conceptnet_all / src /cn_clean.py
DGurgurov's picture
Upload 6 files
1d26da5 verified
raw
history blame contribute delete
No virus
1.11 kB
import json
import re
def clean_prefixes(data):
cleaned_data = {}
prefix_pattern = re.compile(r'^/c/[^/]+/') # Matches the prefix part '/c/<lang_code>/'
for lang, words in data.items():
cleaned_words = {}
for key, values in words.items():
# Clean the key
cleaned_key = prefix_pattern.sub('', key).replace('_', ' ')
# Clean the values
cleaned_values = [prefix_pattern.sub('', value).replace('_', ' ') for value in values]
cleaned_words[cleaned_key] = cleaned_values
cleaned_data[lang] = cleaned_words
return cleaned_data
# Load the relations dictionary from the JSON file
input_file_path = "cn_relations.json"
with open(input_file_path, "r") as json_file:
data = json.load(json_file)
cleaned_data = clean_prefixes(data)
print(cleaned_data['zdj'])
# Save the cleaned data to a new JSON file
output_file_path = "cn_relations_clean.json"
with open(output_file_path, "w") as json_file:
json.dump(cleaned_data, json_file, ensure_ascii=False, indent=4)