first commit

.gitignore

@@ -0,0 +1,8 @@
+.DS_Store
+.idea
+*.log*
+logs
+__pycache__
+问题.txt
+helper.py
+app.py

.pydio

@@ -0,0 +1 @@
+8ff2f803-d830-4ae6-af02-b18c8d3942a2

README.md

@@ -0,0 +1,35 @@
+## 详细使用文档写完再写！
+
+接口文档：http://127.0.0.1:8000/docs。
+
+fairseq 依赖 需要 windows 先安装 vs_BuildTools.exe
+
+decord 用 eva-decord 替代
+https://github.com/dmlc/decord/issues/213
+
+python -B main.py
+
+uvicorn main:app --reload
+
+celery -A src.celery.main.app worker -l info -P eventlet 
+
+celery --broker=redis://:lb714500@127.0.0.1:6379/0 flower --port=5551 
+
+python -mvenv venv
+source venv/bin/activate
+
+windows redis 配置
+redis-server.exe redis.windows.conf
+# 卸载服务：
+redis-server --service-uninstall
+# 开启服务：
+redis-server --service-start
+# 停止服务：
+redis-server --service-stop
+# 重命名服务：
+redis-server --service-name name
+
+mac redis 配置
+软件 /usr/local/Cellar/redis/7.2.4
+redis的配置文件 /usr/local/etc/redis.conf
+启动 redis-server /usr/local/etc/redis.conf

main.py

@@ -0,0 +1,44 @@
+# encoding: utf-8
+from fastapi import FastAPI
+from fastapi.staticfiles import StaticFiles
+from fastapi.middleware.cors import CORSMiddleware
+from src.controller import api_router
+from src.config import *
+from src.utils import *
+
+
+def startApp():
+    # 接口文档配置
+    app: FastAPI = FastAPI(
+        title=settings.PROJECT_NAME,
+        version=settings.PROJECT_VERSION,
+        description=settings.PROJECT_DESC
+    )
+
+    # 配置路由
+    app.include_router(api_router, prefix=settings.API_PREFIX)
+    # 配置静态资源目录
+    app.mount("/" + settings.STATIC_DIR, StaticFiles(directory=settings.STATIC_DIR), name=settings.STATIC_DIR)
+    # 跨域配置
+    app.add_middleware(
+        CORSMiddleware,
+        allow_origins=settings.CORS_ORIGINS,
+        allow_credentials=True,
+        allow_methods=["*"],
+        allow_headers=["*"],
+    )
+
+    return app
+
+
+app = startApp()
+
+if __name__ == "__main__":
+    import uvicorn
+
+    uvicorn.run(
+        app=settings.APP_NAME,
+        host=settings.HOST,
+        port=settings.PORT,
+        reload=settings.RELOAD
+    )

requirements.txt

@@ -0,0 +1,21 @@
+fastapi==0.109.2
+uvicorn==0.27.1
+pydantic==2.6.1
+pydantic-settings==2.2.0
+requests==2.31.0
+
+torch==2.2.0
+opencv-python==4.9.0.80
+open-clip-torch==2.24.0
+transformers==4.37.2
+librosa==0.10.1
+fairseq==0.12.2
+unicodedata2==15.1.0
+zhconv==1.4.3
+modelscope==1.12.0
+rapidfuzz==3.6.1
+eva-decord==0.6.1
+flower==2.0.1
+redis==5.0.1
+celery==5.3.6
+eventlet==0.35.1

requirements_back.txt

@@ -0,0 +1,28 @@
+clip_client>=0.8.3
+clip_server>=0.8.3
+docarray>=0.19.0
+ImageHash==4.2.1
+numpy==1.26.4
+Pillow==9.2.0
+streamlit==1.11.1
+av
+altair<5
+
+fastapi==0.109.2
+uvicorn==0.27.1
+pydantic==2.6.1
+pydantic-settings==2.2.0
+requests==2.31.0
+fastapi-socketio==0.0.10
+
+torch==2.2.0
+opencv-python==4.9.0.80
+open-clip-torch==2.24.0
+transformers==4.37.2
+librosa==0.10.1
+fairseq==0.12.2
+unicodedata2==15.1.0
+zhconv==1.4.3
+modelscope==1.12.0
+rapidfuzz==3.6.1
+decord>=0.6.0

