fix(route): picnob (DIYgod#13986)

* fix(route): picnob

* fix(route): picnob. Use one browser session to do all http requests.

* fix(route): picnob. Use puppeteer as a fallback option when a normal request returns a 403 error.

* fix(route): picnob. Block unnecessary requests when using puppeteer.

* fix(route): picnob. Adaptation of JSON responses when using puppeteer for http requests.

* Update lib/v2/picnob/user.js

---------

lib/v2/picnob/maintainer.js

@@ -1,3 +1,3 @@
 module.exports = {
-    '/user/:id': ['TonyRL'],
+    '/user/:id': ['TonyRL', 'micheal-death'],
 };

lib/v2/picnob/templates/desc.art

@@ -3,11 +3,11 @@
         <source src="{{ item.video }}" type="video/mp4">
     </video>
 {{ else if item.type === 'img_multi' }}
-    {{ each images i }}
-        <a href="{{ i.ori }}"><img src="{{ i.url }}"></a>
+    {{ each item.images i }}
+        <img src="{{ i.url }}">
     {{ /each }}
 {{ else if item.type === 'img_sig' }}
-    <a href="{{ item.pic }}"><img src="{{ item.pic }}"></a>
+    <img src="{{ item.pic }}">
 {{ /if }}
 <br>
 {{@ item.sum }}

lib/v2/picnob/user.js

@@ -3,53 +3,90 @@ const cheerio = require('cheerio');
 const { parseDate } = require('@/utils/parse-date');
 const { art } = require('@/utils/render');
 const path = require('path');
+const { puppeteerGet } = require('./utils');
 
 module.exports = async (ctx) => {
     const baseUrl = 'https://www.picnob.com';
     const { id } = ctx.params;
     const url = `${baseUrl}/profile/${id}/`;
-    const { data: response } = await got(url);
-    const $ = cheerio.load(response);
 
+    const browser = await require('@/utils/puppeteer')();
+    // TODO: can't bypass cloudflare 403 error without puppeteer
+    let html;
+    let usePuppeteer = false;
+    try {
+        const { data } = await got(url, {
+            headers: {
+                accept: 'text/html',
+                referer: 'https://www.google.com/',
+            },
+        });
+        html = data;
+    } catch (e) {
+        if (e.message.includes('code 403')) {
+            html = await puppeteerGet(url, browser);
+            usePuppeteer = true;
+        }
+    }
+    const $ = cheerio.load(html);
     const profileName = $('h1.fullname').text();
     const userId = $('input[name=userid]').attr('value');
 
-    const { data } = await got(`${baseUrl}/api/posts`, {
-        searchParams: {
-            userid: userId,
-        },
-    });
+    let posts;
+    if (!usePuppeteer) {
+        const { data } = await got(`${baseUrl}/api/posts`, {
+            headers: {
+                accept: 'application/json',
+            },
+            searchParams: {
+                userid: userId,
+            },
+        });
+        posts = data.posts;
+    } else {
+        const data = await puppeteerGet(`${baseUrl}/api/posts?userid=${userId}`, browser);
+        posts = data.posts;
+    }
 
-    const list = data.posts.items.map(async (item) => {
-        const { shortcode, type } = item;
-        const link = `${baseUrl}/post/${shortcode}/`;
-        let images = [];
-        if (type === 'img_multi') {
-            images = await ctx.cache.tryGet(link, async () => {
-                const { data } = await got(link);
-                const $ = cheerio.load(data);
-                return [
-                    ...new Set(
-                        $('.post_slide a')
-                            .toArray()
-                            .map((a) => {
-                                a = $(a);
-                                return {
-                                    ori: a.attr('href'),
-                                    url: a.find('img').attr('data-src'),
-                                };
-                            })
-                    ),
-                ];
-            });
-        }
-        return {
-            title: item.sum_pure,
-            description: art(path.join(__dirname, 'templates/desc.art'), { item, images }),
-            link,
-            pubDate: parseDate(item.time, 'X'),
-        };
-    });
+    const list = await Promise.all(
+        posts.items.map(async (item) => {
+            const { shortcode, type } = item;
+            const link = `${baseUrl}/post/${shortcode}/`;
+            if (type === 'img_multi') {
+                item.images = await ctx.cache.tryGet(link, async () => {
+                    let html;
+                    if (!usePuppeteer) {
+                        const { data } = await got(link);
+                        html = data;
+                    } else {
+                        html = await puppeteerGet(link, browser);
+                    }
+                    const $ = cheerio.load(html);
+                    return [
+                        ...new Set(
+                            $('.post_slide a')
+                                .toArray()
+                                .map((a) => {
+                                    a = $(a);
+                                    return {
+                                        ori: a.attr('href'),
+                                        url: a.find('img').attr('data-src'),
+                                    };
+                                })
+                        ),
+                    ];
+                });
+            }
+
+            return {
+                title: item.sum_pure,
+                description: art(path.join(__dirname, 'templates/desc.art'), { item }),
+                link,
+                pubDate: parseDate(item.time, 'X'),
+            };
+        })
+    );
+    await browser.close();
 
     ctx.state.data = {
         title: `${profileName} (@${id}) - Picnob`,

lib/v2/picnob/utils.js

@@ -0,0 +1,24 @@
+const puppeteerGet = async (url, browser) => {
+    let data;
+    const page = await browser.newPage();
+    await page.setRequestInterception(true);
+    page.on('request', (request) => {
+        request.resourceType() === 'document' ? request.continue() : request.abort();
+    });
+    page.on('response', async (response) => {
+        if (response.request().url().includes('/api/posts')) {
+            data = await response.json();
+        } else {
+            data = await response.text();
+        }
+    });
+    await page.goto(url, {
+        waitUntil: 'domcontentloaded',
+    });
+    await page.close();
+    return data;
+};
+
+module.exports = {
+    puppeteerGet,
+};