub(r'_conc.*', '', parts[1])
# 'དབུ།'
b_page = int(parts[2])
sides = {'a': 'ན', 'b': 'བ'}
# 'རྒྱབ་འདུན།' (of 'དབུ།')
b_side = sides[parts[3]]
# 'ཞབས།'
e_page = int(parts[6])
# 'རྒྱབ་འདུན།' (of 'ཞབས།')
import PyTib
from PyTib.common import non_tib_chars, open_file, write_file, tib_sort, bisect_left
import os
import re
in_path = 'input/'
out_path = 'output/'
new = []
for f in os.listdir(in_path):
content = open_file(in_path+f).replace('༌', '་').split('\n')
content = [a.strip() for a in content if a != '']
# find all non-tibetan characters
to_delete = []
for c in content:
for char in c:
if char not in to_delete and non_tib_chars(char):
to_delete.append(char)
# add punctuation to be deleted
to_delete.extend(['།', '༎', '༄', '༅', '༑'])
# replace them with spaces
text = []
for r in range(len(content)-1):
line = content[r]
for t in to_delete:
line = line.replace(t, ' ')
text.append(re.sub(r'\s+', r' ', line))
lexicon = []
for t in text:
lexicon.extend([u.strip('་')+'་' for u in t.split(' ') if u.strip('་') != ''])
def processing_corpus(corpus_path):
print('processing corpus')
# variables for the main xslx
corpus_frequencies = {}
corpus_total_frequency = defaultdict(int)
corpus_origin = defaultdict(list)
# variables for individual xslx files
persons_frequencies = {}
persons_total_frequency = {}
persons_origin = {}
for f in os.listdir(corpus_path):
# A. finding the names for the current file
file_name = f
# finding the name of the section
section = ''
for c in corpus_sections:
if c in f:
section = c
file_name = file_name.replace(c, '')
if section == '':
section = 'ཁུངས་མེད།'
# finding the name of the person
person = re.sub(r'[0-9]* *\([0-9]+\).txt', '', file_name).strip()
# B. initiating the entries in dicts if missing for the current file
# creating an entry in corpus_frequencies{} if it does not exist
if section not in corpus_frequencies.keys():
corpus_frequencies[section] = defaultdict(int)
# creating an entry in persons_frequencies{} if it does not exist
if person not in persons_frequencies.keys():
persons_frequencies[person] = {}
if section not in persons_frequencies[person].keys():
persons_frequencies[person][section] = defaultdict(int)
# creating an entry in persons_total_frequency if it does not exist
if person not in persons_total_frequency.keys():
persons_total_frequency[person] = defaultdict(int)
# creating an entry in persons_origin{} if it does not exist
if person not in persons_origin.keys():
persons_origin[person] = defaultdict(list)
# C. processing the current file
content = open_file(corpus_path+'/'+f).replace('༌', '་').split('\n')
content = [a.strip() for a in content if a != '']
# find all non-tibetan characters
to_delete = chars_to_delete(content)
# replace them with spaces
text = []
for r in range(len(content)):
line = content[r]
for t in to_delete:
line = line.replace(t, ' ')
text.append(re.sub(r'\s+', r' ', line).strip())
# D. filling in the corpus and personal variables
# split the line in words and add it to the persons_frequencies and to the newspaper dict
for t in text:
split_line = [u.rstrip('་')+'་' for u in t.split(' ') if u.rstrip('་') != '']
for word in split_line:
clean_word = word.lstrip('་')
# corpus stats
corpus_frequencies[section][clean_word] += 1
corpus_total_frequency[clean_word] += 1
if f not in corpus_origin[clean_word]:
corpus_origin[clean_word].append(f)
# persons’ stats
persons_frequencies[person][section][clean_word] += 1
persons_total_frequency[person][clean_word] += 1
if f not in persons_origin[person][clean_word]:
persons_origin[person][clean_word].append(f)
