fix(embedding): 彻底解决事件循环冲突导致的嵌入生成异常
通过以下改动修复嵌入生成过程中的事件循环相关问题： - 在 EmbeddingStore._get_embedding 中，改为同步创建-使用-销毁的新事件循环模式，彻底避免嵌套事件循环问题 - 调整批量嵌入 _get_embeddings_batch_threaded，确保每个线程使用独立、短生命周期的事件循环 - 新增 force_new 参数，LLM 请求嵌入任务时强制创建新的客户端实例，减少跨循环对象复用 - 在 OpenAI 客户端的 embedding 调用处补充详细日志，方便排查网络连接异常 - get_embedding() 每次都重建 LLMRequest，降低实例在多个事件循环中穿梭的概率 此次改动虽然以同步风格“硬掰”异步接口，但对现有接口零破坏，确保了向量数据库及相关知识检索功能的稳定性。(还有就是把的脚本文件夹移回来了)
This commit is contained in:
minecraft1024a
217
scripts/info_extraction.py
Normal file
217
scripts/info_extraction.py
Normal file
@@ -0,0 +1,217 @@
|
||||
import json
|
||||
import os
|
||||
import signal
|
||||
from concurrent.futures import ThreadPoolExecutor, as_completed
|
||||
from threading import Lock, Event
|
||||
import sys
|
||||
import datetime
|
||||
|
||||
sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), "..")))
|
||||
# 添加项目根目录到 sys.path
|
||||
|
||||
from rich.progress import Progress # 替换为 rich 进度条
|
||||
|
||||
from src.common.logger import get_logger
|
||||
# from src.chat.knowledge.lpmmconfig import global_config
|
||||
from src.chat.knowledge.ie_process import info_extract_from_str
|
||||
from src.chat.knowledge.open_ie import OpenIE
|
||||
from rich.progress import (
|
||||
BarColumn,
|
||||
TimeElapsedColumn,
|
||||
TimeRemainingColumn,
|
||||
TaskProgressColumn,
|
||||
MofNCompleteColumn,
|
||||
SpinnerColumn,
|
||||
TextColumn,
|
||||
)
|
||||
from raw_data_preprocessor import RAW_DATA_PATH, load_raw_data
|
||||
from src.config.config import global_config, model_config
|
||||
from src.llm_models.utils_model import LLMRequest
|
||||
|
||||
logger = get_logger("LPMM知识库-信息提取")
|
||||
|
||||
|
||||
ROOT_PATH = os.path.abspath(os.path.join(os.path.dirname(__file__), ".."))
|
||||
TEMP_DIR = os.path.join(ROOT_PATH, "temp")
|
||||
# IMPORTED_DATA_PATH = os.path.join(ROOT_PATH, "data", "imported_lpmm_data")
|
||||
OPENIE_OUTPUT_DIR = os.path.join(ROOT_PATH, "data", "openie")
|
||||
|
||||
def ensure_dirs():
|
||||
"""确保临时目录和输出目录存在"""
|
||||
if not os.path.exists(TEMP_DIR):
|
||||
os.makedirs(TEMP_DIR)
|
||||
logger.info(f"已创建临时目录: {TEMP_DIR}")
|
||||
if not os.path.exists(OPENIE_OUTPUT_DIR):
|
||||
os.makedirs(OPENIE_OUTPUT_DIR)
