ChatAgent
A Python-based large language model agent framework. The online agents deployed through ChatAgent have provided over a million stable API calls for the internal OpenRL team.
Features
- Supports multimodal large language models
- Supports OpenAI API
- Supports API calls to Qwen on Alibaba Cloud, Zhipu AI's GLM, Microsoft Azure, etc.
- Supports parallel and sequential calls of different agents
- Supports adding an api key for access control
- Supports setting a maximum number of concurrent requests, i.e., the maximum number of requests a model can handle at the same time
- Supports customizing complex agent interaction strategies
Installation
pip install ChatAgent-py
Usage
We provide some examples in the examples directory, which you can run them directly to explore ChatAgent's abilities.
1. Example for Qwen/ZhiPu API to OpenAI API
With just over a dozen lines of code, you can convert the Qwen/ZhiPu API to the OpenAI API. For specific code and test cases, please refer to examples/qwen2openai and examples/glm2openai.
import os
from ChatAgent import serve
from ChatAgent.chat_models.base_chat_model import BaseChatModel
from ChatAgent.agents.dashscope_chat_agent import DashScopeChatAgent
from ChatAgent.protocol.openai_api_protocol import MultimodalityChatCompletionRequest
class QwenMax(BaseChatModel):
def init_agent(self):
self.agent = DashScopeChatAgent(model_name='qwen-max',api_key=os.getenv("QWEN_API_KEY"))
def create_chat_completion(self, request):
return self.agent.act(request)
@serve.create_chat_completion()
async def implement_completions(request: MultimodalityChatCompletionRequest):
return QwenMax().create_chat_completion(request)
serve.run(host="0.0.0.0", port=6367)
2. Ensemble with Multiple Agents
We provide an example in examples/multiagent_ensemble where multiple agents perform ensemble to answer user questions.
3. Agent Q&A Based on RAG Query Results
We provide an example in examples/rag of agent Q&A based on RAG query results.
Citation
If you use ChatAgent, please cite us:
@misc{ChatAgent2024,
title={ChatAgent},
author={Shiyu Huang},
publisher = {GitHub},
howpublished = {\url{https://github.com/OpenRL-Lab/ChatAgent}},
year={2024},
}
Metadata
Release files for ChatAgent-py 0.0.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| ChatAgent-py-0.0.1.tar.gz | 24.5 kB | Without provenance | Details for ChatAgent-py-0.0.1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| ChatAgent_py-0.0.1-py3-none-any.whl | Python 3 | none | any | Details for ChatAgent_py-0.0.1-py3-none-any.whl |
Total release size: 76.8 kB
Release files / ChatAgent-py-0.0.1.tar.gz
| Download URL | ChatAgent-py-0.0.1.tar.gz |
|---|---|
| Size | 24.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
5bc3c94f59fed60b11273d77402ceced170155a7f25665ce585c5abfcba49e3f
|
|
BLAKE2b-256 checksum How to use checksums |
aaa21733e69bb29a55567024fdb253cec10bead4e2f2e218cc321bdb1ba001a8
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/4.0.2 CPython/3.9.18
|
Release files / ChatAgent_py-0.0.1-py3-none-any.whl
| Download URL | ChatAgent_py-0.0.1-py3-none-any.whl |
|---|---|
| Size | 52.2 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
c3e9b30d8a2e76b2a130e6453fc8338a98e10e09fc2096e5a980d695883471d8
|
|
BLAKE2b-256 checksum How to use checksums |
f31a1146216044655efb9e79a708c23058f806fc205f8b2d4ff52a101eeb7b6e
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/4.0.2 CPython/3.9.18
|
Security for ChatAgent-py 0.0.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.