学习NestJS开发小程序后台（二）检测图片敏感内容

const wxImgCheck = async (token, file) => {
    const url = `https://api.weixin.qq.com/wxa/img_sec_check?access_token=${token}`;
    const stream = fs.createReadStream(file.path);
    const form = new FormData();
    form.append('media', stream);

    const getHeaders = form => {
        return new Promise((resolve, reject) => {
            form.getLength((err, len) => {
                if (err) { reject(err); };
                const headers = Object.assign({ 'Content-Length': len, }, form.getHeaders());
                resolve(headers);
            });
        })
    }

    try {
        const headers = await getHeaders(form);
        const res = await axios.post(url, form, { headers });
        if (res.data.errcode === 40001 || res.data.errcode === 40013) {
            global.AccessToken.flag = true;
            const newToken = await getToken();
            await wxImgCheck(newToken, file);
        }
        
        return res.data;
    } catch (err) {
        return {
            errcode: 500,
            message: err.message
        }
    }
}

const model = nsfwjs.load("/path/to/model/directory/");

import * as nsfwjs from 'nsfwjs';


const imgbody = document.getElementById('img');
const img = new Image();
img.crossOrigin = 'anonymous';
img.src =
'https://xxxxxxxx.jpg';
imgbody?.appendChild(img);
if (img) {
  const model = await nsfwjs.load('MobileNetV2');
  const predictions = await model.classify(img);
  console.log('Predictions: ', predictions);
}

nsfwjs.classify(img: tf.Tensor3D | ImageData | HTMLImageElement | HTMLCanvasElement | HTMLVideoElement, topk?: number): Promise<Array<nsfwjs.predictionType>>

import {
  Controller,
  Post,
  UploadedFile,
  UseInterceptors,
} from '@nestjs/common';
import { FileInterceptor } from '@nestjs/platform-express';
import * as nsfwjs from 'nsfwjs';
import * as tf from '@tensorflow/tfjs-node';
import { readFileSync } from 'fs';
import jpeg from 'jpeg-js';

const convert = async (img) => {
  // Decoded image in UInt8 Byte array
  const image = await jpeg.decode(img, { useTArray: true });

  const numChannels = 3;
  const numPixels = image.width * image.height;
  const values = new Int32Array(numPixels * numChannels);

  for (let i = 0; i < numPixels; i++)
    for (let c = 0; c < numChannels; ++c)
      values[i * numChannels + c] = image.data[i * 4 + c];

  return tf.tensor3d(values, [image.height, image.width, numChannels], 'int32');
};

@Controller('image-check')
export class ImageCheckController {
  @Post('detect')
  @UseInterceptors(FileInterceptor('image'))
  async detectImage(@UploadedFile() file) {
    if (!file) {
      return { error: 'No image uploaded.' };
    }
    const model = await nsfwjs.load('InceptionV3');
    const buffer = readFileSync(file.path);
    const image = await convert(buffer);
    const predictions = await model.classify(image);
    return predictions;
  }
}

let isImageIllegal = false;
if (predictions[0].probability > 0.5 && predictions[0].className === 'Porn') {
  isImageIllegal = true;
}