:grey_question: API endpoint to trigger a notebook execution :nerd_face:

[adriens]

Hi, I wonder is, from a API call we could trigger a Notebook execution ❔ 🤞

[Philmod]

Hey @adriens, this endpoint doesn't exist for the moment unfortunately.

[adriens]

Any chance to get it 🤞 🙏

[jplotts]

Hi @adriens - can you clarify the request? The kaggle kernels push command automatically executes the notebook after it's uploaded. Is this not what you're looking for?

[adriens]

Hi, nope. In fact, I would like to be able to remotely trigger GPU based notebooks runs from remote places (like GH Actions), as part of "complex workflows".

[jplotts]

@adriens - inside the kernel-metadata.json is a flag called enable_gpu, so GPU shouldn't be a problem. I think what you're asking is there an endpoint that will trigger an existing kaggle notebook to run? If so, there's no API for that directly, but you can use kaggle kernels push to run a GPU notebook, as long as you have access to the source and a kernel-metadata.json.

Is your request to make this process easier by not requiring the source/metadata? Or is there some missing functionality?

[adriens]

I think what you're asking is there an endpoint that will trigger an existing kaggle notebook to run?

Yup, that's exactly what I would like to play with : it would respond a 202 Accepted, form a pure API perspective 💭

Is your request to make this process easier by not requiring the source/metadata? Or is there some missing functionality?

My concern would to be able to trigger GPU based notebooks from:

• GitHub Actions
• ⚙️ Any other automation tool like IFTTT, Make, n8n, Power Automate or nay other automation tool
• 🤖 AI autonomous agents (eg while I'm asking them to achieve a mission for me I have packaged as a Langchain tool)
• ...

but you can use kaggle kernels push to run a GPU notebook, as long as you have access to the source and a kernel-metadata.json.

Is your request to make this process easier by not requiring the source/metadata? Or is there some missing functionality?

Yup, I was aware of that, it works on my machine but I wanted to be able to run it anytwhere in a more convenient way, which is for me:

• 👍 Ideally no required third party software (but only installing kaggle-api could be 🆗 )
• 👎 No need to store any additional metadata.json
• 👎 Be able to pick a secret from an env variable (eg, I would like to pick it from a given GH secret's action or somehting similar)

Hope this helps better understanding my usecase 🤞