|
|
|
|
|
import argparse |
|
import json |
|
from pathlib import Path |
|
|
|
from tqdm import tqdm |
|
|
|
from project_settings import project_path |
|
|
|
|
|
def get_args(): |
|
parser = argparse.ArgumentParser() |
|
|
|
parser.add_argument( |
|
"--data_dir", |
|
default=(project_path / "original_data/DoubanConversaionCorpus").as_posix(), |
|
type=str |
|
) |
|
parser.add_argument( |
|
"--output_file", |
|
default=(project_path / "data/douban.jsonl"), |
|
type=str |
|
) |
|
|
|
args = parser.parse_args() |
|
return args |
|
|
|
|
|
def main(): |
|
args = get_args() |
|
|
|
data_dir = Path(args.data_dir) |
|
with open(args.output_file, "w", encoding="utf-8") as fout: |
|
for filename in tqdm(data_dir.glob("*.txt")): |
|
with open(filename, "r", encoding="utf-8") as fin: |
|
for row in fin: |
|
splits = row.split("\t") |
|
label = splits[0] |
|
conversation = splits[1:] |
|
if label != "1": |
|
continue |
|
|
|
row = { |
|
"conversation": [ |
|
{ |
|
"role": "human" if i % 2 == 0 else "assistant", |
|
"message": "".join(str(c).strip().split()) |
|
} |
|
for i, c in enumerate(conversation) |
|
], |
|
"category": filename.stem, |
|
"data_source": "chatterbot", |
|
} |
|
row = json.dumps(row, ensure_ascii=False) |
|
fout.write("{}\n".format(row)) |
|
|
|
return |
|
|
|
|
|
if __name__ == '__main__': |
|
main() |
|
|
