Improve youtube functions

.github/workflows/docker-build.yaml

@@ -4,12 +4,12 @@ name: Build and Push Docker Image
 on:
   push:
     branches:
-      - 'master'
+      - "master"
     tags:
-      - 'v*'
+      - "v*"
   pull_request:
     branches:
-      - 'master'
+      - "master"
 
 jobs:
   build:

Dockerfile

@@ -1,6 +1,6 @@
 # use the official Bun image
 # see all versions at https://hub.docker.com/r/oven/bun/tags
-FROM oven/bun:1.1.4-alpine as base
+FROM oven/bun:1.1.4-alpine AS base
 
 # install dependencies into temp directory
 # this will cache them and speed up future builds
@@ -12,7 +12,7 @@ RUN cd /temp/prod && bun install --frozen-lockfile --production
 
 
 # final image
-FROM base as release
+FROM base AS release
 
 RUN apk add -q --progress --update --no-cache dumb-init
 

src/index.ts

@@ -1,6 +1,11 @@
 import { summaryYouTubeTranscript } from "./langchain-openai.ts";
 import { logger, raindropAPI, Config } from "./config.ts";
-import { getTranscript, getVideoDetails } from "./youtube.ts";
+import {
+  normalizeYouTubeURL,
+  getTranscript,
+  getVideoDetails,
+  extractVideoId,
+} from "./youtube.ts";
 
 async function main() {
   // Get link
@@ -15,37 +20,53 @@ async function main() {
   // Remove the link from Raindrop
   await raindropAPI.removeItem(object._id.toString());
 
+  // Normalize link
+  const link = await normalizeYouTubeURL(object.link);
+  if (link === null) {
+    logger.error(`Invalid YouTube URL: ${object.link}`);
+    return;
+  }
+
+  // Extract video ID
+  const videoId = await extractVideoId(link);
+  if (videoId === null) {
+    logger.error(`Invalid YouTube URL: ${link}`);
+    return;
+  }
+
   // Get details from YouTube
-  const [title, channel, duration, upload_date] = await getVideoDetails(
-    object.link,
-  );
+  const [title, channel, duration, upload_date] =
+    await getVideoDetails(videoId);
+
   // Get transcript from YouTube
-  const transcript = await getTranscript(object.link);
+  const transcript = await getTranscript(videoId);
 
   // Summarize the transcript if available
   let summary = "Podsumowanie: none";
   let usage = "";
-  if (transcript !== "") {
+  if (transcript === null) {
+    logger.error(`Transcription unavailable for: ${link}`);
+  } else {
     [summary, usage] = await summaryYouTubeTranscript(transcript);
   }
 
   // Send the summary to target collection in Raindrop
   const note = `<${channel}> ${title} [${duration}]
 ---
-${object.link} (${upload_date})
+${link} (${upload_date})
 ${summary}
 ${usage}
 `;
   const result = await raindropAPI.addItem(
     Config.raindrop_target_collection_id,
-    object.link,
+    link,
     note,
   );
 
   if (result) {
-    logger.info(`Final success: ${object.link} ${title}`);
+    logger.info(`Final success: ${link} ${title}`);
   } else {
-    logger.error(`Final failure: ${object.link} ${title}`);
+    logger.error(`Final failure: ${link} ${title}`);
   }
 }
 

src/youtube.ts

@@ -1,22 +1,25 @@
 import { Innertube } from "youtubei.js/web";
-import { logger } from "./config.ts";
 
-function extractVideoId(videoUrl: string): string | null {
-  const urlParts = videoUrl.split('v=');
+export async function extractVideoId(videoUrl: string): Promise<string | null> {
+  const urlParts = videoUrl.split("v=");
   if (urlParts.length > 1) {
-      const videoId = urlParts[1].split('&')[0];
-      return videoId;
+    const videoId = urlParts[1].split("&")[0];
+    return videoId;
   }
   return null;
 }
 
-export async function getTranscript(videoUrl: string): Promise<string> {
-  const videoId = extractVideoId(videoUrl);
+export async function normalizeYouTubeURL(
+  videoUrl: string,
+): Promise<string | null> {
+  const videoId = await extractVideoId(videoUrl);
   if (videoId === null) {
-    logger.error(`Invalid YouTube URL: ${videoUrl}`);
-    return "";
+    return null;
   }
+  return `https://www.youtube.com/watch?v=${videoId}`;
+}
 
+export async function getTranscript(videoId: string): Promise<string | null> {
   const youtube = await Innertube.create();
 
   try {
@@ -31,18 +34,13 @@ export async function getTranscript(videoUrl: string): Promise<string> {
         .join(" ") ?? ""
     );
   } catch (error) {
-    logger.error(`Transcription unavailable for: ${videoUrl}`);
-    return "";
+    return null;
   }
 }
 
-export async function getVideoDetails(videoUrl: string): Promise<[string | undefined, string | undefined, string, string]> {
-  const videoId = extractVideoId(videoUrl);
-  if (videoId === null) {
-    logger.error(`Invalid YouTube URL: ${videoUrl}`);
-    return ["", "", "", ""];
-  }
-
+export async function getVideoDetails(
+  videoId: string,
+): Promise<[string | undefined, string | undefined, string, string]> {
   const youtube = await Innertube.create();
 
   const videoBasicInfo = await youtube.getBasicInfo(videoId);
@@ -98,5 +96,5 @@ if (import.meta.main) {
       Run tests
   */
   // await test_getTranscript();
-  await test_getVideoDetails();
+  // await test_getVideoDetails();
 }