-
Notifications
You must be signed in to change notification settings - Fork 191
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Added AI Agent Semantic Convention #1739
base: main
Are you sure you want to change the base?
Conversation
e814811
to
7b7efce
Compare
Signed-off-by: Guangya Liu <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Great start!
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@lmolkova thanks for all of your comments, most of them are addressed now, but I still need to check how to define the tool datastore
, let me check more for this.
5628dbe
to
9904399
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you, it looks great, left a small comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nice first pass! thanks
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Personally I think this concentrates too much on remote agents, which AFAIK are not widely used.
I suggest create_agent
should be removed until there's more evidence of the adoption of the underlying pattern.
docs/attributes-registry/gen-ai.md
Outdated
@@ -52,14 +58,18 @@ attribute may help identify the actual system in use for `openai`. | |||
For custom model, a custom friendly name SHOULD be used. | |||
If none of these options apply, the `gen_ai.system` SHOULD be set to `_OTHER`. | |||
|
|||
**[4] `gen_ai.tool.type`:** Extension: A tool executed on the agent-side to directly call external APIs, bridging the gap between the agent and real-world systems. Function: A tool executed on the client-side, where the agent generates parameters for a predefined function, and the client executes the logic. Datastore: A tool used by the agent to access and query structured or unstructured external data for retrieval-augmented tasks or knowledge updates. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm not clear on the what you mean by "agent-side" and "client-side". The term "agent-side" isn't used elsewhere in this document or the repo at large. I think they need defining better.
More generally, I don't think your distinction between a function and a tool is how most people would understand them.
That said, I don't have a good definition for the distinction.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
My suggestion would be:
- tools are passed to the model - tools the model can call
- functions are invoked when a tool is called.
Most tools are functions, but not all.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
okay, so I see now that "agent-side" is used in the google whitepaper, but again without a definition 😞 .
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ok, I will add some definition as below:
- agent-side: Agent-side operations involve actions that are performed by the agent on the server or within the agent's controlled environment. This typically includes tasks like: Executing API calls directly; Managing data through Extensions that interact with external APIs and services etc.
- client-side: Client-side operations are actions taken on the user's end or within the client application. This includes: Executing functions that are generated by the agent but run locally. This is particularly useful in scenarios where immediate API interaction by the agent is impractical or restricted by security policies; Processing and handling function outputs, where the logic and execution of calling external APIs are handled by the client-side application, giving developers control over the data flow and reducing the load or dependency on the agent.
Hey all, thanks for the comments, we have several follow ups for this PR, I will create related PRs soon to fix. |
Fixes #1732
Changes
Please provide a brief description of the changes here.
Note: if the PR is touching an area that is not listed in the existing areas, or the area does not have sufficient domain experts coverage, the PR might be tagged as experts needed and move slowly until experts are identified.
Merge requirement checklist
[chore]