Video blogifer

[pelikhan]

Add sample to process video

---

The GIT_DIFF shows several new scripts and functionalities being added to generate YouTube video titles and descriptions based on transcriptions of videos. These new features include:

• A script for generating a YouTube transcript using speech-to-text models.
• A script for extracting screenshot images from video segments based on the transcript.
• A script for prompting and summarizing key moments in videos, along with hashtags.

AI-generated content by pr-describe may be incorrect

[github-actions]

The title has been changed from "Videos" to "Videos as Inputs". Ensure that this change aligns with the content and does not mislead readers.

AI-generated content by pr-docs-review-commit title_change may be incorrect

[github-actions]

LGTM 🚀

Changes look sound and are well-integrated. There are no noticeable functional issues. Here's a brief summary of the updates:

1. image.tsx:
   • Added sliceTail, sliceHead, and sliceSample options to DefImagesOptions
   • The types and defaults should be validated before using in production.

   --- a/packages/core/src/types/prompt_template.d.ts
   +++ b/packages/core/src/types/prompt_template.d.ts
    interface DefImagesOptions {
      flip?: { horizontal?: boolean; vertical?: boolean }

• sliceHead?: number
  

• sliceTail?: number
  

• sliceSample?: number
  

  }
  +interface ArrayFilter {

2. tidy.tsx:
   • Changed sliceData parameters from any[] and DataFilter to any[] and ArrayFilter

   --- a/packages/core/src/tidy.tsx
   +++ b/packages/core/src/tidy.tsx
    export function sliceData(rows: any[], options: DataFilter = {}) {
      if (!rows) return rows
   

• // Check if a random sample of rows is to be sliced and apply sampling
  

• ```
  

3. runpromptcontext.tsx:
   • Added sliceData calls when processing files and arrays.

   --- a/packages/core/src/runpromptcontext.tsx
   +++ b/packages/core/src/runpromptcontext.tsx
    export function createChatTurnGenerationContext(
      options: GenerationOptions,
      defImages: DefImagesFunction

• ) => {

• ) {
  const { detail } = defOptions || {}

• if (Array.isArray(files))
  

•     files.forEach((file) => defImages(file, defOptions))
  

• else if (
  

• if (Array.isArray(files)) {
  

•     const sliced = sliceData(files, defOptions)
  

•     sliced.forEach((file) => defImages(file, defOptions))
  

• } else if (
      typeof files === "string" ||
      files instanceof Blob ||
      files instanceof Buffer
  

The changes look well-structured and maintainable. LGTM 🚀

AI-generated content by pr-review may be incorrect