run.bat

@@ -0,0 +1,59 @@
+@echo off
+
+:menu
+cls
+echo. 
+echo. AI<41><49><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>˵<EFBFBD>
+echo. 
+echo.  1: <20><><EFBFBD><EFBFBD> FastApi 
+echo.  2: <20><><EFBFBD><EFBFBD> Celery 
+echo.  3: <20><><EFBFBD><EFBFBD> Flower WebUi
+echo.  4: <20><><EFBFBD><EFBFBD> Redis Server
+echo.  5: ȫ<><C8AB> <20><><EFBFBD><EFBFBD>
+ECHO.
+
+set/p option="<EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>ѡ<EFBFBD><EFBFBD>:"
+if "%option%"=="1" (
+    call:fastapi 
+    goto menu
+)
+if "%option%"=="2" (
+    call:celery 
+    goto menu
+)
+if "%option%"=="3" (
+    call:flower 
+    goto menu
+)
+if "%option%"=="4" (
+    call:redis 
+    goto menu
+)
+if "%option%"=="5" (
+    call:all 
+    goto menu
+)
+
+:fastapi
+start cmd /k "echo <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> FastApi<70><69><EFBFBD><EFBFBD><EFBFBD>Ե<EFBFBD>... && python -B main.py"
+goto:eof
+
+:celery
+start cmd /k "echo <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> Celery<72><79><EFBFBD><EFBFBD><EFBFBD>Ե<EFBFBD>... && celery -A src.celery.main.app worker -l info -P eventlet"
+goto:eof
+
+:flower
+start cmd /k "echo <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> Flower<65><72><EFBFBD><EFBFBD><EFBFBD>Ե<EFBFBD>... && celery --broker=redis://:lb714500@127.0.0.1:6379/0 flower --port=5551"
+goto:eof
+
+:redis
+start cmd /k "echo <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> Redis Server<65><72><EFBFBD><EFBFBD><EFBFBD>Ե<EFBFBD>... && cd C:\Program Files\Redis && redis-server.exe redis.windows.conf"
+goto:eof
+
+:all
+call:redis
+call:fastapi
+call:celery
+call:flower
+goto menu
+goto:eof

run.sh

@@ -0,0 +1,74 @@
+#!/bin/bash
+
+
+
+newTerminal() {
+  osascript 2>/dev/null <<EOF
+    tell application "System Events"
+      tell process "Terminal" to keystroke "t" using command down
+    end
+    tell application "Terminal"
+      activate
+      do script with command "cd \"$PWD\"; $*" in window 1
+    end tell
+EOF
+}
+
+fastapi(){
+  newTerminal "echo 正在启动 FastApi，请稍等... && $PWD/venv/bin/python -B main.py"
+}
+
+celery(){
+    echo "echo 正在启动 Celery，请稍等... && celery -A src.celery.main.app worker -l info -P eventlet"
+}
+
+flower(){
+    echo "echo 正在启动 Flower，请稍等... && celery --broker=redis://:lb714500@127.0.0.1:6379/0 flower --port=5551"
+}
+
+redis(){
+    echo "echo 正在启动 Redis Server，请稍等... && cd C:\Program Files\Redis && redis-server.exe redis.windows.conf"
+}
+
+all(){
+    redis
+    fastapi
+    celery
+    flower
+}
+
+name='0'
+menu() {
+  echo
+  echo  AI爬虫 批处理启动菜单
+  echo
+  echo    1: 启动 FastApi
+  echo    2: 启动 Celery
+  echo    3: 启动 Flower WebUi
+  echo    4: 启动 Redis Server
+  echo    5: 全部 启动
+  echo
+  read -p "请输入你的选择:" name
+}
+
+menu
+
+if [ $name == "1" ]
+then
+   fastapi
+elif [ $name == "2" ]
+then
+   celery
+elif [ $name == "3" ]
+then
+   flower
+elif [ $name == "4" ]
+then
+   redis
+elif [ $name == "5" ]
+then
+   all
+else
+   echo "没有符合的条件"
+fi
+

src/.pydio

@@ -0,0 +1 @@
+3ac36a1c-7d73-47bf-af4e-1e05933b65bd

src/ai/.pydio

@@ -0,0 +1 @@
+0e6c5784-911f-49db-b759-70844e9d3d6e

src/ai/ai_test.py

@@ -0,0 +1,26 @@
+import torch
+from modelscope.utils.constant import Tasks
+from modelscope.pipelines import pipeline
+from modelscope.preprocessors.image import load_image
+
+pipeline = pipeline(task=Tasks.multi_modal_embedding, model='damo/multi-modal_clip-vit-large-patch14_336_zh', model_revision='v1.0.1', device='cpu')
+input_img = load_image('https://p1.a.yximgs.com/upic/2023/12/27/13/BMjAyMzEyMjcxMzAwMzdfMjkwOTEwNDc5N18xMjA2NTYyODIyNDRfMl8z_Be33b730fcb8aaf000ba5e58bf8c3681e.jpg?tag=1-1708460644-unknown-0-huswbkiuiq-0ac88680ace16101&clientCacheKey=3x9ghy3dj8y2z5a.jpg&di=70208a94&bp=14764') # 支持皮卡丘示例图片路径/本地图片 返回PIL.Image
+input_texts = ["小火龙", "白色衬衫"]
+
+# 支持一张图片(PIL.Image)或多张图片(List[PIL.Image])输入，输出归一化特征向量
+img_embedding = pipeline.forward({'img': input_img})['img_embedding'] # 2D Tensor, [图片数, 特征维度]
+
+# 支持一条文本(str)或多条文本(List[str])输入，输出归一化特征向量
+text_embedding = pipeline.forward({'text': input_texts})['text_embedding'] # 2D Tensor, [文本数, 特征维度]
+
+# 计算图文相似度
+with torch.no_grad():
+    # 计算内积得到logit，考虑模型temperature
+    logits_per_image = (img_embedding / pipeline.model.temperature) @ text_embedding.t()
+    # 根据logit计算概率分布
+    probs = logits_per_image.softmax(dim=-1).cpu().numpy()
+
+# similarityNum = probs[0][1] * 100
+print("图文匹配概率:", probs[0])
+# result = True if similarityNum >= 85 else False
+# print("图文匹配概率:", result)

