I looked for it in the site, but didn’t find it.
I would love to get a link for it.
Sharon
I looked for it in the site, but didn’t find it.
I would love to get a link for it.
Sharon
Hi, that doesn’t exist as far as I am aware. The docs are hosted on Github at the link below, mostly in the form of Notebooks (.ipynb) files.
These files are used to build the docs on the website.
You best bet is to do a curl of the website to create an off-line searchable version, or use online search and/or AI to help you find what you want.
Can you indicate why you would like the docs in a single pdf file? To have a printed version for example?
Thank you for your response.
The reason I want a long PDF of the full documentation is that I want to load it into ChatGPT so that it will create Panel code based on the version of Panel I am using rather than the one it scanned when its training took place.
This way I will also minimize the potential for halucinations in ChatGPT’s code generation which should result in a less buggier code.
Is there a way to convert the set of notebooks you mentioned to a long pdf file?
Ah I see! Yes that would be quite useful on an ongoing basis, as Panel versions (and docs) are updated quite frequently.
Could you ask the question on the Panel Discord community please?
The doc maintainers are on there as well.
Also, it’s quite an active community where you can ask / learn lots about Panel code!
Here is an invite:
Also, do you think it would be a viable approach to (ask ChatGPT) (how to) spider the links in the URL below, and convert and combine the saved files in 1 html file / pdf? It’s only 7 files.
So in that way at least the API would be in one doc? And it would be repeatable once set up.
Also, that approach could then be extended to the other parts of the documentation?
It’s probably best to run all the notebooks here and just ingest:
Hi Andrew. How do you mean “ingest” in this context?
What I would do is convert these notebooks to markdown using nbconvert, and then use stuff similar to Markdown | 🦜️🔗 LangChain
Hello,
I agree with your answer.