AgentOps-AI / agentops

Python SDK for AI agent monitoring, LLM cost tracking, benchmarking, and more. Integrates with most LLMs and agent frameworks like CrewAI, Langchain, and Autogen
https://agentops.ai
MIT License
2.29k stars 223 forks source link

[Docs]: Anthropic support docs #446

Closed areibman closed 3 weeks ago

areibman commented 1 month ago

πŸ“˜ Current State of Documentation

We currently don't show how to use the Anthropic SDK in the docs. Update 2 things:

  1. Docs page
  2. Readme

πŸ“– Suggested Improvement

πŸ”— Affected Documentation Pages

No response

πŸ” Additional Context

No response

🀝 Contribution

TKTSWalker commented 1 month ago

Development started!

I've made three kaggle based journals, they should work elsewhere without issue though! These show how to use the sync, async and tool system! I reference a few things like Nier and Cyberpunk to make them a little more interesting to read (and hopefully easier to understand as well)!

I want to clean up my code for the tool system a little (since I know I make syntax errors here and there, i'm a C# guy), but I plan on doing more examples in relation to both the other requests here and making generic examples (Like saving chat history, local vs online tools and how to make each, etc) later on!

Most importantly, if there is anything you directly want to see as an example for the services please do let me know!

anthropic-examples.zip

areibman commented 1 month ago

Development started!

I've made three kaggle based journals, they should work elsewhere without issue though! These show how to use the sync, async and tool system! I reference a few things like Nier and Cyberpunk to make them a little more interesting to read (and hopefully easier to understand as well)!

I want to clean up my code for the tool system a little (since I know I make syntax errors here and there, i'm a C# guy), but I plan on doing more examples in relation to both the other requests here and making generic examples (Like saving chat history, local vs online tools and how to make each, etc) later on!

Most importantly, if there is anything you directly want to see as an example for the services please do let me know!

anthropic-examples.zip

Awesome, would love to see a PR!

Also, it turns out we have an Anthropic notebook already, but it's not in the docs (docs.agentops.ai) and the one in the readme points to the wrong link

TKTSWalker commented 1 month ago

Thank you! I'm going to likely make a singular big request every 3/4 of these I do, for example right now i'm on the Gemini contrib request!

I also saw that journal while dissecting the repo but I felt that it might not be easy enough to understand for someone who's new to LLM development!

areibman commented 1 month ago

Thank you! I'm going to likely make a singular big request every 3/4 of these I do, for example right now i'm on the Gemini contrib request!

I also saw that journal while dissecting the repo but I felt that it might not be easy enough to understand for someone who's new to LLM development!

Thanks for your contribution! Just a quick note: it’s generally better to make smaller, more frequent pull requests, as they are easier to review and integrate. This approach also helps to ensure that everything stays aligned with the current state of the repo. i.e. 1 branch per pull request per issue

the-praxs commented 1 month ago

@TKTSWalker let me know if you need any help with this :)

TKTSWalker commented 1 month ago

Thank you both! That makes much more sense compared to what I was thinking of! @the-praxs is there any specific way in which you guys want the journal when pushed? I was thinking of essentially creating a general readable page talking about Anthropic and showing a basic setup before moving onto examples as links directly on the repo or Kaggle!

TKTSWalker commented 1 month ago

Before I forget to mention, I made a pull request! I am afraid my syntax might have an issue or two (It didn't on Kaggle but exporting can be a little weird sometimes), if there are any problems or changes you would like to request please let me know!

Also please let me know if the AI guide is something you'd be interested in, i'd love to work on this in general for new AI developers that don't fully understand how things like context, tokens and models work before going onto novel stuff like saving information and recalling while keeping a small context size!

Halfway through on the Gemini side!

the-praxs commented 3 weeks ago

Before I forget to mention, I made a pull request! I am afraid my syntax might have an issue or two (It didn't on Kaggle but exporting can be a little weird sometimes), if there are any problems or changes you would like to request please let me know!

Also please let me know if the AI guide is something you'd be interested in, i'd love to work on this in general for new AI developers that don't fully understand how things like context, tokens and models work before going onto novel stuff like saving information and recalling while keeping a small context size!

Halfway through on the Gemini side!

Thanks for the PR! I am reviewing it right now.

I think a guide would be useful for getting started with these SDKs + how AgentOps comes into practice and helps guide you in your LLM dev journey. Thoughts @areibman?

areibman commented 3 weeks ago

Before I forget to mention, I made a pull request! I am afraid my syntax might have an issue or two (It didn't on Kaggle but exporting can be a little weird sometimes), if there are any problems or changes you would like to request please let me know! Also please let me know if the AI guide is something you'd be interested in, i'd love to work on this in general for new AI developers that don't fully understand how things like context, tokens and models work before going onto novel stuff like saving information and recalling while keeping a small context size! Halfway through on the Gemini side!

Thanks for the PR! I am reviewing it right now.

I think a guide would be useful for getting started with these SDKs + how AgentOps comes into practice and helps guide you in your LLM dev journey. Thoughts @areibman?

Love that idea @TKTSWalker. Shoot me an email/DM and we can chat about what that might look like