src/ai/imageCaptioning.py

@@ -0,0 +1,53 @@
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+
+class imageCaptioning:
+    # AI模型的本地地址
+    modelName = 'damo/mplug_image-captioning_coco_base_zh'
+    # 使用 cpu 还是 GPU 计算，GPU 比 CPU快，但是 GPU 只支持 英伟达 显卡，至少4G显存以上
+    deviceType = 'cpu' # cpu, gpu
+    #  外部传入 需要检索的关键词数组
+    keyword = []
+    # 外部传入 的视频帧 截图数据，[ 'http地址 或者 本地地址', '' ]
+    # 地址可以用 video标签 poster 预览图，或者使用 ffmpeg，docarray，这里推荐 使用 多种模态数据结构工具包 docarray
+    # 具体 docarray 封装看 utils/screenshot.py
+    imgArr = []
+    # 符合 条件 数据的计数
+    count = 0
+    # 符合 条件 数据的list下标
+    haveIndex = []
+
+    def __init__(self, key, imgList):
+        self.keyword = key
+        self.imgArr = imgList
+
+    def start(self):
+        self.count = 0
+        self.haveIndex = []
+    
+        for index, item in enumerate(self.imgArr):
+            print()
+            print("ai 得到的图片为：", item)
+            res = self.mplugImage(item)
+            print("ai 解析图片内容为：", res)
+            for v in self.keyword:
+                if v in res:
+                    self.count += 1
+                    self.haveIndex.append(index)
+
+
+        if self.count != 0 and len(self.haveIndex) != 0:
+            # print('包含次数:', self.count)
+            return {'count': self.count, 'index': self.haveIndex}
+        else:
+            return {'count': 0, 'index': []}
+
+      
+
+
+
+    def mplugImage(self, url):
+        pipeline_caption = pipeline(Tasks.image_captioning, model=self.modelName, device=self.deviceType)
+        result = pipeline_caption(url)
+        str = ''.join(result['caption'].split())
+        return str.replace("。", "")

src/ai/multimModaImageTextSearch.py

@@ -0,0 +1,75 @@
+import torch
+from modelscope.utils.constant import Tasks
+from modelscope.pipelines import pipeline
+from modelscope.preprocessors.image import load_image
+from fastapi import WebSocket
+
+class multimModaImageTextSearch:
+    # AI模型的本地地址
+    modelName: str = 'damo/multi-modal_clip-vit-large-patch14_336_zh'
+    # 使用 cpu 还是 GPU 计算，GPU 比 CPU快，但是 GPU 只支持 英伟达 显卡，至少4G显存以上
+    deviceType: str = 'cpu' # cpu, gpu
+    # 干扰词，干扰词越少见，搜索结果的准确度越高
+     # 但不是越少见越多就越好，需要用数据测试，找到合适的
+    noiseWord: list[str] = ['小火龙']
+    # 阈值，数值越大，搜索结果的准确度越高，
+    # 但不是越大越好，需要用数据测试，找到合适的
+    thresholdValue: int = 85
+    #  外部传入 需要检索的关键词数组
+    keyword: str = ""
+    # 外部传入 的视频帧 截图数据，[ 'http地址 或者 本地地址', '' ]
+    # 地址可以用 video标签 poster 预览图，或者使用 ffmpeg，docarray，这里推荐 
+    # 使用 多种模态数据结构工具包 docarray，具体 docarray 封装看 utils/screenshot.py
+    imgArr: list[str] = []
+    # 符合 条件 数据的计数
+    count: int = 0
+    # 符合 条件 数据的list下标
+    haveIndex = []
+
+    def __init__(self, key, imgList):
+        self.keyword = key
+        self.imgArr = imgList
+
+    def start(self):
+        self.count = 0
+        self.haveIndex = []
+
+        for index, item in enumerate(self.imgArr):
+            print()
+            print("关键词: ", self.keyword)
+            print(f"下标：{index}，图片地址为：{item}")
+            res = self.imageTextSearch(item)
+            print(f"下标：{index}，图片&文字是否相似：{res}")
+            if res:
+                self.count += 1
+                self.haveIndex.append(index)
+
+        if self.count != 0 and len(self.haveIndex) != 0:
+            return {'count': self.count, 'index': self.haveIndex}
+        else:
+            return {'count': 0, 'index': []}
+        
+    def imageTextSearch(self,url):
+        pipeline_caption = pipeline(
+            task=Tasks.multi_modal_embedding, 
+            model=self.modelName, 
+            model_revision='v1.0.1', 
+            device=self.deviceType
+        )
+        # 支持网络地址 & 本地地址
+        input_img = load_image(url) 
+        # input_texts = [self.noiseWord, self.keyword]
+        input_texts = [self.keyword]+[item for item in self.noiseWord]
+
+        img_embedding = pipeline_caption.forward({'img': input_img})['img_embedding'] 
+        text_embedding = pipeline_caption.forward({'text': input_texts})['text_embedding']
+
+        with torch.no_grad():
+            logits_per_image = (img_embedding / pipeline_caption.model.temperature) @ text_embedding.t()
+            probs = logits_per_image.softmax(dim=-1).cpu().numpy()
+
+        similarityNum = probs[0][0] * 100
+        print(f"图片&文字相似度数值：{similarityNum}")
+        result = True if similarityNum >= self.thresholdValue else False
+        return result
+              

