def provide_corpus_functionwords(numOfFunctionwords):
global top_corpus_functionwords
top_corpus_functionwords={}
synchronized_functionwords_file_path = Path("vectors_handling/vectors/synchronized_functionwords/synchronized_functionwords.txt")
if not util.exists(synchronized_functionwords_file_path): # can't find the file in memory
log('Cannot find synchronized_functionwords file') # redundant
corpus_functionwords = {}
for domain_dir in os.scandir(setup.database):
if domain_dir.name == 'europe_data' and setup.domain == 'in':
for country_dir in os.scandir(domain_dir):
country_name = str.split(os.path.basename(country_dir), '.')[1]
log('Counting function words in ' + country_name)
for user_dir in os.scandir(country_dir):
for file_dir in os.scandir(user_dir):
file = open(file_dir, "r", encoding="utf-8")
lines = file.readlines()
for line in lines:
words = line.split()
for word in words:
if word in function_words_map.keys():
if word not in corpus_functionwords.keys():
corpus_functionwords[word] = 1
else:
corpus_functionwords[word] += 1
top_corpus_functionwords = heapq.nlargest(numOfFunctionwords, corpus_functionwords,
key=corpus_functionwords.get)
util.save_file(synchronized_functionwords_file_path, top_corpus_functionwords)
top_corpus_functionwords = util.load_file(synchronized_functionwords_file_path)
return top_corpus_functionwords
def provide_top_spelling_errors():
spelling_file_path = Path("vectors_handling/vectors/spelling_errors/top_spelling_errors.txt")
if not util.exists(spelling_file_path): # can't find the file in memory
log('Cannot find top bipos file') # redundant
generate_top_spelling_errors(spelling_file_path)
top_spelling_errors = util.load_file(spelling_file_path)
return top_spelling_errors
def provide_top_unigram():
unigram_file_path = Path("vectors_handling/vectors/unigrams/top_unigrams.txt")
if not util.exists(unigram_file_path): # can't find the file in memory
log('Cannot find top unigrams file') # redundant
generate_top_unigrams(unigram_file_path)
top_unigrams = util.load_file(unigram_file_path)
return top_unigrams
def provide_top_tripos():
tripos_file_path = Path("vectors_handling/vectors/pos/top_tripos.txt")
if not util.exists(tripos_file_path): # can't find the file in memory
log('Cannot find top tripos file') # redundant
generate_top_tripos(tripos_file_path)
top_tripos = util.load_file(tripos_file_path)
return top_tripos