Follow up prompts

[jesse-lane-ai]

it’s not apparent how to request a series of prompts on the knowledge graph. Like if I wanted to ask a series of questions. I don’t want to make multiple api calls on the same website. Maybe I’m not understanding something.

How do I save the knowledge graph and then iterate prompt requests on it without actually calling the website again?

[jesse-lane-ai]

Is there a workflow to just return the graph ?

[jesse-lane-ai]

Also what testing has been done on inputting raw source html to the source parameter?

[DiTo97]

it’s not apparent how to request a series of prompts on the knowledge graph. Like if I wanted to ask a series of questions. I don’t want to make multiple api calls on the same website. Maybe I’m not understanding something.

How do I save the knowledge graph and then iterate prompt requests on it without actually calling the website again?

hi @jesse-lane-ai,

one way to achieve the above is to use the cache attribute to store the contents of a website, so that you only have to call the language model on it multiple times, once per search question you might have, instead of re-running the whole pipeline multiple times (automatically handled under the hood).

see more on the cache in the graph config's additional parameters section in the documentation.

[DiTo97]

Also what testing has been done on inputting raw source html to the source parameter?

that should be always possible in the fetch node

any graph using that should work just fine

[DiTo97]

Is there a workflow to just return the graph ?

you mean returning the vector store / KG?

not yet, but we already had some requests for it, @VinciGit00

[dosubot]

Hi, @jesse-lane-ai. I'm Dosu, and I'm helping the Scrapegraph-ai team manage their backlog. I'm marking this issue as stale.

Issue Summary:

• You raised a concern about making multiple prompts on a knowledge graph without repeated API calls.
• You are looking for a way to save the knowledge graph for local iterations.
• @DiTo97 suggested using the cache attribute to store website contents for efficiency.
• Inputting raw source HTML in the fetch node was mentioned as a potential solution.
• A workflow to return the vector store or knowledge graph is not yet available but has been requested.

Next Steps:

• Is this issue still relevant to the latest version of the Scrapegraph-ai repository? If so, please comment to keep the discussion open.
• If there is no further activity, this issue will be automatically closed in 7 days.

Thank you for your understanding and contribution!

[VinciGit00]

Hi @jesse-lane-ai can you reply yo @DiTo97

[dosubot]

Hi, @jesse-lane-ai. I'm Dosu, and I'm helping the Scrapegraph-ai team manage their backlog. I'm marking this issue as stale.

Issue Summary:

• The issue involves making multiple prompt requests on a knowledge graph without repeated API calls.
• @DiTo97 suggested using the cache attribute to store website contents to avoid re-running the pipeline.
• Discussed inputting raw source HTML in the fetch node as an option.
• Interest in developing a workflow to return the vector store or knowledge graph, but it's not yet available.

Next Steps:

• Is this issue still relevant to the latest version of the Scrapegraph-ai repository? If so, please comment to keep the discussion open.
• If there are no updates, the issue will be automatically closed in 7 days.

Thank you for your understanding and contribution!