src/ai/test.py

@@ -0,0 +1,22 @@
+from multimModaImageTextSearch import multimModaImageTextSearch
+
+imgList=[
+        'https://p3.a.yximgs.com/upic/2024/02/17/17/BMjAyNDAyMTcxNzUwMjVfMTIxMjEyNTM0OF8xMjUyMDI5MzE4NjlfMV8z_B0817bda8480e6981ee36e3e03e430b0e.jpg?tag=1-1708294678-unknown-0-cixqi9cbxk-c689c31c32f2ffdc&clientCacheKey=3xq7x3n3qn37ebk.jpg&di=70208a94&bp=14764',
+        'https://p3.a.yximgs.com/upic/2024/01/06/07/BMjAyNDAxMDYwNzMwNTFfMjc2MjcyOTk5N18xMjE0NDE2NDk4MDlfMl8z_B05a11d8cf566915c63399ed4e5344807.jpg?tag=1-1708294678-unknown-0-gwntq0bmmi-e2f16df83fcd9f1d&clientCacheKey=3xncgpqst6dcquk.jpg&di=70208a94&bp=14764',
+        'https://p2.a.yximgs.com/upic/2022/09/07/10/BMjAyMjA5MDcxMDI3MzJfMTQzMjQ3NzA2XzgzNjQzMjMwNzA3XzJfMw==_B010362f63eebfbf8bf83c119e56adc32.jpg?tag=1-1708300460-unknown-0-eevoqueweh-a6cf9d3f78363d23&clientCacheKey=3x9455m5a6kijdu.jpg&di=70208a94&bp=14764',
+
+        'https://p1.a.yximgs.com/upic/2023/11/25/16/BMjAyMzExMjUxNjEyMjhfMTkwMDU1MjI2Ml8xMTgwMzUwNjE5OTdfMV8z_B3b6eaa07a54295924840b2676e4fcf5c.jpg?tag=1-1708300460-unknown-0-xmixtw3cxl-725ff520d02821b3&clientCacheKey=3xfave2rujw6mhk.jpg&di=70208a94&bp=14764',
+        'https://p1.a.yximgs.com/upic/2023/10/23/19/BMjAyMzEwMjMxOTM0NDVfMzQ3NTg0NzAyMV8xMTU2MjM4NDU3MjFfMV8z_B6cda66f76f7c07bafcae503472960ecf.jpg?tag=1-1708300527-unknown-0-dnxzlo1pni-409704b08667fd22&clientCacheKey=3xwa8mv9hcr53xc.jpg&di=70208a94&bp=14764',
+    ]
+
+image = multimModaImageTextSearch(
+    key="牛仔裤",
+    imgList=imgList
+)
+
+res = image.start()
+
+print("输入AI的总视频数据长度: ", len(imgList))
+print("AI匹配关键词后得到数据为: ", res)
+
+

src/celery/.pydio

@@ -0,0 +1 @@
+8743ab62-0f5d-4dea-8f07-40d2740a8f88

src/celery/ai_image/.pydio

@@ -0,0 +1 @@
+ef314510-8071-42fa-8309-fbfcc2040b93

src/celery/ai_image/__init__.py

@@ -0,0 +1 @@
+from .task import start_ai_img

src/celery/ai_image/task.py

@@ -0,0 +1,36 @@
+from src.celery.main import app
+from src.ai.multimModaImageTextSearch import multimModaImageTextSearch
+import celery
+from src.utils import *
+
+class MyTask(celery.Task):
+    # 任务失败时执行
+    def on_failure(self, exc, task_id, args, kwargs, einfo):
+        print('{0!r} failed: {1!r}'.format(task_id, exc))
+    # 任务成功时执行
+    def on_success(self, retval, task_id, args, kwargs):
+        # print("--------------------------: ",socketio.app())
+        print("task_id: ", task_id)
+        print("retval: ", retval)
+        print("args: ", args)
+        print("kwargs: ", kwargs)
+        print("任务成功时执行")
+        pass
+    # 任务重试时执行
+    def on_retry(self, exc, task_id, args, kwargs, einfo):
+        pass
+
+@app.task(base=MyTask, bind=True)
+def start_ai_img(self, keyWord: str, imgList: list[str]):
+    try:
+        image = multimModaImageTextSearch(
+            key=keyWord,
+            imgList=imgList
+        )
+        res = image.start()
+        return res
+    
+    except Exception as exc:
+        raise self.retry(exc=exc)
+  
+  

