FastAPI Python照片打马赛克API

提示：文章写完后，目录可以自动生成，如何生成可参考右边的帮助文档

文章目录

前言
一、FastAPI是什么？
二、使用步骤
- 1.安装FastAPI及Uvicorn
- 2.编写马赛克核心代码
- 3.编写照片上传接口
- 4.编写马赛克处理接口
三、测试
总结

前言

现代社会大家对于自己的隐私越来越注重，我们平时在外拍摄风景照片的时候难免会有其他人入镜。在发朋友圈或者微博的时候，为了保证这些人的肖像权我一般都会做马赛克处理。
使用python库加fastapi可以做一个快速处理的webapi，具体介绍如下。

一、FastAPI是什么？

FastAPI 是一个用于构建 API 的现代、快速（高性能）的 web 框架，使用 Python 3.6+ 并基于标准的 Python 类型提示。它具有快速，高效，简单及标准化的特点。

二、使用步骤

1.安装FastAPI及Uvicorn

命令如下：

pip install fastapi
pip install "uvicorn[standard]"

2.编写马赛克核心代码

代码如下：

from fastapi import HTTPException
import cv2
import osdef do_mosaic(frame, x, y, w, h, neighbor=9):fh, fw = frame.shape[0], frame.shape[1]if (y + h > fh) or (x + w > fw):returnfor i in range(0, h - neighbor, neighbor): for j in range(0, w - neighbor, neighbor):rect = [j + x, i + y, neighbor, neighbor]color = frame[i + y][j + x].tolist()left_up = (rect[0], rect[1])right_down = (rect[0] + neighbor - 1, rect[1] + neighbor - 1)cv2.rectangle(frame, left_up, right_down, color, -1)

3.编写照片上传接口

代码如下：

from fastapi import FastAPI, UploadFile, Form, File, HTTPException, Request
from model import PicItem, PicReq
import time
import os
import mosaic
import logging
import pickleapp = FastAPI()# setup loggers
logger = logging.getLogger("uvicorn.access")
logger.setLevel(logging.INFO)
handler = logging.handlers.RotatingFileHandler("api.log",mode="a",maxBytes = 100*1024, backupCount = 5)
handler.setFormatter(logging.Formatter("%(asctime)s - %(levelname)s - %(message)s"))
logger.addHandler(handler)@app.get("/")
async def hello():logger.info("enter hello")logger.error("error test")return {"message":"hello world!"}@app.post("/uploadfile/")
async def create_upload_file(file: UploadFile = File(), uid: str = Form()):logger.info("enter create_upload_file uid:%s" % uid)#checkif uid == "":logger.error("create_upload_file 未获取到UID")raise HTTPException(status_code=500, detail="未获取到UID")if not file:logger.error("create_upload_file 未获取到文件")raise HTTPException(status_code=500, detail="未获取到文件")mime = file.content_typeif mime != "image/jpeg":logger.error("create_upload_file 只允许图片格式文件")raise HTTPException(status_code=500, detail="只允许图片格式文件")#set file save path and namedata = await file.read()timestr = time.strftime("%Y%m%d%H%M%S",time.localtime())fname = file.filename.split('.')folderpath = "files/" + uidnewfname = "%s.%s" % (timestr, fname[1])try:if not os.path.exists(folderpath):os.makedirs(folderpath)filepath = "%s/%s" % (folderpath, newfname)with open(filepath,"wb") as f:#save filef.write(data)f.close()except e:raise HTTPException(status_code=500, detail= e)#set result bodyresult = PicItem()result.uid = uidresult.filename = newfname#resultstr = pickle.dumps(result).decode('UTF-8','ignore')#logger.info("return create_upload_file uid:%s response:%s" % (uid, resultstr))return result

4.编写马赛克处理接口

代码如下：

@app.post("/go/")
async def mosaic_pic(picReq: PicReq):logger.info("enter go request:%s" % picReq)savedname =  mosaic.use_mosaic(picReq.uid, picReq.filename)logger.info("return go savedname:%s" % savedname)return savednamedef use_mosaic(uid, fname):print("enter use_mosaic")filepath = "files/%s/%s" % (uid, fname)if not os.path.exists(filepath):raise HTTPException(status_code=500, detail="文件不存在")img = cv2.imread(filepath)fixsize = 128 # declare fix sizegray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)  # convert to greyprint("read file success")# laod opencv schemaclassifier = cv2.CascadeClassifier("haarcascade_frontalface_default.xml")color = (0, 255, 0)# begin to identify facefaceRects = classifier.detectMultiScale(gray, scaleFactor=1.3, minNeighbors=3, minSize=(32, 32))print(faceRects)if len(faceRects):  # get faces if above zeroprint("get faces")for faceRect in faceRects:  # loop each facex, y, w, h = faceRectdo_mosaic(img, x, y, w, h)print("finish use_mosaic")sfname = fname.split(".")donename = "%s_done.%s" % (sfname[0], sfname[1])cv2.imwrite("files/%s/%s" % (uid, donename) , img)return donename