.dockerignore

@@ -1 +0,0 @@
-node_modules

.env.example

@@ -1,6 +0,0 @@
-## DB [MongoDb](https://www.mongodb.com/) ##
-MONGO_USERNAME=akamai
-MONGO_PASSWORD=password
-MONGO_DATABASE=mydatabase
-MONGO_HOST=mongo
-MONGO_PORT=27017

.gitignore

@@ -4,7 +4,7 @@
 # Reference: https://intellij-support.jetbrains.com/hc/en-us/articles/206544839
 
 #poppeteer's output
-sites_assets/**/*
+crawler_assets/**/*
 
 
 # User-specific stuff:

Dockerfile

@@ -1,27 +0,0 @@
-FROM node:14-alpine AS base
-
-# Set non-interactive mode
-ENV DEBIAN_FRONTEND=noninteractive
-ENV NODE_ENV=production
-
-# Install chromium
-RUN apk add --no-cache chromium chromium-chromedriver
-
-# Create a symbolic link for google-chrome
-RUN ln -s /usr/bin/chromium-browser /usr/bin/google-chrome
-
-# Set the working directory
-WORKDIR /app
-
-# Copy the package.json and package-lock.json files
-COPY package*.json ./
-
-# Install dependencies
-RUN npm install -g npm@9.6.3 && \
-    npm install
-
-# Copy the source code
-COPY . .
-
-# Start the application
-CMD ["npm", "run", "start"]

README.md

@@ -1,34 +1 @@
-# Crawing & Serving #
-
-The crawler is a simple crawler that crawls the web and stores the results in a database and assets in a file system. The server is a simple server that serves the results of the crawler.
-
-
-## Crawler ##
-
-
-### Usage ###
-Post a JSON object to the crawler with the following format:
-
-  `domain.com/crawl`
-    {
-        "url": "http://www.example.com",
-    }
-
-The crawler will then crawl the given url and store the results in a database and assets in a file system
-`crawler_assests/www.example.com/`.
-
-
-# API # 
-
-The API is a simple API that serves the results of the crawler.
-
-# Routes # 
-  ## GET ## 
-    /sites - Returns a list of all sites
-    /sites/:id - Returns the site object for the given site Id
-    /sites/domain/:domain - Returns the domain object for the given domain
-  ## DELETE ##
-    /sites/:id - Deletes the site object for the given site Id
-    /sites/domain/:domain - Deletes the domain object for the given domain
-  ## Post ## 
-    /sites/:id - Updates the site object for the given site Id
+# Crawing && Serving #

docker-compose.yaml

@@ -1,32 +0,0 @@
-version: '3'
-services:
-  web:
-    build: .
-    ports:
-      - '3000:3000'
-    depends_on:
-      - mongo
-    environment:
-      - MONGO_USERNAME=${MONGO_USERNAME}
-      - MONGO_PASSWORD=${MONGO_PASSWORD}
-      - MONGO_HOST=${MONGO_HOST}
-      - MONGO_PORT=${MONGO_PORT}
-      - MONGO_DATABASE=${MONGO_DATABASE}
-    networks:
-      - appnet
-  mongo:
-    image: mongo
-    environment:
-      - MONGO_INITDB_DATABASE=${MONGO_DATABASE}
-      - MONGO_INITDB_ROOT_USERNAME=${MONGO_USERNAME}
-      - MONGO_INITDB_ROOT_PASSWORD=${MONGO_PASSWORD}
-    volumes:
-      - new:/data/db
-    networks:
-      - appnet
-networks:
-  appnet:
-    driver: bridge
-volumes:
-  new:
-    driver: local

package.json

@@ -21,18 +21,13 @@
   "dependencies": {
     "@nestjs/common": "^9.0.0",
     "@nestjs/core": "^9.0.5",
-    "@nestjs/mongoose": "^9.2.2",
     "@nestjs/platform-express": "^9.0.0",
-    "axios": "^1.3.5",
-    "dotenv": "^16.0.3",
-    "minimatch": "^9.0.0",
-    "mongoose": "^7.0.4",
     "puppeteer": "^19.9.1",
     "reflect-metadata": "^0.1.13",
-    "@nestjs/cli": "^9.0.0",
     "rxjs": "^7.5.5"
   },
   "devDependencies": {
+    "@nestjs/cli": "^9.0.0",
     "@nestjs/schematics": "^9.0.0",
     "@nestjs/testing": "^9.0.0",
     "@types/express": "^4.17.13",

src/api/api.controller.spec.ts

@@ -0,0 +1,18 @@
+import { Test, TestingModule } from '@nestjs/testing';
+import { ApiController } from './api.controller';
+
+describe('ApiController', () => {
+  let controller: ApiController;
+
+  beforeEach(async () => {
+    const module: TestingModule = await Test.createTestingModule({
+      controllers: [ApiController],
+    }).compile();
+
+    controller = module.get<ApiController>(ApiController);
+  });
+
+  it('should be defined', () => {
+    expect(controller).toBeDefined();
+  });
+});

src/api/api.controller.ts

@@ -1,68 +1,11 @@
-import { Body, Controller, Delete, Get, Param, Post } from '@nestjs/common';
+import { Body, Controller, Post } from '@nestjs/common';
 import { CrawlerService } from '../crawler/crawler.service';
-import { InjectModel } from '@nestjs/mongoose';
-import { Model } from 'mongoose';
-import { Site } from 'src/interfaces/site.interface';
 