src/celery/config.py

@@ -0,0 +1,10 @@
+# config.py => 配置文件
+# 配置中的CELERY_为特定前缀
+# https://docs.celeryproject.org/en/latest/genindex.html
+BROKER_URL = 'redis://:lb714500@127.0.0.1:6379/0' # Broker配置
+CELERY_RESULT_BACKEND = 'redis://:lb714500@127.0.0.1:6379/0' # BACKEND配置
+CELERY_RESULT_SERIALIZER = 'json' # 结果序列化方案
+CELERY_TASK_RESULT_EXPIRES = 60 * 60 * 24 # 任务过期时间
+CELERY_TIMEZONE='Asia/Shanghai' # 时区配置
+
+# CELERY_IMPORTS = ('src.celery.ai_image.task',) # 指定导入的任务模块

src/celery/main.py

@@ -0,0 +1,10 @@
+from celery import Celery
+
+# 创建Celery实例且名称为 ai_image
+app = Celery('ai_image') 
+# 从配置文件config.py中加载配置参数
+app.config_from_object('src.celery.config') 
+# 注册任务
+app.autodiscover_tasks([
+    'src.celery.ai_image.task'
+])

src/config/.pydio

@@ -0,0 +1 @@
+13b4bc8f-2ce7-468e-afff-c5db791c1c28

src/config/__init__.py

@@ -0,0 +1 @@
+from .setting import *

src/config/setting.py

@@ -0,0 +1,41 @@
+from pydantic import AnyHttpUrl
+from fastapi import FastAPI
+from pydantic_settings import BaseSettings
+
+IS_DEV = True  # 是否开发环境
+
+class Settings(BaseSettings):
+   
+    PROJECT_NAME: str = "AiReptile"  # 项目名称 必填
+    PROJECT_DESC: str = "🎉 接口汇总 🎉"  # 描述
+    PROJECT_VERSION: int | str = 1.0  # 版本
+    API_PREFIX: str = "/api/v1"  # 接口前缀
+
+    APP_NAME: str = "main:app"
+    HOST: str = "0.0.0.0"     # 允许访问程序的ip， 只允许本地访问使用 127.0.0.1， 只在直接允许程序时候生效
+    PORT: int = 8000    # 程序端口，只在直接运行程序的时候生效
+    RELOAD: bool = True     # 是否自动重启，只在直接运行程序时候生效
+
+    CORS_ORIGINS: list[str] = ['*']  # 跨域请求(务必指定精确ip, 不要用localhost)
+    MOUNT_LOCATION: str = '/ws'  
+
+    STATIC_DIR: str = "static"  # 静态文件目录
+    BASE_URL: AnyHttpUrl = "http://127.0.0.1:8000"  # 开发环境(为了存放图片全路径)
+
+
+class DevelopmentConfig(Settings):
+    pass
+
+class ProductionConfig(Settings):
+    BASE_URL: AnyHttpUrl = "http://114.115.143.81:8000"  # 生产环境(为了存放图片全路径)
+    CORS_ORIGINS: list[AnyHttpUrl] = ["http://114.115.143.81"]  # 跨域请求
+
+    # REDIS_URI: str = "redis://:123456@redis:6379/0"  # Redis
+    # DATABASE_URI: str = "mysql://root:123456@mysql:3306/demo?charset=utf8"  # MySQL
+    # DATABASE_ECHO: bool = True  # 是否打印数据库日志 (可看到创建表、表数据增删改查的信息)
+
+    # LOGGER_LEVEL: str = 'INFO'  # 日志等级: ['DEBUG' | 'INFO']
+
+
+appInatanc: FastAPI
+settings = DevelopmentConfig() if IS_DEV else ProductionConfig()

src/controller/.pydio

@@ -0,0 +1 @@
+92111a44-fa81-47ee-91d8-c1021373be45

src/controller/__init__.py

@@ -0,0 +1,12 @@
+from fastapi import APIRouter
+
+from .kuaishou.kuaishou import router as kuaishou_api
+from .ai.ai import router as ai_api
+
+
+api_router = APIRouter()
+
+api_router.include_router(kuaishou_api, prefix="/ks")
+api_router.include_router(ai_api, prefix="/ai")
+
+__all__ = ['api_router']

