macula-grass / process_csvs.py
jcuenod
Initial commit
6075d60
import os
import csv
GRASS_DIR = "grass"
def process_csvs(output_file):
files = sorted([f for f in os.listdir(GRASS_DIR) if f.endswith(".csv")])
sentence_id_map = {}
clause_id_map = {}
phrase_id_map = {}
next_sentence_id = 1
next_clause_id = 1
next_phrase_id = 1
all_rows = []
header = None
for filename in files:
path = os.path.join(GRASS_DIR, filename)
print(f"Processing {path}...")
with open(path, newline='', encoding='utf-8') as csvfile:
reader = csv.DictReader(csvfile)
if header is None:
header = reader.fieldnames
for row in reader:
# Sentence ID
orig_sentence = row['sentence_id']
if orig_sentence not in sentence_id_map:
sentence_id_map[orig_sentence] = str(next_sentence_id)
next_sentence_id += 1
row['sentence_id'] = sentence_id_map[orig_sentence]
# Clause ID
orig_clause = row['clause_id']
if orig_clause not in clause_id_map:
clause_id_map[orig_clause] = str(next_clause_id)
next_clause_id += 1
row['clause_id'] = clause_id_map[orig_clause]
# Phrase ID
orig_phrase = row['phrase_id']
if orig_phrase not in phrase_id_map:
phrase_id_map[orig_phrase] = str(next_phrase_id)
next_phrase_id += 1
row['phrase_id'] = phrase_id_map[orig_phrase]
all_rows.append(row)
# Write output
if header is None:
# Fallback header if no files found
header = ["sentence_id", "clause_id", "phrase_id", "word_id", "ref", "text", "lemma", "strong", "morph"]
with open(output_file, 'w', newline='', encoding='utf-8') as outcsv:
writer = csv.DictWriter(outcsv, fieldnames=header)
writer.writeheader()
for row in all_rows:
writer.writerow(row)
if __name__ == "__main__":
process_csvs("macula_grass.csv")