关于后端:爬了10000张NASA关于火星探索的图片我发现了一个秘密

5次阅读

共计 6274 个字符，预计需要花费 16 分钟才能阅读完成。

最近，我应用爬虫技术，爬取了美国航空航天局，也就是你电影里常常见到的 NASA, 火星摸索的相干图片，有 10000 张吧。

嗯嗯，小事件，小事件。

完事儿之后，有点小冲动，于是就有了这篇文章，将有以下内容：

我为什么要爬取 NASA 的图片
我是如何爬取 NASA 图片的（超具体）
我失去了什么(高清大图)
我发现了什么机密（超劲爆）

我曾经超过 35 了，瑟瑟发抖啥时候被开了。

天天想着万一哪天就业了干点啥，想着玩个自媒体吧，天天给大家瞎文言。文言点历史谜团，宇宙神秘啥的，于是我就盯上了 NASA。

NASA 有各种宇宙摸索工作，并且有相干的文章，访谈，图片，视频公开，这是不可多得的资源库。

NASA 的网站是能够公开拜访的，地址是

https://www.nasa.gov/

关上当前，它首页长这样，能够看到各种内容。右上角还有个搜寻框，咱们输出 Mars 也就是 火星

稍等片刻，会展现出与 Mars 相干的各种内容，其中有一项 Mars Exploration 也就是 火星摸索

点开之后，就到了一个新页面，而后找到 Images 图片，就到了咱们爬取的指标页

https://www.nasa.gov/mission_pages/mars/images/index.html

页面下拉，你将会看到有个大大的按钮，下面写着 MORE IMAGES，点击试试你就发现：

页面的内容，并不是页面间接加载的，而是通过 api 申请后，异步渲染的

F12, 关上浏览器开发者模式, 从新执行方才的步骤，察看申请信息，发现都会有如下的状况

看上去这个 url 地址十分重要，咱们先看他的申请地址：

https://www.nasa.gov/api/2/ubernode/_search?size=24&from=24&sort=promo-date-time%3Adesc&q=((ubernode-type%3Aimage)%20AND%20(topics%3A3152))&_source_include=promo-date-time%2Cmaster-image%2Cnid%2Ctitle%2Ctopics%2Cmissions%2Ccollections%2Cother-tags%2Cubernode-type%2Cprimary-tag%2Csecondary-tag%2Ccardfeed-title%2Ctype%2Ccollection-asset-link%2Clink-or-attachment%2Cpr-leader-sentence%2Cimage-feature-caption%2Cattachments%2Curi

留神看外面的参数

size=24&from=24

很显然，size 就是每次申请图片的数量，from 通过试验，是查问初始地位，咱们能够改它来获取其余内容

咱们再来看看它的返回信息：

{
    "took": 3,
    "timed_out": false,
    "_shards": {
        "total": 5,
        "successful": 5,
        "skipped": 0,
        "failed": 0
    },
    "hits": {
        "total": 659,
        "max_score": null,
        "hits": [{
            "_index": "nasa-public",
            "_type": "ubernode",
            "_id": "450040",
            "_score": null,
            "_source": {
                "image-feature-caption": "Mars 2020 rover underwent an eye exam after several cameras were installed on the rover.",
                "topics": ["3140", "3152"],
                "nid": "450040",
                "title": "NASA'Optometrists'Verify Mars 2020 Rover's 20/20 Vision","type":"ubernode","uri":"/image-feature/jpl/nasa-optometrists-verify-mars-2020-rovers-2020-vision","collections": ["4525","5246"],"link-or-attachment":"link","missions": ["6336"],"primary-tag":"6336","cardfeed-title":"NASA 'Optometrists' Verify Mars 2020 Rover's 20/20 Vision",
                "promo-date-time": "2019-08-05T17:49:00-04:00",
                "secondary-tag": "3140",
                "master-image": {
                    "fid": "603128",
                    "alt": "Engineers test cameras on the top of the Mars 2020 rover’s mast and front chassis.",
                    "width": "1600",
                    "id": "603128",
                    "title": "Engineers test cameras on the top of the Mars 2020 rover’s mast and front chassis.",
                    "uri": "public://thumbnails/image/pia23314-16.jpg",
                    "height": "900"
                },
                "ubernode-type": "image"
            },
            "sort": [1565041740000]
        }, {
            "_index": "nasa-public",
            "_type": "ubernode",
            "_id": "433172",
            "_score": null,
            "_source": {"image-feature-caption": "NASA still hasn't heard from the Opportunity rover, but at least we can see it again.","topics": ["3152"],"nid":"433172","title":"Opportunity Emerges in a Dusty Picture","type":"ubernode","uri":"/image-feature/opportunity-emerges-in-a-dusty-picture","collections": ["7628"],"link-or-attachment":"link","missions": ["3639"],"primary-tag":"3152","cardfeed-title":"Opportunity Emerges in a Dusty Picture","promo-date-time":"2018-09-26T12:39:00-04:00","secondary-tag":"7628","master-image": {"fid":"584263","alt":"NASA's Opportunity rover appears as a blip in the center of this square",
                    "width": "1400",
                    "id": "584263",
                    "title": "NASA's Opportunity rover appears as a blip in the center of this square","uri":"public://thumbnails/image/pia22549-16.jpg","height":"788"},"ubernode-type":"image"},"sort": [1537979940000]
        }]
    }
}

