pq_video_cover/common/cover_gpt4o.py

import time

import requests
import json


class CoverAnalyzer:
    @staticmethod
    def cover_list_data(cover_list):
        list_data = []
        for cover in cover_list:
            data_dict = CoverAnalyzer.analyze_covers(str(cover))
            if data_dict:
                list_data.append(data_dict)
        if list_data:
            return list_data
        else:
            return None

    @staticmethod
    def analyze_covers(cover):
        for i in range(3):
            url = "http://aigc-api.cybertogether.net//aigc/dev/test/gpt"
            payload = json.dumps({
                "imageList": [cover],
                # "prompt": "请帮我完成图片分析工作",
                "prompt": "请帮我完成图片分析工作：你只需要理解图片正中心600 × 480的区域画面，根据该区域画面和特征，给出相应的回答",
                "responseFormat": {
                    "type": "json_schema",
                    "json_schema": {
                        "strict": True,
                        "name": "comprehension_result",
                        "schema": {
                            "type": "object",
                            "properties": {
                                "是否包含老年人为画面主体": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否包含老年人为画面主体，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否有多人为画面主体": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否有多人为画面主体，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否有超过3人为画面主体": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否有超过3人为画面主体，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否充斥画面的密集人群为画面主体": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否充斥画面的密集人群为画面主体，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否存在知名历史、近代人物": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否存在知名历史、近代人物，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否存在人物脸部、头部未完整出现在画面的情况": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否存在人物脸部、头部未完整出现在画面的情况，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否是不以人为主体的鲜花、美景、知名建筑或风景": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否是不以人为主体的鲜花、美景、知名建筑或风景，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否是老照片、怀旧风格": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否是老照片、怀旧风格，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否是农村、军事、综艺演出、历史画面": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否是农村、军事、综艺演出、历史画面，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否有趣味、惊奇的形象或画面为主体": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否有趣味、惊奇的形象或画面为主体，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否以大号文字或密集文字为主体并且不包含人物": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否以大号文字或密集文字为主体并且不包含人物，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否是不包含人物的纯色画面": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否是不包含人物的纯色画面，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                },
                                "是否是模糊的或清晰度、像素较低的": {
                                    "type": "object",
                                    "properties": {
                                        "result": {
                                            "description": "请分析所有图片，是否是模糊的或清晰度、像素较低的，如果是返回1，如果不是返回0",
                                            "type": "number",
                                            "enum": [1, 0]
                                        },
                                        "reason": {
                                            "description": "得出结论的客观的分析原因和依据",
                                            "type": "string"
                                        }
                                    },
                                    "required": ["result", "reason"],
                                    "additionalProperties": False
                                }
                            },
                            "additionalProperties": False,
                            "required": [
                                "是否包含老年人为画面主体",
                                "是否有多人为画面主体",
                                "是否有超过3人为画面主体",
                                "是否充斥画面的密集人群为画面主体",
                                "是否存在知名历史、近代人物",
                                "是否存在人物脸部、头部未完整出现在画面的情况",
                                "是否是不以人为主体的鲜花、美景、知名建筑或风景",
                                "是否是老照片、怀旧风格",
                                "是否是农村、军事、综艺演出、历史画面",
                                "是否有趣味、惊奇的形象或画面为主体",
                                "是否以大号文字或密集文字为主体并且不包含人物",
                                "是否是不包含人物的纯色画面",
                                "是否是模糊的或清晰度、像素较低的"
                            ]
                        }
                    }
                }
            })
            headers = {'Content-Type': 'application/json'}
            try:
                response = requests.post(url, headers=headers, data=payload)
                response_data = response.json()

                data = json.loads(response_data.get('data', '{}'))

                result_sum = sum(item['result'] for item in data.values())  # 获取总和
                data_dict = {
                    "data": str(data),
                    "cover_url": cover,
                    "cover_sum": result_sum
                }
                return data_dict
            except requests.exceptions.RequestException as e:
                print(f"请求出错: {e}，封面链接:{cover}")
                time.sleep(1)
                continue
            except json.JSONDecodeError as e:
                print(f"JSON 解析错误: {e}，封面链接:{cover}")
                time.sleep(1)
                continue
            except Exception as e:
                print(f"其他错误: {e}，封面链接:{cover}")
                time.sleep(1)
                continue

        return None