|
import os |
|
import json |
|
from collections import defaultdict |
|
from transformers import LlamaTokenizer |
|
from tqdm import tqdm |
|
import argparse |
|
|
|
def build_entity_triples(triples): |
|
entity_triples = defaultdict(list) |
|
for triple in triples: |
|
subj, relation, obj = triple |
|
|
|
triple_str = f"{subj} - {relation} - {obj}" |
|
entity_triples[subj].append(triple_str) |
|
return entity_triples |
|
|
|
def create_paragraphs_with_alignment(subj, old_triples_list, new_triples_list, tokenizer, max_tokens=512): |
|
paragraphs_old = [] |
|
paragraphs_new = [] |
|
boundaries = [] |
|
|
|
index = 0 |
|
start_idx = 0 |
|
while index < len(old_triples_list): |
|
|
|
candidate_old_triples = old_triples_list[start_idx:index+1] |
|
paragraph_text_old = ', '.join(candidate_old_triples) |
|
tokens_in_paragraph = len(tokenizer.encode(paragraph_text_old, add_special_tokens=False)) |
|
|
|
if tokens_in_paragraph > max_tokens: |
|
if index == start_idx: |
|
|
|
boundaries.append(index+1) |
|
start_idx = index+1 |
|
index += 1 |
|
else: |
|
boundaries.append(index) |
|
start_idx = index |
|
else: |
|
index += 1 |
|
|
|
|
|
if start_idx < len(old_triples_list): |
|
boundaries.append(len(old_triples_list)) |
|
|
|
|
|
start = 0 |
|
for end in boundaries: |
|
|
|
old_paragraph_triples = old_triples_list[start:end] |
|
paragraph_text_old = ', '.join(old_paragraph_triples) |
|
paragraphs_old.append({ |
|
"title": subj, |
|
"contents": paragraph_text_old |
|
}) |
|
|
|
|
|
new_paragraph_triples = new_triples_list[start:end] |
|
paragraph_text_new = ', '.join(new_paragraph_triples) |
|
paragraphs_new.append({ |
|
"title": subj, |
|
"contents": paragraph_text_new |
|
}) |
|
|
|
start = end |
|
|
|
return paragraphs_old, paragraphs_new |
|
|
|
def main(args): |
|
os.makedirs(os.path.dirname(args.old_output_file), exist_ok=True) |
|
os.makedirs(os.path.dirname(args.new_output_file), exist_ok=True) |
|
|
|
tokenizer = LlamaTokenizer.from_pretrained(args.model_path) |
|
|
|
with open(args.old_input_json_file, 'r') as f: |
|
old_triples = json.load(f) |
|
|
|
with open(args.new_input_json_file, 'r') as f: |
|
new_triples = json.load(f) |
|
|
|
old_entity_triples = build_entity_triples(old_triples) |
|
new_entity_triples = build_entity_triples(new_triples) |
|
|
|
|
|
all_paragraphs_old = [] |
|
all_paragraphs_new = [] |
|
entities = set(old_entity_triples.keys()).union(new_entity_triples.keys()) |
|
|
|
for subj in tqdm(entities, desc="Processing entities", unit="entity"): |
|
old_triples_list = old_entity_triples.get(subj, []) |
|
new_triples_list = new_entity_triples.get(subj, []) |
|
|
|
|
|
max_length = max(len(old_triples_list), len(new_triples_list)) |
|
if len(old_triples_list) < max_length: |
|
old_triples_list.extend([''] * (max_length - len(old_triples_list))) |
|
if len(new_triples_list) < max_length: |
|
new_triples_list.extend([''] * (max_length - len(new_triples_list))) |
|
|
|
paragraphs_old, paragraphs_new = create_paragraphs_with_alignment( |
|
subj, old_triples_list, new_triples_list, tokenizer, args.max_tokens |
|
) |
|
all_paragraphs_old.extend(paragraphs_old) |
|
all_paragraphs_new.extend(paragraphs_new) |
|
|
|
with open(args.old_output_file, 'w') as out_f: |
|
json.dump(all_paragraphs_old, out_f, ensure_ascii=False, indent=4) |
|
|
|
with open(args.new_output_file, 'w') as out_f: |
|
json.dump(all_paragraphs_new, out_f, ensure_ascii=False, indent=4) |
|
|
|
print(f"old {args.old_output_file}") |
|
print(f"new {args.new_output_file}") |
|
|
|
if __name__ == "__main__": |
|
parser = argparse.ArgumentParser(description="Process old and new triples") |
|
parser.add_argument('--old_input_json_file', type=str, required=True, help='旧的三元组JSON文件的路径。') |
|
parser.add_argument('--new_input_json_file', type=str, required=True, help='新的三元组JSON文件的路径。') |
|
parser.add_argument('--old_output_file', type=str, required=True, help='保存处理后的旧三元组输出文件的完整路径。') |
|
parser.add_argument('--new_output_file', type=str, required=True, help='保存处理后的新三元组输出文件的完整路径。') |
|
parser.add_argument('--model_path', type=str, required=True, help='预训练的Tokenizer模型的路径。') |
|
parser.add_argument('--max_tokens', type=int, default=512, help='每个段落的最大令牌数。') |
|
|
|
args = parser.parse_args() |
|
main(args) |
|
|