下面的 json 内容太长，我删除了一些反复的，实际上 hits 这个数组，也是 24 个，和页面上显示的图片数量是一样的。那根本能够判定，页面上的信息，就是从这个数组来的。

进而比照发现，master-image字段下，就是咱们须要的信息，包含 图片地址 , 图片尺寸 , 图片题目。

上面就上代码，组装申请 URL, 获取内容，下载图片三步走

我应用了 Dart 语言，你们随便

import 'dart:convert';
import 'package:dio/dio.dart';

main() async {
  // 每页数量是固定 24 个, 扭转初始值即可
  for (int from = 0; from < 24 * 100; from = from + 24) {await getPage(from);
  }
}

// 获取每一页的信息并且下载
Future<void> getPage(int from) async {
  String url = 'https://www.nasa.gov/api/2/ubernode/_search?size=24&from=' +
      from.toString() +
      '&sort=promo-date-time%3Adesc&q=((ubernode-type%3Aimage)%20AND%20(topics%3A3152))&_source_include=promo-date-time%2Cmaster-image%2Cnid%2Ctitle%2Ctopics%2Cmissions%2Ccollections%2Cother-tags%2Cubernode-type%2Cprimary-tag%2Csecondary-tag%2Ccardfeed-title%2Ctype%2Ccollection-asset-link%2Clink-or-attachment%2Cpr-leader-sentence%2Cimage-feature-caption%2Cattachments%2Curi';

  // 获取到内容
  var res = await Dio().get(url);
  var map = jsonDecode(res.toString());

  (map['hits']['hits'] as List<dynamic>).forEach((element) async {Uri fileUri = Uri.parse(getUri(element));
    String savePath = getSavePath(element);

    await Dio().downloadUri(fileUri, savePath);
    print('已下载:' + savePath);
  });
}

// 获取图片下载地址
String getUri(dynamic element) {String uri = element['_source']['master-image']['uri'].toString();

  uri = uri.replaceAll('public://',
      'https://www.nasa.gov/sites/default/files/styles/full_width_feature/public/');

  return uri;
}

// 解决信息，并且返回图片保留地址
String getSavePath(dynamic element) {String id = element['_id'];
  String fid = element['_source']['master-image']['fid'].toString();
  String title = element['_source']['master-image']['title'].toString();
  String uri = element['_source']['master-image']['uri'].toString();

  String savePath =
      id + '_' + fid + '_' + title.trim() + '.' + uri.split('.').last;
  savePath = savePath.replaceAll('/', '');
  savePath = savePath.replaceAll('\\', '');
  savePath = savePath.replaceAll('"','');
  savePath = 'images/' + savePath;

  return savePath;
}

下面的代码，还是很简略的，有教训的同学应该一看就懂。

走起来吧。

已下载: images/470436_643588_This is the third color image taken by NASA’s Ingenuity helicopter.jpg
已下载: images/470435_643587_This is the second color image taken by NASA’s Ingenuity helicopter.jpg
已下载: images/468546_639327_This is the first high-resolution, color image to be sent back by the Hazard Cameras (Hazcams).jpg
已下载: images/452007_605784_Danielson Crater on Mars.jpg
已下载: images/458478_615132_Gullies on Mars.jpg
已下载: images/469416_641582_A field of sand dunes occupies this frosty 5-kilometer diameter crater in the high-latitudes of the northern plains of Mars..jpeg
已下载: images/458075_614251_Mars 2020 With Sample Tubes (Artist's Concept).jpg
已下载: images/470381_643473_CME.jpg
已下载: images/458813_615896_Mars.jpg
已下载: images/467026_635309_Illustration of NASA’s Perseverance rover begins its descent through the Martian atmosphere.jpg
已下载: images/470438_643591_This black and white image was taken by NASA’s Ingenuity helicopter during its third flight on April 25, 2021.jpg
已下载: images/465488_631398_Cliffs in Ancient Ice on Mars.jpg
已下载: images/463659_626874_Avalanche on Mars.jpg
已下载: images/470251_643164_This image from NASA’s Perseverance rover shows the agency’s Ingenuity Mars Helicopter right after it successfully completed a high-speed spin-up test..jpeg
已下载: images/468636_639726_Mars' Jezero Crater.jpg

这些图片

还有这些

图片，图片题目都有了, 够看一个月了我预计。

这张图，我最喜爱了。一个如此清晰，一个如此浑浊，这是为什么？火星人的裂缝产生器？

好吧，真正的机密是：

NASA 的网站居然没有防采集，不信你也试试。。。

正文完

后端

发表至：后端

2021-06-30

0

关于后端:北京某大公司SpringBean生命周期

关于后端:我们正在错误的组织代码

关于后端:函数式编程的Java编码实践利用惰性写出高性能且抽象的代码

关于云原生:深度解读畅捷通云原生架构转型实战历程

关于后端:爬了10000张NASA关于火星探索的图片我发现了一个秘密

前言

我为什么要爬 NASA 的图片

我是如何爬取 NASA 图片的（超具体）

我失去了什么

我发现了什么机密

站内搜索