return corpus_frequencies, corpus_total_frequency, corpus_origin, persons_frequencies, persons_total_frequency, persons_origin
def process(in_path, template_path, total_stats):
global collection_eds, file, debug
raw_template = open_file(template_path)
verbs = jp.decode(open_file('./resources/monlam_verbs.json'))
all_ngrams = open_ngrams()
files = find_file_path(in_path, '../1-a-reinsert_notes/output/conc_yaml')
# print(files)
for filename in files:
# if 'N5000' not in filename:
# continue
f = filename.split('/')[-1]
print(f)
if debug and f != file:
continue
work_name = f.replace('_conc.txt', '').replace('.txt', '')
raw = open_file(filename)
# setting collection_eds for the current file
collection_eds = list({a for a in re.findall(r' ([^ ]+): ', raw)})
if len(collection_eds) > 4:
print(collection_eds)
data = prepare_data(raw)
profiles, profile_cats = find_profiles(data)
# prepare
prepared = find_all_parts(data)
# categorise
categorised_notes = jp.decode(raw_template)
# find ngram frequencies
frequencies = ngram_frequency(prepared, all_ngrams)
if debug:
if file == f and note_num != 0:
for note in prepared:
if note[0] == note_num:
categorise(note, categorised_notes, verbs)
elif file == f:
for note in prepared:
categorise(note, categorised_notes, verbs)
else:
for note in prepared:
categorise(note, categorised_notes, verbs)
# finally write the json file
stats = {}
total = 0
for key1, item1 in sorted(categorised_notes.items()):
if type(item1) == list:
if len(item1) != 0:
stats[key1] = len(item1)
total += len(item1)
else:
stats[key1] = {}
for key2, item2 in sorted(item1.items()):
if type(item2) == list:
if len(item2) != 0:
stats[key1][key2] = len(item2)
total += len(item2)
else:
stats[key1][key2] = {}
for key3, item3 in sorted(item2.items()):
if type(item3) == list:
if len(item3) != 0:
stats[key1][key2][key3] = len(item3)
total += len(item3)
else:
stats[key1][key2][key3] = {}
for key4, item4 in sorted(item3.items()):
if type(item4) == list:
if len(item4) != 0:
stats[key1][key2][key3][
key4] = len(item4)
total += len(item4)
stats['Notes’ total'] = total
categorised = total
if 'long_diff' in stats['dunno'].keys():
categorised -= stats['dunno']['long_diff']
if 'short_diff' in stats['dunno'].keys():
categorised -= stats['dunno']['short_diff']
if 'no_diff' in stats['dunno'].keys():
categorised -= stats['dunno']['no_diff']
if total == 0:
percentage = 0
print('the notes were not processed!')
else:
percentage = categorised * 100 / total
stats['Categorised'] = '{} notes ({:02.2f}%)'.format(
categorised, percentage)
stats['Profiles'] = profile_cats
total_stats.append('{}\n{}'.format(work_name, jp.encode(stats)))
encoded = jp.encode(categorised_notes)
if encoded != raw_template:
categorised_notes['Stats'] = stats
categorised_notes['profile'] = profiles
categorised_notes['ngram_freq'] = frequencies
write_file('output/{}_cats.json'.format(work_name),
jp.encode(categorised_notes))
def create_base_text(raw_path):
for f in os.listdir('output'):
content = open_file('{}/{}_raw.txt'.format(raw_path, f))
# put back in a single line
content = content.replace('\n', ' ')
write_file('output/{}/{}_base.txt'.format(f, f), content)
def copy_final_version(final_path):
for f in os.listdir('output'):
if f+'_final.txt' in os.listdir('../4-a-final_formatting/output/3-3-final'):
write_file('output/{}/{}'.format(f, f+'_final.txt'), open_file('{}/{}_final.txt'.format(final_path, f)))
def copy_cat_json_file(json_path):
for f in os.listdir('output'):
write_file('output/{}/{}'.format(f, f+'_cats.json'), open_file('{}/{}_cats.json'.format(json_path, f)))
