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

---

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]"
1
2

2.编写马赛克核心代码

代码如下：

from fastapi import HTTPException
import cv2
import os

def 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):
        return
    for 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)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

---

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 pickle

app = 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)
    #check
    if 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_type
    if 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 name
    data = await file.read()
    timestr = time.strftime("%Y%m%d%H%M%S",time.localtime())
    fname = file.filename.split('.')
    folderpath = "files/" + uid
    newfname = "%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 file
            f.write(data)
            f.close()
    except e:
        raise HTTPException(status_code=500, detail= e)

    #set result body
    result = PicItem()
    result.uid = uid
    result.filename = newfname
    #resultstr = pickle.dumps(result).decode('UTF-8','ignore')
    #logger.info("return create_upload_file uid:%s response:%s" % (uid, resultstr))
    return result
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62

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 savedname


def 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 size
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)  # convert to grey
    print("read file success")
    # laod opencv schema
    classifier = cv2.CascadeClassifier("haarcascade_frontalface_default.xml")
    color = (0, 255, 0)
    # begin to identify face
    faceRects = classifier.detectMultiScale(gray, scaleFactor=1.3, minNeighbors=3, minSize=(32, 32))
    print(faceRects)
    if len(faceRects):  # get faces if above zero
        print("get faces")
        for faceRect in faceRects:  # loop each face
            x, y, w, h = faceRect
            do_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
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34

三、测试

原照片：

处理后的照片：

总结

以上就是今天要讲的内容，本文仅仅简单介绍了python是如何将照片马赛克化的，然后使用fastapi开发接口暴露给用户使用。