amechama/sharkey
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

merge: follow-up to !390 (!397)

Browse Source 
View MR for information: https://activitypub.software/TransFem-org/Sharkey/-/merge_requests/397

Approved-by: Amelia Yukii <amelia.yukii@shourai.de>
Approved-by: Marie <marie@kaifa.ch>
This commit is contained in:
Marie 2024-02-03 17:05:21 +00:00
commit 65a7623944
1 changed files with 28 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

54
packages/backend/src/queue/processors/ImportNotesProcessorService.ts
View File

@ -1,4 +1,5 @@
import * as fs from 'node:fs';
import * as fsp from 'node:fs/promises';
import * as vm from 'node:vm';
import * as crypto from 'node:crypto';
import { Inject, Injectable } from '@nestjs/common';
@ -51,7 +52,7 @@ export class ImportNotesProcessorService {
@bindThis
private async uploadFiles(dir: string, user: MiUser, folder?: MiDriveFolder['id']) {
const fileList = fs.readdirSync(dir);
const fileList = await fsp.readdir(dir);
for await (const file of fileList) {
const name = `${dir}/${file}`;
if (fs.statSync(name).isDirectory()) {
@ -130,14 +131,16 @@ export class ImportNotesProcessorService {
return typeof obj[Symbol.iterator] === 'function';
}
private parseTwitterFile(str : string) : null | [{ tweet: any }] {
const removed = str.replace(new RegExp('window\\.YTD\\.tweets\\.part0 = ', 'g'), '');
@bindThis
private parseTwitterFile(str : string) : { tweet: object }[] {
const jsonStr = str.replace(/^\s*window\.YTD\.tweets\.part0\s*=\s*/, '');
try {
return JSON.parse(removed);
return JSON.parse(jsonStr);
} catch (error) {
//The format is not what we expected. Either this file was tampered with or twitters exports changed
return null;
this.logger.warn('Failed to import twitter notes due to malformed file');
throw error;
}
}
@ -173,7 +176,7 @@ export class ImportNotesProcessorService {
const destPath = path + '/twitter.zip';
try {
fs.writeFileSync(destPath, '', 'binary');
await fsp.writeFile(destPath, '', 'binary');
await this.downloadService.downloadUrl(file.url, destPath);
} catch (e) { // TODO: 何度か再試行
if (e instanceof Error || typeof e === 'string') {
@ -185,21 +188,13 @@ export class ImportNotesProcessorService {
const outputPath = path + '/twitter';
try {
this.logger.succ(`Unzipping to ${outputPath}`);
ZipReader.withDestinationPath(outputPath).viaBuffer(await fs.promises.readFile(destPath));
ZipReader.withDestinationPath(outputPath).viaBuffer(await fsp.readFile(destPath));
const unprocessedTweetJson = this.parseTwitterFile(fs.readFileSync(outputPath + '/data/tweets.js', 'utf-8'));
//Make sure that it isnt null (because if something went wrong in parseTwitterFile it returns null)
if (unprocessedTweetJson) {
const tweets = Object.keys(unprocessedTweetJson).reduce((m, key, i, obj) => {
return m.concat(unprocessedTweetJson[i].tweet);
}, []);
const unprocessedTweets = this.parseTwitterFile(await fsp.readFile(outputPath + '/data/tweets.js', 'utf-8'));
const tweets = unprocessedTweets.map(e => e.tweet);
const processedTweets = await this.recreateChain(['id_str'], ['in_reply_to_status_id_str'], tweets, false);
this.queueService.createImportTweetsToDbJob(job.data.user, processedTweets, null);
} else {
this.logger.warn('Failed to import twitter notes due to malformed file');
}
} finally {
cleanup();
}
@ -211,7 +206,7 @@ export class ImportNotesProcessorService {
const destPath = path + '/facebook.zip';
try {
fs.writeFileSync(destPath, '', 'binary');
await fsp.writeFile(destPath, '', 'binary');
await this.downloadService.downloadUrl(file.url, destPath);
} catch (e) { // TODO: 何度か再試行
if (e instanceof Error || typeof e === 'string') {
@ -223,8 +218,8 @@ export class ImportNotesProcessorService {
const outputPath = path + '/facebook';
try {
this.logger.succ(`Unzipping to ${outputPath}`);
ZipReader.withDestinationPath(outputPath).viaBuffer(await fs.promises.readFile(destPath));
const postsJson = fs.readFileSync(outputPath + '/your_activity_across_facebook/posts/your_posts__check_ins__photos_and_videos_1.json', 'utf-8');
ZipReader.withDestinationPath(outputPath).viaBuffer(await fsp.readFile(destPath));
const postsJson = await fsp.readFile(outputPath + '/your_activity_across_facebook/posts/your_posts__check_ins__photos_and_videos_1.json', 'utf-8');
