#2025/7/22 #用vl模型进行预警

shidong

2025-07-22 2c3655dcfa20d6ac36219b879debef5590f936d4

 qwen_thread.py

@@ -1,18 +1,15 @@
import time
from concurrent.futures import ThreadPoolExecutor
import threading

import torch
from PIL import Image
from pymilvus import connections, Collection
from datetime import datetime
import os
import requests
import asyncio
import logging
import re
from logging.handlers import RotatingFileHandler

from qwen_vl_utils import process_vision_info
from transformers import AutoModelForVision2Seq, AutoProcessor


@@ -29,19 +26,22 @@
        # 加载集合
        self.collection = Collection(name="smartobject")
        self.collection.load()

        if config.get('cuda') == None or config.get('cuda') == '0':
            self.device = f"cuda"
        else:
            self.device = f"cuda:{config.get('cuda')}"
        self.model_pool = []
        self.lock_pool = [threading.Lock() for _ in range(int(config.get("threadnum")))]
        for i in range(int(config.get("threadnum"))):
            model = AutoModelForVision2Seq.from_pretrained(
                config.get("qwenaddr"),
                device_map=f"cuda:{config.get('cuda')}",
                device_map=self.device,
                trust_remote_code=True,
                use_safetensors=True,
                torch_dtype=torch.float16

            ).eval()
            model = model.to(f"cuda:{config.get('cuda')}")
            model = model.to(self.device)
            self.model_pool.append(model)

        # 共享的处理器 (线程安全)
@@ -73,19 +73,14 @@
            # 1. 从集合A获取向量和元数据
            is_waning = 0
            is_desc = 2

            # 生成图片描述
            ks_time = datetime.now()
            desc = self.image_desc(res)
            desc_time = datetime.now() - ks_time
            current_time = datetime.now()
            risk_description = ""
            suggestion = ""
            # 图片描述生成成功
            if desc:
                is_desc = 2
                # 调用规则匹配方法,判断是否预警
                is_waning = self.image_rule_chat(desc, res['waning_value'], ragurl,rag_mode,max_tokens)
            is_waning = self.image_rule(res)
                # 如果预警,则生成隐患描述和处理建议
                if is_waning == 1:
                    # 获取规章制度数据
@@ -94,11 +89,42 @@
                    risk_description = self.image_rule_chat_with_detail(filedata, res['waning_value'], ragurl,rag_mode,max_tokens)
                    # 生成处理建议
                    suggestion = self.image_rule_chat_suggestion(filedata, res['waning_value'], ragurl,rag_mode,max_tokens)
                    self.logger.info(
                        f"{res['video_point_id']}执行完毕：{res['id']}:是否预警{is_waning},安全隐患：{risk_description}\n处理建议：{suggestion}")
                #self.logger.info(f"{res['video_point_id']}执行完毕：{res['id']}:是否预警{is_waning},安全隐患：{risk_description}\n处理建议：{suggestion}")
                # 数据组
            data = {
                "event_level_id": res['event_level_id'],  # event_level_id
                "event_level_name": res['event_level_name'],  # event_level_id
                "rule_id": res["rule_id"],
                "video_point_id": res['video_point_id'],  # video_point_id
                "video_point_name": res['video_point_name'],
                "is_waning": is_waning,
                "is_desc": 1,
                "zh_desc_class": res['zh_desc_class'],  # text_vector
                "bounding_box": res['bounding_box'],  # bounding_box
                "task_id": res['task_id'],  # task_id
                "task_name": res['task_name'],  # task_id
                "detect_id": res['detect_id'],  # detect_id
                "detect_time": res['detect_time'],  # detect_time
                "detect_num": res['detect_num'],
                "waning_value": res['waning_value'],
                "image_path": res['image_path'],  # image_path
                "image_desc_path": res['image_desc_path'],  # image_desc_path
                "video_path": res['video_path'],
                "text_vector": res['text_vector'],
                "risk_description": risk_description,
                "suggestion": suggestion,
                "knowledge_id": res['knowledge_id']
            }
            self.collection.delete(f"id == {res['id']}")
            # 保存到milvus
            image_id = self.collection.insert(data).primary_keys
            res['id'] = image_id[0]
            # 图片描述生成成功
            desc = self.image_desc(res)
            if desc:
                is_desc = 2
            else:
                is_desc = 3