|
||||
logger.info(f"已创建输出目录: {OPENIE_OUTPUT_DIR}")
|
||||
if not os.path.exists(RAW_DATA_PATH):
|
||||
os.makedirs(RAW_DATA_PATH)
|
||||
logger.info(f"已创建原始数据目录: {RAW_DATA_PATH}")
|
||||
|
||||
# 创建一个线程安全的锁，用于保护文件操作和共享数据
|
||||
file_lock = Lock()
|
||||
open_ie_doc_lock = Lock()
|
||||
|
||||
# 创建一个事件标志，用于控制程序终止
|
||||
shutdown_event = Event()
|
||||
|
||||
lpmm_entity_extract_llm = LLMRequest(
|
||||
model_set=model_config.model_task_config.lpmm_entity_extract,
|
||||
request_type="lpmm.entity_extract"
|
||||
)
|
||||
lpmm_rdf_build_llm = LLMRequest(
|
||||
model_set=model_config.model_task_config.lpmm_rdf_build,
|
||||
request_type="lpmm.rdf_build"
|
||||
)
|
||||
def process_single_text(pg_hash, raw_data):
|
||||
"""处理单个文本的函数，用于线程池"""
|
||||
temp_file_path = f"{TEMP_DIR}/{pg_hash}.json"
|
||||
|
||||
# 使用文件锁检查和读取缓存文件
|
||||
with file_lock:
|
||||
if os.path.exists(temp_file_path):
|
||||
try:
|
||||
# 存在对应的提取结果
|
||||
logger.info(f"找到缓存的提取结果：{pg_hash}")
|
||||
with open(temp_file_path, "r", encoding="utf-8") as f:
|
||||
return json.load(f), None
|
||||
except json.JSONDecodeError:
|
||||
# 如果JSON文件损坏，删除它并重新处理
|
||||
logger.warning(f"缓存文件损坏，重新处理：{pg_hash}")
|
||||
os.remove(temp_file_path)
|
||||
|
||||
entity_list, rdf_triple_list = info_extract_from_str(
|
||||
lpmm_entity_extract_llm,
|
||||
lpmm_rdf_build_llm,
|
||||
raw_data,
|
||||
)
|
||||
if entity_list is None or rdf_triple_list is None:
|
||||
return None, pg_hash
|
||||
doc_item = {
|
||||
"idx": pg_hash,
|
||||
"passage": raw_data,
|
||||
"extracted_entities": entity_list,
|
||||
"extracted_triples": rdf_triple_list,
|
||||
}
|
||||
# 保存临时提取结果
|
||||
with file_lock:
|
||||
try:
|
||||
with open(temp_file_path, "w", encoding="utf-8") as f:
|
||||
json.dump(doc_item, f, ensure_ascii=False, indent=4)
|
||||
except Exception as e:
|
||||
logger.error(f"保存缓存文件失败：{pg_hash}, 错误：{e}")
|
||||
# 如果保存失败，确保不会留下损坏的文件
|
||||
if os.path.exists(temp_file_path):
|
||||
os.remove(temp_file_path)
|
||||
sys.exit(0)
|
||||
return None, pg_hash
|
||||
return doc_item, None
|
||||
|
||||
|
||||
def signal_handler(_signum, _frame):
|
||||
"""处理Ctrl+C信号"""
|
||||
logger.info("\n接收到中断信号，正在优雅地关闭程序...")
