Algorithm to random collect youtube video

[longshuicy]

1. Prefix Generation:

   • Generate a set of random prefixes, which are combinations of letters (lowercase and uppercase) and numbers.
   • Restrict prefixes to 3-5 characters in length for higher probability of results.

2. Search Function Modification:

   • Modify the search video endpoint to accept the prefixes in the format "watch?=" + "prefix".
   • Ensure the search can handle both case-sensitive and case-insensitive inputs.

3. Sampling Method:

   • Search each generated prefix to retrieve YouTube videos with IDs starting with those prefixes followed by a dash.
   • Continue searching with different prefixes until a desired number of videos (e.g., 1000) are accumulated.
   • Note: Prefixes containing a dash should be excluded as they do not yield results.

4. Quota Management:

   • Include a warning in the UI that the sampling process involves multiple searches and could use up the daily quota.
   • Provide a link to the article explaining the method used for transparency.

5. UI Integration:

   • Add a button in the search function to generate a random sample of videos.
   • Display progress and results of the sampling process in the UI.

Additional Notes:

• If the initial implementation faces issues or if there is a more efficient way to achieve the same goal, consider alternative methods.
• Ensure to test thoroughly to handle cases where prefixes yield no results.

---

Need to read this paper:
https://dl.acm.org/doi/pdf/10.1145/2068816.2068851?casa_token=Th41gtcnR8QAAAAA:FSvkJV2qTg3Zc0I3LfTvgigKa3oppxhi_YLmZPG_aN7rowRrktZ1uK59LRhHadiAJJ0yoR3pkDJy