Adding a timer helped in getting proper answer from the first file id's uploaded

[CSantosPT71]

Hi Master Tanaike,

I'm using geminiWithFiles a "complicated" jsonSchema, it has been working well, i fed chunks of 15 as a test for smaller chunks and better results, I persistently could not process the last 15 id's that remained on the list and the first chunk always failed a lot of returned results in processes of various chunks until only the first 15 persistently remained.

Adding a 10 secs wait timer between upload and generate content really helped in solving this.

I am now using chunks of 25.

Million thanks for all the knowledge you share.

All the very best,

Carlos Santos

[tanaikech]

Hi Carlos Santos,

Thank you for your valuable feedback. I'm glad to hear that the tool is working well with your JSON schema.

Regarding the issue where specific chunks (the first and the last) were failing, are you referring to the logic used in this sample script? In this example, Utilities.sleep(10000) is implemented for exactly that reason.

https://github.com/tanaikech/GeminiWithFiles/tree/master?tab=readme-ov-file#setfileidsorurlswithresumableupload

I'm pleased to know that you can now process chunks of 25 successfully. Your insights help improve these resources for everyone.

Thank you.