|
||||
sys.exit(0)
|
||||
|
||||
|
||||
def main(): # sourcery skip: comprehension-to-generator, extract-method
|
||||
# 设置信号处理器
|
||||
signal.signal(signal.SIGINT, signal_handler)
|
||||
ensure_dirs() # 确保目录存在
|
||||
# 新增用户确认提示
|
||||
print("=== 重要操作确认，请认真阅读以下内容哦 ===")
|
||||
print("实体提取操作将会花费较多api余额和时间，建议在空闲时段执行。")
|
||||
print("举例：600万字全剧情，提取选用deepseek v3 0324，消耗约40元，约3小时。")
|
||||
print("建议使用硅基流动的非Pro模型")
|
||||
print("或者使用可以用赠金抵扣的Pro模型")
|
||||
print("请确保账户余额充足，并且在执行前确认无误。")
|
||||
confirm = input("确认继续执行？(y/n): ").strip().lower()
|
||||
if confirm != "y":
|
||||
logger.info("用户取消操作")
|
||||
print("操作已取消")
|
||||
sys.exit(1)
|
||||
print("\n" + "=" * 40 + "\n")
|
||||
ensure_dirs() # 确保目录存在
|
||||
logger.info("--------进行信息提取--------\n")
|
||||
|
||||
# 加载原始数据
|
||||
logger.info("正在加载原始数据")
|
||||
all_sha256_list, all_raw_datas = load_raw_data()
|
||||
|
||||
failed_sha256 = []
|
||||
open_ie_doc = []
|
||||
|
||||
workers = global_config.lpmm_knowledge.info_extraction_workers
|
||||
with ThreadPoolExecutor(max_workers=workers) as executor:
|
||||
future_to_hash = {
|
||||
executor.submit(process_single_text, pg_hash, raw_data): pg_hash
|
||||
for pg_hash, raw_data in zip(all_sha256_list, all_raw_datas, strict=False)
|
||||
}
|
||||
|
||||
with Progress(
|
||||
SpinnerColumn(),
|
||||
TextColumn("[progress.description]{task.description}"),
|
||||
BarColumn(),
|
||||
TaskProgressColumn(),
|
||||
MofNCompleteColumn(),
|
||||
"•",
|
||||
TimeElapsedColumn(),
|
||||
"<",
|
||||
TimeRemainingColumn(),
|
||||
transient=False,
|
||||
) as progress:
|
||||
task = progress.add_task("正在进行提取：", total=len(future_to_hash))
|
||||
try:
|
||||
for future in as_completed(future_to_hash):
|
||||
if shutdown_event.is_set():
|
||||
for f in future_to_hash:
|
||||
if not f.done():
|
||||
f.cancel()
|
||||
break
|
||||
|
||||
doc_item, failed_hash = future.result()
|
||||
if failed_hash:
|
||||
failed_sha256.append(failed_hash)
|
||||
logger.error(f"提取失败：{failed_hash}")
|
||||
elif doc_item:
|
||||
with open_ie_doc_lock:
|
||||
open_ie_doc.append(doc_item)
|
||||
progress.update(task, advance=1)
|
||||
except KeyboardInterrupt:
|
||||
logger.info("\n接收到中断信号，正在优雅地关闭程序...")
|
||||
shutdown_event.set()
|
||||
for f in future_to_hash:
|
||||
if not f.done():
|
||||
f.cancel()
|
||||
|
||||
# 合并所有文件的提取结果并保存
|
||||
if open_ie_doc:
|
||||
sum_phrase_chars = sum([len(e) for chunk in open_ie_doc for e in chunk["extracted_entities"]])
|
||||
sum_phrase_words = sum([len(e.split()) for chunk in open_ie_doc for e in chunk["extracted_entities"]])
|
||||
num_phrases = sum([len(chunk["extracted_entities"]) for chunk in open_ie_doc])
|
||||
openie_obj = OpenIE(
|
||||
open_ie_doc,
|
||||
round(sum_phrase_chars / num_phrases, 4) if num_phrases else 0,
|
||||
round(sum_phrase_words / num_phrases, 4) if num_phrases else 0,
|
||||
)
|
||||
# 输出文件名格式：MM-DD-HH-ss-openie.json
|
||||
now = datetime.datetime.now()
|
||||
filename = now.strftime("%m-%d-%H-%S-openie.json")
|
||||
output_path = os.path.join(OPENIE_OUTPUT_DIR, filename)
|
||||
with open(output_path, "w", encoding="utf-8") as f:
|
||||
json.dump(
|
||||
openie_obj.to_dict() if hasattr(openie_obj, "to_dict") else openie_obj.__dict__,
|
||||
f,
|
||||
ensure_ascii=False,
|
||||
indent=4,
|
||||
)
|
||||
logger.info(f"信息提取结果已保存到: {output_path}")
|
||||
else:
|
||||
logger.warning("没有可保存的信息提取结果")
|
||||
|
||||
logger.info("--------信息提取完成--------")
|
||||
logger.info(f"提取失败的文段SHA256：{failed_sha256}")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
Reference in New Issue
Block a user