src/controller/ai/.pydio

@@ -0,0 +1 @@
+da633ebd-90ac-4bea-80b9-b86147bdc5fd

src/controller/ai/ai.py

@@ -0,0 +1,46 @@
+from fastapi import APIRouter
+from src.celery.ai_image import start_ai_img
+from src.controller.ai.schemas import textSearchVideoSchemas
+from src.utils import *
+
+router = APIRouter()
+
+
+# http://127.0.0.1:8000/api/v1/ai/text_search_video
+@router.post("/text_search_video")
+async def textSearchVideo(params: textSearchVideoSchemas):
+    task = start_ai_img.delay(params.keyWord, params.imgList)
+    return { 
+        'task_id': task.id,
+        'message': 'ai任务已添加后台'
+    }
+
+# @router.websocket("/ws")
+# async def task_status(websocket: WebSocket):
+#     await websocket.accept()
+#     while True:
+#         data = await websocket.receive_json()
+#         while True:
+#             res = getTaskStaus(data['task_id'])
+#             print("2 循环查找任务 res：", res)
+#             if res['status']:
+#                 await websocket.send_json(res)
+#                 break;
+#             time.sleep(3)
+        
+
+# http://127.0.0.1:8000/api/v1/ai/task_status?task_id=c4b57ff1-0ca4-448e-ad95-e3d549841842
+@router.get("/task_status")
+def taskStatus(task_id: str):
+    return getTaskStaus(task_id)
+    
+
+def getTaskStaus(id: str):
+    async_result = start_ai_img.AsyncResult(id)
+    if async_result.successful():
+        return async_result.get() | { 'status': True }
+    else:
+        return {
+            'status': False,
+            'message': tool.getState(async_result.status)
+        }

src/controller/ai/schemas/.pydio

@@ -0,0 +1 @@
+f407fdd2-8a42-450d-9051-57c793797b75

src/controller/ai/schemas/__init__.py

@@ -0,0 +1 @@
+from src.controller.ai.schemas.ai_schemas import *

src/controller/ai/schemas/ai_schemas.py

@@ -0,0 +1,5 @@
+from pydantic import BaseModel, HttpUrl
+
+class textSearchVideoSchemas(BaseModel):
+    keyWord: str | None = None
+    imgList: list[str]

src/controller/kuaishou/.pydio

@@ -0,0 +1 @@
+c9933e83-8290-4b3c-aaaf-da0611a06ac4

src/controller/kuaishou/kuaishou.py

@@ -0,0 +1,37 @@
+from fastapi import APIRouter
+from ...http import *
+from src.celery.ai_image import start_ai_img
+
+
+router = APIRouter()
+
+# http://127.0.0.1:8000/api/v1/ks/like
+@router.get("/like")
+async def LikeVideo(pcursor: str = ''):
+    """
+      获取账号下点过赞的所有视频
+
+      - **pcursor**: 下一页的页标
+      \f
+      :param item: User input.
+      """
+    res = await http.post(
+        { 'page': "profile", 'pcursor': pcursor }, 
+        httpConfig.visionProfileLikePhotoList
+    )
+
+    # feeds = res['feeds']
+
+    # 处理原始数据，构造视频 & 图片
+    # imgSourceUrl = []
+    # for index, item in enumerate(feeds):
+    #     imgSourceUrl.append(item['photo']['coverUrl'])
+        # imgSourceUrl.append({ 
+        #   'id': item['photo']['id'],
+        #   'imgSrc': item['photo']['coverUrl'], 
+        #   'videoSrc': item['photo']['photoUrl'] 
+        # })
+
+    # task = start_ai_img.delay(keyWord,imgSourceUrl)
+        
+    return res

