From 4fce7e9287aed1078857f2a59e1cea449e6e97cf Mon Sep 17 00:00:00 2001
From: Jackson Harper <jacksonh@gmail.com>
Date: Mon, 22 Jan 2024 15:35:14 +0800
Subject: [PATCH 1/3] Add new job for handling upload content

---
 packages/pdf-handler/src/job.ts | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/packages/pdf-handler/src/job.ts b/packages/pdf-handler/src/job.ts
index 3ba1c1849..563c92506 100644
--- a/packages/pdf-handler/src/job.ts
+++ b/packages/pdf-handler/src/job.ts
@@ -4,6 +4,13 @@ import { redisDataSource } from './redis_data_source'
 const QUEUE_NAME = 'omnivore-backend-queue'
 const JOB_NAME = 'update-pdf-content'
 
+interface savePageJob {
+  userId: string
+  data: unknown
+  isRss: boolean
+  isImport: boolean
+}
+
 const queue = new Queue(QUEUE_NAME, {
   connection: redisDataSource.queueRedisClient,
 })

From f8264bfc10cd4f9f7e5b68d46cedcefd96548a69 Mon Sep 17 00:00:00 2001
From: Jackson Harper <jacksonh@gmail.com>
Date: Tue, 23 Jan 2024 23:45:28 +0800
Subject: [PATCH 2/3] Hard code tweet url to avoid import

---
 packages/api/src/services/save_page.ts | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/packages/api/src/services/save_page.ts b/packages/api/src/services/save_page.ts
index bb887fd83..a91f90549 100644
--- a/packages/api/src/services/save_page.ts
+++ b/packages/api/src/services/save_page.ts
@@ -18,7 +18,6 @@ import {
   cleanUrl,
   generateSlug,
   stringToHash,
-  TWEET_URL_REGEX,
   validatedDate,
   wordsCount,
 } from '../utils/helpers'
@@ -32,7 +31,7 @@ import { createLibraryItem, updateLibraryItem } from './library_item'
 
 // where we can use APIs to fetch their underlying content.
 const FORCE_PUPPETEER_URLS = [
-  TWEET_URL_REGEX,
+  /twitter\.com\/(?:#!\/)?(\w+)\/status(?:es)?\/(\d+)(?:\/.*)?/,
   /^((?:https?:)?\/\/)?((?:www|m)\.)?((?:youtube\.com|youtu.be))(\/(?:[\w-]+\?v=|embed\/|v\/)?)([\w-]+)(\S+)?$/,
 ]
 const ALREADY_PARSED_SOURCES = [
@@ -58,7 +57,10 @@ const createSlug = (url: string, title?: string | null | undefined) => {
 const shouldParseInBackend = (input: SavePageInput): boolean => {
   return (
     ALREADY_PARSED_SOURCES.indexOf(input.source) === -1 &&
-    FORCE_PUPPETEER_URLS.some((regex) => regex.test(input.url))
+    FORCE_PUPPETEER_URLS.some((regex) => {
+      console.log('REGEX: ', regex)
+      return regex.test(input.url)
+    })
   )
 }
 

From c7cfb8693f95ec7875b9d64eb776e5458d722afa Mon Sep 17 00:00:00 2001
From: Jackson Harper <jacksonh@gmail.com>
Date: Tue, 23 Jan 2024 23:46:59 +0800
Subject: [PATCH 3/3] Remove unused change

---
 packages/pdf-handler/src/job.ts | 7 -------
 1 file changed, 7 deletions(-)

diff --git a/packages/pdf-handler/src/job.ts b/packages/pdf-handler/src/job.ts
index 563c92506..3ba1c1849 100644
--- a/packages/pdf-handler/src/job.ts
+++ b/packages/pdf-handler/src/job.ts
@@ -4,13 +4,6 @@ import { redisDataSource } from './redis_data_source'
 const QUEUE_NAME = 'omnivore-backend-queue'
 const JOB_NAME = 'update-pdf-content'
 
-interface savePageJob {
-  userId: string
-  data: unknown
-  isRss: boolean
-  isImport: boolean
-}
-
 const queue = new Queue(QUEUE_NAME, {
   connection: redisDataSource.queueRedisClient,
 })