 @Controller('/')
 export class ApiController {
-
-  constructor(private crawlerService: CrawlerService, @InjectModel('Site') private readonly siteModel: Model<Site> ) {}
-  
+  constructor(private crawlerService: CrawlerService) {}
   @Post('crawl')
   async crawl(@Body() body: { url: string }) {
-
-    const results = this.crawlerService.crawl(body.url);
-    results.then((data) => {
-      console.log("Done crawling !", data);
-      const newSite = new this.siteModel(data);
-      newSite.save().then((result) => {
-        console.log("Site saved !", result);
-      }).catch((err) => {
-        console.log("Error saving site !", err.message);
-      });
-    }).catch((err) => {
-      console.log("** Error crawling ! **", err);
-      console.log(err);
-    });
-    return {
-      message: 'Got your request for ' + body.url
+    return this.crawlerService.crawl(body.url);
   }
 }
-
-  // Get all
-  @Get('sites')
-  async getSites() {
-    const sites = await this.siteModel.find().exec();
-    return sites || {};
-  }
-
-  // Get by id
-  @Get('sites/:id')
-  async getSite(@Param('id') id: string) {
-    const site = await this.siteModel.findById(id).exec();
-    return site || {};
-  }
-
-  // Get by domain
-  @Get('sites/domain/:domain')
-  async getSiteByDomain(@Param('domain') domain: string) {
-    const site = await this.siteModel.findOne({ domain }).exec();
-    return site || {};
-  }
-
-  // Delete by domain
-  @Delete('sites/domain/:domain')
-  async deleteSiteByDomain(@Param('domain') domain: string) {
-    const site = await this.siteModel.findOneAndDelete({ domain }).exec();
-    return site || {};
-  }
-
-  // Delete by id
-  @Delete('sites/:id')
-  async deleteSite(@Param('id') id: string) {
-    const site = await this.siteModel.findByIdAndDelete(id).exec();
-    return site || {};
-  }
-  
-}

src/api/api.module.ts

@@ -1,11 +1,8 @@
 import { Module } from '@nestjs/common';
 import { ApiController } from './api.controller';
 import { CrawlerService } from '../crawler/crawler.service';
-import { SitesSchema } from '../schema/sites.schema';
-import { MongooseModule } from '@nestjs/mongoose';
 
 @Module({
-  imports: [MongooseModule.forFeature([{ name: 'Site', schema: SitesSchema }])],
   controllers: [ApiController],
   providers: [CrawlerService]
 

src/app.module.ts

@@ -1,20 +1,8 @@
 import { Module } from '@nestjs/common';
-import { MongooseModule } from '@nestjs/mongoose';
 import { ApiModule } from './api/api.module';
 import { CrawlerModule } from './crawler/crawler.module';
-const dotenv = require('dotenv');
 
-dotenv.config();
-console.log("mongodb://" + process.env.MONGO_USERNAME + ":" + process.env.MONGO_PASSWORD + "@" + process.env.MONGO_HOST + ":" + process.env.MONGO_PORT);
 @Module({
-  imports: [
-    MongooseModule.forRoot("mongodb://" + process.env.MONGO_USERNAME + ":" + process.env.MONGO_PASSWORD + "@" + process.env.MONGO_HOST + ":" + process.env.MONGO_PORT, {
-      useNewUrlParser: true,
-      useUnifiedTopology: true,
-    }),
-    ApiModule,
-    CrawlerModule
-  ]
-
+  imports: [ApiModule, CrawlerModule]
 })
 export class AppModule {}

src/crawler/crawler.service.spec.ts

@@ -0,0 +1,18 @@
+import { Test, TestingModule } from '@nestjs/testing';
+import { CrawlerService } from './crawler.service';
+
+describe('CrawlerService', () => {
+  let service: CrawlerService;
+
+  beforeEach(async () => {
+    const module: TestingModule = await Test.createTestingModule({
+      providers: [CrawlerService],
+    }).compile();
+
+    service = module.get<CrawlerService>(CrawlerService);
+  });
+
+  it('should be defined', () => {
+    expect(service).toBeDefined();
+  });
+});

src/crawler/crawler.service.ts

@@ -3,102 +3,25 @@ import { mkdirSync } from 'fs';
 import * as fs from 'fs';
 import puppeteer from 'puppeteer';
 import { URL } from 'url';
-import axios from 'axios';
-import { Site } from '../interfaces/site.interface';
 