            # 数据组
            data = {
                "event_level_id": res['event_level_id'],  # event_level_id
@@ -141,7 +167,9 @@
            # 调用rag
            asyncio.run(self.insert_json_data(ragurl, data))
            rag_time = datetime.now() - current_time
            self.logger.info(f"{res['video_point_id']}执行完毕：{image_id}运行结束总体用时:{datetime.now() - ks_time},图片描述用时{desc_time}，RAG用时{rag_time}")
            self.logger.info(f"{res['video_point_id']}执行完毕：{image_id}运行结束总体用时:{datetime.now() - ks_time},图片描述用时{desc_time},RAG用时{rag_time}")
            if is_waning == 1:
                self.logger.info(f"{res['video_point_id']}执行完毕：{image_id},图片描述：{desc}\n隐患：{risk_description}\n建议：{suggestion}")
        except Exception as e:
            self.logger.info(f"线程：执行模型解析时出错::{e}")
            return 0
@@ -173,12 +201,58 @@
                return_tensors="pt",
            )
            inputs = inputs.to(model.device)
            with torch.inference_mode(),torch.cuda.amp.autocast():
                outputs = model.generate(**inputs,max_new_tokens=200)
            with torch.inference_mode(), torch.amp.autocast(device_type=self.device, dtype=torch.float16):
                outputs = model.generate(**inputs,max_new_tokens=200,do_sample=False,num_beams=1,temperature=None,top_p=None,top_k=1,use_cache=True,repetition_penalty=1.0)
            generated_ids = outputs[:, len(inputs.input_ids[0]):]
            image_text = self.processor.batch_decode(
                generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True
            )
            image_des = (image_text[0]).strip()
            #self.logger.info(f"{res_data['video_point_id']}:{res_data['id']}:{res_data['detect_time']}:{image_des}")
            return image_des
        except Exception as e:
            self.logger.info(f"线程：执行图片描述时出错:{e}")
        finally:
            # 4. 释放模型
            self._release_model(model)
            torch.cuda.empty_cache()

    def image_rule(self, res_data):
        try:
            model, lock = self._acquire_model()
            image = Image.open(res_data['image_desc_path']).convert("RGB").resize((600, 600), Image.Resampling.LANCZOS)

            messages = [
                {
                    "role": "user",
                    "content": [
                        {"type": "image", "image": image},
                        {"type": "text", "text": f"图片中是否有{res_data['waning_value']}?请回答yes或no"},
                    ],
                }
            ]

            # Preparation for inference
            text = self.processor.apply_chat_template(
                messages, tokenize=False, add_generation_prompt=True
            )
            image_inputs, video_inputs = process_vision_info(messages)
            inputs = self.processor(
                text=[text],
                images=image_inputs,
                videos=video_inputs,
                padding=True,
                return_tensors="pt",
            )
            inputs = inputs.to(model.device)

            with torch.no_grad():
                outputs = model.generate(**inputs, max_new_tokens=10)
            generated_ids = outputs[:, len(inputs.input_ids[0]):]
            image_text = self.processor.batch_decode(
                generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True
            )

            image_des = (image_text[0]).strip()
            return image_des
        except Exception as e:
@@ -239,9 +313,9 @@
            }
            response = requests.post(ragurl + "/chat", json=search_data)
            results = response.json().get('data')
            # self.logger.info(results)
            ret = re.sub(r'<think>.*?</think>', '', results, flags=re.DOTALL)
            ret = ret.replace(" ", "").replace("\t", "").replace("\n", "")
            #self.logger.info(f"{rule_text}:{ret}")
            is_waning = 0
            if len(ret) > 2:
                is_waning = 1
@@ -271,6 +345,7 @@
                "max_tokens": max_tokens
            }
        }
        #self.logger.info(content)
        response = requests.post(ragurl + "/chat", json=search_data)
        # 从json提取data字段内容
        ret = response.json()["data"]