const posts = JSON.parse(postsJson);
const facebookFolder = await this.driveFoldersRepository.findOneBy({ name: 'Facebook', userId: job.data.user.id, parentId: folder?.id });
if (facebookFolder == null && folder) {
@ -244,7 +239,7 @@ export class ImportNotesProcessorService {
const destPath = path + '/unknown.zip';
try {
fs.writeFileSync(destPath, '', 'binary');
await fsp.writeFile(destPath, '', 'binary');
await this.downloadService.downloadUrl(file.url, destPath);
} catch (e) { // TODO: 何度か再試行
if (e instanceof Error || typeof e === 'string') {
@ -256,11 +251,11 @@ export class ImportNotesProcessorService {
const outputPath = path + '/unknown';
try {
this.logger.succ(`Unzipping to ${outputPath}`);
ZipReader.withDestinationPath(outputPath).viaBuffer(await fs.promises.readFile(destPath));
ZipReader.withDestinationPath(outputPath).viaBuffer(await fsp.readFile(destPath));
const isInstagram = type === 'Instagram' || fs.existsSync(outputPath + '/instagram_live') || fs.existsSync(outputPath + '/instagram_ads_and_businesses');
const isOutbox = type === 'Mastodon' || fs.existsSync(outputPath + '/outbox.json');
if (isInstagram) {
const postsJson = fs.readFileSync(outputPath + '/content/posts_1.json', 'utf-8');
const postsJson = await fsp.readFile(outputPath + '/content/posts_1.json', 'utf-8');
const posts = JSON.parse(postsJson);
const igFolder = await this.driveFoldersRepository.findOneBy({ name: 'Instagram', userId: job.data.user.id, parentId: folder?.id });
if (igFolder == null && folder) {
@ -270,16 +265,16 @@ export class ImportNotesProcessorService {
}
this.queueService.createImportIGToDbJob(job.data.user, posts);
} else if (isOutbox) {
const actorJson = fs.readFileSync(outputPath + '/actor.json', 'utf-8');
const actorJson = await fsp.readFile(outputPath + '/actor.json', 'utf-8');
const actor = JSON.parse(actorJson);
const isPleroma = actor['@context'].some((v: any) => typeof v === 'string' && v.match(/litepub(.*)/));
if (isPleroma) {
const outboxJson = fs.readFileSync(outputPath + '/outbox.json', 'utf-8');
const outboxJson = await fsp.readFile(outputPath + '/outbox.json', 'utf-8');
const outbox = JSON.parse(outboxJson);
const processedToots = await this.recreateChain(['object', 'id'], ['object', 'inReplyTo'], outbox.orderedItems.filter((x: any) => x.type === 'Create' && x.object.type === 'Note'), true);
this.queueService.createImportPleroToDbJob(job.data.user, processedToots, null);
} else {
const outboxJson = fs.readFileSync(outputPath + '/outbox.json', 'utf-8');
const outboxJson = await fsp.readFile(outputPath + '/outbox.json', 'utf-8');
const outbox = JSON.parse(outboxJson);
let mastoFolder = await this.driveFoldersRepository.findOneBy({ name: 'Mastodon', userId: job.data.user.id, parentId: folder?.id });
if (mastoFolder == null && folder) {
@ -302,7 +297,7 @@ export class ImportNotesProcessorService {
this.logger.info(`Temp dir is ${path}`);
try {
fs.writeFileSync(path, '', 'utf-8');
await fsp.writeFile(path, '', 'utf-8');
await this.downloadService.downloadUrl(file.url, path);
} catch (e) { // TODO: 何度か再試行
if (e instanceof Error || typeof e === 'string') {
@ -311,7 +306,7 @@ export class ImportNotesProcessorService {
throw e;
}
const notesJson = fs.readFileSync(path, 'utf-8');
const notesJson = await fsp.readFile(path, 'utf-8');
const notes = JSON.parse(notesJson);
const processedNotes = await this.recreateChain(['id'], ['replyId'], notes, false);
this.queueService.createImportKeyNotesToDbJob(job.data.user, processedNotes, null);
@ -590,7 +585,8 @@ export class ImportNotesProcessorService {
try {
const date = new Date(tweet.created_at);
const textReplaceURLs = tweet.entities.urls && tweet.entities.urls.length > 0 ? await replaceTwitterUrls(tweet.full_text, tweet.entities.urls) : tweet.full_text;
const decodedText = tweet.full_text.replaceAll('&gt;', '>').replaceAll('&lt;', '<').replaceAll('&amp;', '&');
const textReplaceURLs = tweet.entities.urls && tweet.entities.urls.length > 0 ? await replaceTwitterUrls(decodedText, tweet.entities.urls) : decodedText;
const text = tweet.entities.user_mentions && tweet.entities.user_mentions.length > 0 ? await replaceTwitterMentions(textReplaceURLs, tweet.entities.user_mentions) : textReplaceURLs;
const files: MiDriveFile[] = [];