 @Injectable()
 export class CrawlerService {
 
-  constructor() {
-    if (!fs.existsSync('sites_assets')) {
-      mkdirSync('sites_assets');
-    }
-  }
-
-  async crawl(url: string): Promise<Site> {
-
-    console.log("start crawl website", url);
-    const browser = await puppeteer.launch({ executablePath: '/usr/bin/chromium-browser', headless: true, args: ['--no-sandbox'] });
+  async crawl(url: string){
+    const browser = await puppeteer.launch();
     const page = await browser.newPage();
-    const domain = this.extractDomain(url);
     await page.goto(url);
-    const directory = `sites_assets/${domain}/`;
+    const directory = `crawler_assets/${this.extractDomain(url)}/`;
+    console.log(directory)
     if (!fs.existsSync(directory)) {
       mkdirSync(directory);
     }
-
-    // STYLESHEETS //
-    console.log("start stylesheets")
-    const stylesheetsUrls = await page.$$eval('link[rel="stylesheet"]', links => links.map(link => link.href));
-    let cssDir = `${directory}/css/`
-    const cssSheetsLocation = await this.downloadFiles(stylesheetsUrls, cssDir);
-    console.log(`cssSheetsLocation: `, cssSheetsLocation);
-    // STYLESHEETS //
-
-    // SCRIPTS //
-    const scriptsUrls = await page.$$eval('script', scripts => scripts.map(script => script.src));
-    let scriptsDir = `${directory}/scripts/`
-    const scriptsSheetsLocation = await this.downloadFiles(scriptsUrls, scriptsDir);
-    console.log(`scriptsSheetsLocation: `, scriptsSheetsLocation)
-    // SCRIPTS //
-
-    // SCREENSHOT //
-    const screenshotBuffer: Buffer = await page.screenshot({ fullPage: true });
-    await new Promise((resolve, reject) => {
-      fs.writeFile(`${directory}screenshot.png`, screenshotBuffer, (err) => {
-        if (err) {
-          reject(err);
-        } else {
-          resolve(true);
-        }
-      });
-    });
-    // SCREENSHOT //
-
-    // URLS //
-    const urls = await page.$$eval('a', links => links.map(link => link.href));
-    const urlsList = urls.filter((url) => url.startsWith('http'));
-    console.log(urlsList);
-    // URLS //
-
+    await page.pdf({path: `${directory}/page.pdf`, format: 'A4'});
+    await page.screenshot({path: `${directory}/screenshot.png`});
     await browser.close();
-    return {
-      domain,
-      cssSheetsLocation,
-      scriptsSheetsLocation,
-      urlsList
-    }
   }
 
-  async downloadFiles(urls: string[], path: string) {
-    const finalUrls: string[] = [];
-    if (!fs.existsSync(path)) {
-      mkdirSync(path);
-    }
-
-    await Promise.all(
-      urls.map(async (url) => {
-        if (!url.startsWith('http')) return;
-        const response = await axios.get(url);
-        const content = response.data;
-        // trim / from end of url string
-        let fileLocation = url.replace(/\/$/, "");
-        // get last part of url
-        fileLocation = fileLocation.substring(fileLocation.lastIndexOf('/') + 1);
-        // save file
-        if (fileLocation.length > 10) {
-          fileLocation = fileLocation.substring(0, 10);
-        }
-        finalUrls.push(`${path}${fileLocation}`);
-        console.log(`Saving file ${path}${fileLocation}`);
-        fs.writeFileSync(`${path}${fileLocation}`, content);
-      })
-    );
-
-    return finalUrls;
-  }
-
-  extractDomain(urlString: string) {
+  extractDomain(urlString) {
     const url = new URL(urlString);
     return url.hostname;
   }

src/interfaces/site.interface.ts

@@ -1,6 +0,0 @@
-export interface Site {
-  domain: string;
-  cssSheetsLocation: string[];
-  scriptsSheetsLocation: string[];
-  urlsList: string[];
-}

src/schema/sites.schema.ts

@@ -1,21 +0,0 @@
-import { Prop, Schema, SchemaFactory } from '@nestjs/mongoose';
-import { Document } from 'mongoose';
-
-export type SitesDocument = Sites & Document;
-
-@Schema()
-export class Sites {
-  @Prop({ required: true, unique: true})
-  domain: string;
-  
-  @Prop()
-  cssSheetsLocation: string[];
-
-  @Prop()
-  scriptsSheetsLocation: string[];
-
-  @Prop({ required: true})
-  urlsList: string[];
-}
-
-export const SitesSchema = SchemaFactory.createForClass(Sites);