src/controller/kuaishou/schemas/.pydio

@@ -0,0 +1 @@
+63696c28-b9ab-493f-a1ef-f01fef0a0334

src/http/.pydio

@@ -0,0 +1 @@
+43a32678-2504-4c0b-9355-bccef42511ca

src/http/__init__.py

@@ -0,0 +1,2 @@
+from ._config import httpConfig
+from .http import http

src/http/_config.py

@@ -0,0 +1,54 @@
+class HttpConfig:
+    url: str = "https://www.kuaishou.com/graphql"
+    header: list[str,str] = {
+        'Content-Type': 'application/json',
+        'Origin': 'https://www.kuaishou.com',
+        'Host': 'www.kuaishou.com',
+        'Referer': 'https://www.kuaishou.com/profile/3x2347wkfukzx4g',
+        'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36',
+        'Cookie': 'kpf=PC_WEB; clientid=3; did=web_fca770657d0a24abb53ddc92bb42e612; didv=1700407978000; userId=58344290; kuaishou.server.web_st=ChZrdWFpc2hvdS5zZXJ2ZXIud2ViLnN0EqABFs7urbq3vvcrvB-IRBSKHXfdVRzblSFx7udepTzeAtgCompa6-aZiv1f_Bb66SubMkCbLODJQ3cn_GxlWWWbPgbQEyul9f7SDKsGphyvK1Pk3aj-yn899D9gR-xj1wiqT_0nA9x3XcPmvuJ8qQzIeLl3236npeMV0i9uYEAeIHDp9RtvPlQ4o0piktgGjMZH13Uhk7-ZfqDuUPQ-mEH7axoSS2UWrTnDahVDKzRYjzjLpJM-IiDXKnEX31rxM9oVovhBLMPSocxSvtN7wyo1oS1nNdvjSCgFMAE; kuaishou.server.web_ph=5b6a8c20ee92347d220cbc98f705f92042fb; kpn=KUAISHOU_VISION'
+    }
+    visionProfileLikePhotoList: str = ''' fragment photoContent on PhotoEntity { 
+        __typename 
+        id duration 
+        caption originCaption 
+        likeCount viewCount commentCount realLikeCount 
+        coverUrl photoUrl photoH265Url manifest 
+        manifestH265 videoResource coverUrls 
+        { 
+            url __typename 
+        } 
+        timestamp expTag animatedCoverUrl distance 
+        videoRatio liked stereoType profileUserTopPhoto 
+        musicBlocked riskTagContent riskTagUrl 
+    } 
+    fragment recoPhotoFragment on recoPhotoEntity { 
+        __typename id duration caption originCaption likeCount 
+        viewCount commentCount realLikeCount coverUrl photoUrl photoH265Url manifest
+        manifestH265 videoResource coverUrls 
+        { url __typename } 
+        timestamp expTag animatedCoverUrl distance videoRatio liked stereoType 
+        profileUserTopPhoto musicBlocked riskTagContent riskTagUrl 
+    } 
+    fragment feedContent on Feed { 
+        type author 
+        { 
+            id name headerUrl following headerUrls 
+            { 
+                url __typename 
+            } __typename 
+        } photo { 
+            ...photoContent ...recoPhotoFragment __typename 
+        } canAddComment llsid status currentPcursor tags 
+        { type name __typename } 
+        __typename 
+    } 
+    query visionProfileLikePhotoList($pcursor: String, $page: String, $webPageArea: String) { 
+        visionProfileLikePhotoList(pcursor: $pcursor, page: $page, webPageArea: $webPageArea) { 
+            result llsid webPageArea feeds {
+                ...feedContent __typename 
+                } hostName pcursor __typename 
+        } 
+    }'''
+
+httpConfig = HttpConfig

src/http/http.py

@@ -0,0 +1,17 @@
+import requests
+from . import *
+
+class Http:
+
+    async def post(variables, query):
+        res = requests.post(
+            httpConfig.url,
+            json={
+                'query': query,
+                'variables': variables
+            },
+            headers=httpConfig.header,
+        )
+        return res.json()['data']['visionProfileLikePhotoList']
+    
+http = Http 

src/utils/.pydio

@@ -0,0 +1 @@
+176e28e8-8250-4bc9-8589-5b9a605af8c6

src/utils/__init__.py

@@ -0,0 +1,3 @@
+# from .setInterval import *
+from .tool import *
+# from .socketio import *

src/utils/screenshot.py

@@ -0,0 +1,37 @@
+import os, shutil
+from docarray import Document, DocumentArray
+
+# https://blog.csdn.net/Jina_AI/article/details/128475707
+
+class Screenshot:
+    # 视频名称
+    videoName = ''
+    # 视频所在的 父级路径
+    videoPath = '../../source/'
+    # 视频帧截图后输出的 父级路径
+    outputKeyframes = '../../dist/'
+    d = ''
+    keyframes = DocumentArray()
+
+    def __init__(self, video_name):
+        self.videoName = video_name
+        self.d = Document(uri=self.videoPath + self.videoName).load_uri_to_video_tensor(only_keyframes=False)
+        self.createdDir()
+        self.start()
+
+    def start(self):
+        for i in range(len(self.d.tensor)):
+            if i in self.d.tags['keyframe_indices']:
+                keyframe = Document(
+                    tensor=self.d.tensor[i], tags={'index': len(self.keyframes)}
+                )
+                keyframe.save_image_tensor_to_file(file=f'{self.outputKeyframes}/{self.videoName}/{len(self.keyframes)}.png')
+                self.keyframes.append(keyframe)
+
+
+    def createdDir(self):
+        if not os.path.isdir(self.outputKeyframes + self.videoName):
+            os.makedirs(self.outputKeyframes + self.videoName, exist_ok=True)
+
+
+Screenshot('yjkon.mp4')

src/utils/socketio.py

@@ -0,0 +1,34 @@
+from fastapi import FastAPI
+from fastapi_socketio import SocketManager
+
+class Socketio:
+
+    def __init__(self):
+        pass
+
+    def init(self, app: FastAPI):
+        print("++++++++++++++++++++++++++++++++++: ", app)
+        self._app = app
+        self._manager = SocketManager(app=app, mount_location="/")
+
+        self._manager._sio.on('task_status', self.task_status)
+
+    async def task_status(sid, args, msg):
+        print("task_status sid ++++++++++++++ ", sid)
+        print("task_status args ++++++++++++++ ", args)
+        print("task_status kwargs ++++++++++++++ ", msg)
+
+
+
+    def app(self):
+        return self._app
+    
+    def manager(self):
+        return self._manager
+    
+    @property
+    def emit(self):
+        return self._app.emit
+
+
+socketio = Socketio()   

src/utils/tool.py

@@ -0,0 +1,14 @@
+class Tool:
+    def getState(state):
+        if state == "PENDING":
+            return "任务正在等待执行"
+        elif state == "STARTED":
+            return "任务已启动."
+        elif state == "RETRY":
+            return "将重试该任务，可能是因为失败"
+        elif state == "FAILURE":
+            return "该任务引发异常，或已超过重试限制"
+        elif state == "SUCCESS":
+            return "任务执行成功"
+        
+tool = Tool

