feat: update optimize;

data/batch/batch.py

@@ -99,7 +99,7 @@ def process_single_file(
         )
         )
         model = "Qwen2-72B-Instruct"
         model = "Qwen2-72B-Instruct"
         # 合同审核Excel工作流处理
         # 合同审核Excel工作流处理
-        logger.info(" 第{}个文件,处理文件: {}".format(counter, original_file))
+        # logger.info(" 第{}个文件,处理文件: {}".format(counter, original_file))
 
 
         result = fastgpt_openai_chat(
         result = fastgpt_openai_chat(
             url,
             url,
@@ -124,9 +124,9 @@ def process_single_file(
                 ),
                 ),
                 des_doc_file,
                 des_doc_file,
             )
             )
-            logger.info(
-                f"第{counter}个文件下载:{excel_url}到{des_excel_file} {des_doc_file}"
-            )
+            # logger.info(
+            #     f"第{counter}个文件下载:{excel_url}到{des_excel_file} {des_doc_file}"
+            # )
     except Exception as e:
     except Exception as e:
         logger.error(f"{original_file} 处理异常 第{counter}个文件: {e}")
         logger.error(f"{original_file} 处理异常 第{counter}个文件: {e}")
         logger.error(traceback.print_exc())
         logger.error(traceback.print_exc())

data/optimize_review_prompt_flow.py

@@ -514,6 +514,7 @@ def parse_args() -> argparse.Namespace:
     parser.add_argument(
     parser.add_argument(
         "--batch-output",
         "--batch-output",
         type=Path,
         type=Path,
+        # default='/data/home/htsc/jp-contract/data/benchmark/results/jp-output-20260618-163339-976082',
         help="Use an existing batch output dir for the first round instead of running batch.py.",
         help="Use an existing batch output dir for the first round instead of running batch.py.",
     )
     )
     parser.add_argument(
     parser.add_argument(
@@ -537,7 +538,7 @@ def main() -> int:
             if not args.skip_app_restart:
             if not args.skip_app_restart:
                 restart_app(args.app_session, args.app_command)
                 restart_app(args.app_session, args.app_command)
             batch_output_dir = run_batch()
             batch_output_dir = run_batch()
-
+        print(f'round={round_idx} 完成批处理任务={batch_output_dir}', flush=True)
         # 2. 测评结果分析，生成优化建议，写回规则文件，记录优化报告
         # 2. 测评结果分析，生成优化建议，写回规则文件，记录优化报告
         eval_excel = run_eval(batch_output_dir)
         eval_excel = run_eval(batch_output_dir)
         eval_items = read_eval_items(
         eval_items = read_eval_items(
@@ -546,15 +547,16 @@ def main() -> int:
             f1_threshold=args.f1_threshold,
             f1_threshold=args.f1_threshold,
             example_limit=args.example_limit,
             example_limit=args.example_limit,
         )
         )
+        print(f'round={round_idx} 解析评测结果，待优化审查项数量={len(eval_items)}', flush=True)
         rule_tasks = build_rule_tasks(eval_items)
         rule_tasks = build_rule_tasks(eval_items)
         optimizations = asyncio.run(optimize_rule_tasks(rule_tasks, max_workers=args.max_workers))
         optimizations = asyncio.run(optimize_rule_tasks(rule_tasks, max_workers=args.max_workers))
-
+        print(f'round={round_idx} 完成LLM优化建议生成，优化建议数量={len(optimizations)}', flush=True)
         backup_file: Path | None = None
         backup_file: Path | None = None
         applied: list[dict[str, Any]] = []
         applied: list[dict[str, Any]] = []
         if optimizations and not args.dry_run:
         if optimizations and not args.dry_run:
             backup_file = backup_rules(rules_file, report_file.parent / "rule_backups", round_idx)
             backup_file = backup_rules(rules_file, report_file.parent / "rule_backups", round_idx)
             applied = update_rules_file(rules_file, optimizations)
             applied = update_rules_file(rules_file, optimizations)
-
+        # 
         append_round_report(
         append_round_report(
             report_file=report_file,
             report_file=report_file,
             round_idx=round_idx,
             round_idx=round_idx,
@@ -566,13 +568,7 @@ def main() -> int:
             applied=applied,
             applied=applied,
             f1_threshold=args.f1_threshold,
             f1_threshold=args.f1_threshold,
         )
         )
-
-        print(
-            f"round={round_idx} batch_output={batch_output_dir} eval_excel={eval_excel} "
-            f"optimized_items={len(eval_items)} optimized_rules={len(rule_tasks)} "
-            f"applied_updates={len(applied)} report={report_file}",
-            flush=True,
-        )
+        print('*' * 80, flush=True)
 
 
         if not eval_items:
         if not eval_items:
             break
             break

utils/http_util.py

@@ -116,7 +116,7 @@ def download_file(url, path, input_url_to_inner=True):
     if not url.startswith("http:") and not url.startswith("https:"):
     if not url.startswith("http:") and not url.startswith("https:"):
         url = base_fastgpt_url + url
         url = base_fastgpt_url + url
     url = url.replace(outer_backend_url, base_backend_url)
     url = url.replace(outer_backend_url, base_backend_url)
-    logger.info(f"url准备下载:{url}")
+    # logger.info(f"url准备下载:{url}")
     # 发送一个HTTP请求到URL
     # 发送一个HTTP请求到URL
     response = requests.get(url)
     response = requests.get(url)
     # 确保请求成功
     # 确保请求成功
@@ -131,10 +131,10 @@ def download_file(url, path, input_url_to_inner=True):
         with open(target_path, "wb") as f:
         with open(target_path, "wb") as f:
             # 写入响应的内容
             # 写入响应的内容
             f.write(response.content)
             f.write(response.content)
-        logger.info(f"{url}文件下载成功，保存到{target_path}")
+        # logger.info(f"{url}文件下载成功，保存到{target_path}")
         return str(target_path)
         return str(target_path)
     else:
     else:
-        logger.error(f"{url}文件下载失败. HTTP Status Code: {response.status_code}")
+        # logger.error(f"{url}文件下载失败. HTTP Status Code: {response.status_code}")
         return None
         return None