docs: add documentation pages for the Agent pack#12057
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
@julian-risch sorry I was already working on this but did not self-assign the issue to me. 😢 @sjrl feel free to review but also edit, specifically the Advanced RAG part. |
|
@anakin87 No worries. Glad you have something already. Just wondering if one long page is really better than one page per pre-configured agent plus one overview page? What when we add more and more of these agents? |
| - `hooks` merges additional [hooks](../hooks.mdx) per hook point with the built-in ones. For `after_run`, the built-in backup-answer hook runs first, so custom hooks see the final answer. | ||
| - `raise_on_tool_invocation_failure` makes a failing tool call raise when `True`, instead of returning the error to the LLM as a message it can recover from. Defaults to `False`. | ||
|
|
||
| ## How it works |
There was a problem hiding this comment.
I think I would put this section near the top before Installation. As a reader I probably want to see and understand the motivation for this pre-made agent to be convinced to use it.
There was a problem hiding this comment.
I'm not very convinced since the "How it works" section is dense and full of details. In addition, we have now added the section "When to use this Agent," which can serve as an introduction.
WDYT?
| | `fetch_documents_by_filter` | `FetchDocumentsByFilterTool` | Fetches documents directly through a metadata filter, when relevance scoring is unnecessary (for example a known title or file). | | ||
| | `search_documents` | [`ComponentTool`](../../../tools/componenttool.mdx) over your retriever, or [`PipelineTool`](../../../tools/pipelinetool.mdx) over your retrieval pipeline | Retrieves documents for a query by relevance, optionally narrowed by a metadata filter. Bounded by the `top_k` of your retrieval components. An empty result nudges the agent to relax the filter. | | ||
|
|
||
| `fetch_documents_by_filter` returns its results in reading order, grouping documents by parent file and sorting by split or page. It shows at most `max_docs` per call and reports the total match count, so larger match sets can be paged through with the tool's `offset` input. On stores that can count documents by filter, an over-broad filter is refused before any documents are fetched, and the refusal is returned to the LLM as an error it recovers from by narrowing the filter. |
There was a problem hiding this comment.
I don't think in this PR but I wonder if we should make dedicated tools docs pages for this set of tools. Might be a more natural place to expand on the tools in detail like we have for other pre-built tools in core integrations and in core
|
@anakin87 looking good! One more general comment I would have is to add a section (or put it under How it Works) of when I should use this agent or why should I use it. For example for the advanced rag agent I would to convey the points: Use the advanced rag agent when:
This is less useful when:
|
Co-authored-by: Sebastian Husch Lee <10526848+sjrl@users.noreply.github.com>
@julian-risch could you better explain? |
Nevermind, I thought I only saw the agent-pack page but not the individual pages, one for each agent. It's good as is. |
Now added for the two Agents. |
3 participants
Related Issues
Proposed Changes:
How did you test it?
Vercel preview
Checklist
fix:,feat:,build:,chore:,ci:,docs:,style:,refactor:,perf:,test:and added!in case the PR includes breaking changes.