static/.pydio

@@ -0,0 +1 @@
+9f9458e7-77f6-4706-bec7-91af6384b28b

static/dist/.pydio

@@ -0,0 +1 @@
+6f35b2a1-e5f9-483a-97b2-ddc726412f0f

static/dist/jk.mp4/.pydio

@@ -0,0 +1 @@
+b8b52532-fea6-4df8-a7bd-3c3bffd863d5

static/dist/linenz.mp4/.pydio

@@ -0,0 +1 @@
+8ccd5d21-5849-4cfd-85cc-99125a20ba3d

static/dist/nz.mp4/.pydio

@@ -0,0 +1 @@
+12c149cd-72c0-43fc-bda3-a192e85a5519

static/dist/rf.mp4/.pydio

@@ -0,0 +1 @@
+d59c3e02-4216-44e8-a5ae-411d771ad1bb

static/dist/yj.mp4/.pydio

@@ -0,0 +1 @@
+95c87bb7-9825-45a5-bc6f-2031d2d96ec3

static/dist/yjkon.mp4/.pydio

@@ -0,0 +1 @@
+5f27e668-eea2-4a2d-9731-cd35a7233a85

static/source/.pydio

@@ -0,0 +1 @@
+6723c002-05dc-4ca3-9174-0a0e7fda82e8

test/.pydio

@@ -0,0 +1 @@
+9f11425b-4d81-41f5-93d7-2609b17d0a1b

test/index.html

@@ -0,0 +1,99 @@
+<!DOCTYPE html>
+<html lang="en">
+
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Document</title>
+    <style>
+        * {
+            margin: 0;
+            padding: 0;
+            list-style: none;
+            text-decoration: none;
+        }
+
+        .list {
+            column-count: 5;
+            column-gap: 10px;
+        }
+
+        .list li {
+            margin-bottom: 10px;
+        }
+
+        .list li img {
+            width: 100%;
+        }
+    </style>
+</head>
+
+<body>
+
+    <input class="search" type="text" value="牛仔裤" placeholder="输入关键词" />
+    <button class="submit">搜索</button>
+
+    <ul class="list">
+
+    </ul>
+
+</body>
+<script src="https://cdn.bootcdn.net/ajax/libs/jquery/3.7.1/jquery.min.js"></script>
+<script src="https://cdn.bootcdn.net/ajax/libs/axios/1.5.0/axios.js"></script>
+<script>
+    let pcursorNum = '1703970033000', feedsList = [], filterList = [], intervalId = 0;
+    
+    axios.defaults.baseURL = 'http://127.0.0.1:8000/api/v1';
+    axios.interceptors.response.use(response => response.data, error => Promise.reject(error));
+
+    // 点击 搜索按钮 使用【视频预览图地址】进行 视频&文字的多模态搜索，
+    // 会在后端进行 ai批量任务计算，接口返回 task_id
+    $(".submit").click(async () => {
+        let { task_id } = await axios.post('/ai/text_search_video', {
+            keyWord: $(".search").val(),
+            imgList: feedsList.map(v => v.photo.coverUrl)
+        })
+        checkTaskStatus(task_id)
+    })
+
+    // 使用 task_id 轮询，查询 任务计算 结果，如果为 true
+    // 结束 轮询，根据后端返回的 搜索结果 index 下标集合从原始数据中，
+    // 取出数据，存储 filterList 并渲染页面
+    async function checkTaskStatus(task_id) {
+        intervalId = setInterval(async () => {
+            let { status, index } = await axios.get('/ai/task_status', {
+                params: { task_id }
+            });
+            if (status) {
+                clearInterval(intervalId)
+                renderItem(index.map(v => feedsList[v]))
+            }
+        }, 3000)
+    }
+
+    async function getIndex() {
+        let { feeds, pcursor } = await axios.get('/ks/like', {
+            params: { pcursor: pcursorNum }
+        })
+
+        pcursorNum = pcursor
+        feedsList = feeds
+        renderItem(feedsList)
+    }
+
+    function renderItem(data) {
+        $(".list").empty()
+        data.forEach((v, i) => {
+            $(".list").append(`
+                <li>
+                    <img src="${v.photo.coverUrl}" alt="">
+                </li>
+            `)
+        });
+    }
+
+    getIndex()
+
+</script>
+
+</html>

test/wsc.py

@@ -0,0 +1,24 @@
+import asyncio
+import socketio
+
+sio = socketio.AsyncClient()
+
+@sio.event
+async def connect():
+    print('connection established')
+
+@sio.event
+async def my_message(data):
+    print('message received with ', data)
+    await sio.emit('my response', {'response': 'my response'})
+
+@sio.event
+async def disconnect():
+    print('disconnected from server')
+
+async def main():
+    await sio.connect('http://localhost:8000')
+    await sio.wait()
+
+if __name__ == '__main__':
+    asyncio.run(main())