167 lines
5.4 KiB
Python
167 lines
5.4 KiB
Python
import json
|
||
from openai import OpenAI
|
||
from kg_build_v2.Prompt import Prompt_anli
|
||
from kg_build_v2.extract_html import extract_fault_sections,extract_by_unit,extract_cases,extract_content_slices
|
||
import re
|
||
from modelsAPI.model_api import OpenaiAPI
|
||
|
||
def read_markdown_file(file_path):
|
||
"""
|
||
读取 Markdown 文件并返回其全部内容作为字符串。
|
||
|
||
参数:
|
||
file_path (str): Markdown 文件的路径（支持相对或绝对路径）
|
||
|
||
返回:
|
||
str 或 None: 成功时返回文件内容；出错时返回 None
|
||
"""
|
||
try:
|
||
with open(file_path, 'r', encoding='utf-8') as f:
|
||
content = f.read()
|
||
return content
|
||
except FileNotFoundError:
|
||
print(f"错误：文件未找到 - {file_path}")
|
||
except PermissionError:
|
||
print(f"错误：没有权限读取文件 - {file_path}")
|
||
except UnicodeDecodeError:
|
||
print(f"错误：文件编码不是 UTF-8，无法读取 - {file_path}")
|
||
except Exception as e:
|
||
print(f"读取文件时发生未知错误: {e}")
|
||
return None
|
||
|
||
|
||
def safe_json_loads(s: str):
|
||
try:
|
||
return json.loads(s)
|
||
except json.JSONDecodeError as e:
|
||
# 尝试修复常见问题
|
||
s = s.replace("\\n", "\\\n").replace("\r", "\\r").replace("\t", "\\t")
|
||
# 或者直接清理非法字符
|
||
import re
|
||
s = re.sub(r'[\x00-\x1f\x7f]', '', s) # 移除控制字符
|
||
try:
|
||
return json.loads(s)
|
||
except:
|
||
raise e
|
||
|
||
# def openai_chat(query:str, model="Qwen3-32B"):
|
||
|
||
# """调用OpenAI API"""
|
||
# try:
|
||
# client = OpenAI(
|
||
# api_key= "none",
|
||
# base_url= "http://192.168.0.46:59800/v1"
|
||
# )
|
||
|
||
# response = client.chat.completions.create(
|
||
# model=model,
|
||
# messages=[
|
||
# {"role": "user", "content": query},
|
||
# ],
|
||
# temperature=0.1,
|
||
# stream=False,
|
||
# response_format = {"type": "json_object"},
|
||
# extra_body={"chat_template_kwargs": {"enable_thinking": False}}
|
||
# )
|
||
# return response.choices[0].message.content
|
||
# except Exception as e:
|
||
# print(f"调用OpenAI API时出错: {e}")
|
||
# return None
|
||
|
||
def extract_wxanli(chunks):
|
||
"""
|
||
批量处理一个文档的多个切片（chunks），每个切片调用大模型结构化，并返回全部结果。
|
||
|
||
:param chunks: List[str] - 同一文档的多个文本切片
|
||
:return: List[Dict] - 结构化结果列表
|
||
"""
|
||
all_results = []
|
||
all_entities = []
|
||
all_relationships = []
|
||
|
||
for idx, block in enumerate(chunks, 1):
|
||
print(f"=== 维修案例 Block {idx} ===")
|
||
print(block)
|
||
print("\\n" + "="*60 + "\\n")
|
||
|
||
final_prompt = Prompt_anli.replace("text",block)
|
||
|
||
raw_response = OpenaiAPI.openai_chat(final_prompt,timeout=180)
|
||
|
||
# 尝试解析 JSON
|
||
try:
|
||
cleaned = raw_response.strip()
|
||
if cleaned.startswith("```json"):
|
||
cleaned = cleaned[7:].lstrip()
|
||
if cleaned.endswith("```"):
|
||
cleaned = cleaned[:-3].rstrip()
|
||
|
||
# ✅ 修复：移除 ensure_ascii 和 indent
|
||
json_result = safe_json_loads(cleaned)
|
||
entities = json_result.get("entities", [])
|
||
relationships = json_result.get("relationships", [])
|
||
for entity in entities:
|
||
# 确保 entity 有 "properties" 键，且其值为字典
|
||
if "properties" in entity and isinstance(entity["properties"], dict):
|
||
entity["properties"]["切片"] = block
|
||
# 打印美化版
|
||
all_entities.extend(entities)
|
||
all_relationships.extend(relationships)
|
||
# 将当前结果追加到总列表
|
||
# all_results.append({
|
||
# "entities": entities, # 已添加切片属性的实体列表
|
||
# "relationships": relationships, # 关系列表
|
||
# })
|
||
# 可选：打印美化版（使用 dumps）
|
||
|
||
success = True
|
||
except Exception as e:
|
||
print(f" ⚠️ 第 {idx} 切片 JSON 解析失败: {e}")
|
||
print(f" 原始响应: {repr(raw_response)}")
|
||
continue # 跳过保存失败项
|
||
|
||
|
||
result = {
|
||
"entities": all_entities,
|
||
"relationships": all_relationships
|
||
}
|
||
return result
|
||
|
||
|
||
|
||
|
||
def get_wxanli_entity_rela(markdown_content):
|
||
keywords = ["一、故障现象", "故障名称：", "单位："]
|
||
chunks = extract_content_slices(markdown_content,keywords)
|
||
|
||
valid_chunks = [chunk for chunk in chunks if chunk and len(chunk.strip()) > 10]
|
||
if valid_chunks:
|
||
# 调用处理函数
|
||
results = extract_wxanli(valid_chunks)
|
||
|
||
# 这里可以将 results 保存到文件，例如按文件名保存为 JSON
|
||
|
||
else:
|
||
results = []
|
||
|
||
|
||
return results
|
||
|
||
|
||
|
||
|
||
|
||
|
||
# if __name__ == "__main__":
|
||
# input_directory = r"F:\zklnlp\HJ\hj_kg_code\kgrag\table_extract\files-1\图解目录_shenghna_output11_full_content.md"
|
||
|
||
# # 调用目录处理函数
|
||
# mdcontent = read_markdown_file(input_directory)
|
||
|
||
# if mdcontent is None:
|
||
# print("无法读取文件，程序退出")
|
||
# exit(1)
|
||
|
||
# res = get_wxanli_entity_rela(markdown_content=mdcontent)
|
||
# print(res)
|