fix(baoyu-fetch): extract X video variants

This commit is contained in:
Jim Liu 宝玉
2026-04-21 13:39:01 -05:00
parent d7f623158c
commit fe26c2b639
4 changed files with 261 additions and 22 deletions
@@ -339,4 +339,105 @@ describe("x article extraction", () => {
expect(content.markdown).toContain("https://example.com/report");
expect(content.markdown).not.toContain("https://t.co/example");
});
test("renders article video media as the highest bitrate mp4 link", () => {
const payload = {
data: {
tweetResult: {
result: {
rest_id: "2046628728210350366",
legacy: {
full_text: "Fallback text",
favorite_count: 12,
retweet_count: 3,
reply_count: 1,
created_at: "Tue Apr 21 16:34:47 +0000 2026",
},
core: {
user_results: {
result: {
legacy: {
name: "Google AI Studio",
screen_name: "GoogleAIStudio",
},
},
},
},
article: {
article_results: {
result: {
title: "Article with video",
media_entities: [
{
media_id: "2046627051822530560",
media_info: {
__typename: "ApiVideo",
variants: [
{
bit_rate: 2176000,
content_type: "video/mp4",
url: "https://video.twimg.com/amplify_video/2046627051822530560/vid/avc1/1280x720/medium.mp4",
},
{
content_type: "application/x-mpegURL",
url: "https://video.twimg.com/amplify_video/2046627051822530560/pl/playlist.m3u8",
},
{
bit_rate: 10368000,
content_type: "video/mp4",
url: "https://video.twimg.com/amplify_video/2046627051822530560/vid/avc1/1920x1080/high.mp4",
},
],
},
},
],
content_state: {
blocks: [
{
type: "atomic",
text: " ",
data: {},
entityRanges: [{ key: 0, length: 1, offset: 0 }],
inlineStyleRanges: [],
},
],
entityMap: [
{
key: "0",
value: {
type: "MEDIA",
mutability: "Immutable",
data: {
mediaItems: [{ mediaId: "2046627051822530560" }],
},
},
},
],
},
},
},
},
},
},
},
};
const document = extractArticleDocumentFromPayload(
payload,
"2046628728210350366",
"https://x.com/GoogleAIStudio/status/2046628728210350366",
);
expect(document).not.toBeNull();
const content = document?.content[0];
expect(content?.type).toBe("markdown");
if (!content || content.type !== "markdown") {
throw new Error("Expected markdown content");
}
expect(content.markdown).toBe(
"[video](https://video.twimg.com/amplify_video/2046627051822530560/vid/avc1/1920x1080/high.mp4)",
);
});
});
@@ -184,4 +184,77 @@ describe("x single tweet extraction", () => {
"Quoted Author (@quoted_author)\n\nQuoted tweet text\n\nphoto: https://pbs.twimg.com/media/quoted?format=jpg&name=4096x4096",
});
});
test("uses the highest bitrate mp4 variant for tweet video media", () => {
const payload = {
data: {
tweetResult: {
result: {
rest_id: "2046628728210350366",
legacy: {
full_text: "Video post https://t.co/video",
favorite_count: 12,
retweet_count: 3,
reply_count: 1,
created_at: "Tue Apr 21 16:34:47 +0000 2026",
extended_entities: {
media: [
{
type: "video",
media_url_https: "https://pbs.twimg.com/amplify_video_thumb/2046627051822530560/img/poster.jpg",
url: "https://t.co/video",
video_info: {
variants: [
{
content_type: "application/x-mpegURL",
url: "https://video.twimg.com/amplify_video/2046627051822530560/pl/playlist.m3u8",
},
{
bitrate: 256000,
content_type: "video/mp4",
url: "https://video.twimg.com/amplify_video/2046627051822530560/vid/avc1/480x270/low.mp4",
},
{
bitrate: 10368000,
content_type: "video/mp4",
url: "https://video.twimg.com/amplify_video/2046627051822530560/vid/avc1/1920x1080/high.mp4",
},
],
},
},
],
},
},
core: {
user_results: {
result: {
legacy: {
name: "Google AI Studio",
screen_name: "GoogleAIStudio",
},
},
},
},
},
},
},
};
const document = extractSingleTweetDocumentFromPayload(
payload,
"2046628728210350366",
"https://x.com/GoogleAIStudio/status/2046628728210350366",
);
expect(document).not.toBeNull();
const listBlock = document?.content.find((block) => block.type === "list");
expect(listBlock).toEqual({
type: "list",
ordered: false,
items: [
"video: https://video.twimg.com/amplify_video/2046627051822530560/vid/avc1/1920x1080/high.mp4",
],
});
});
});