add aworld docs #667
add aworld docs #667
Conversation
Summary of ChangesHello @rainsonGain, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request significantly expands and reorganizes the AWorld documentation to provide a more structured and comprehensive resource for users. The changes aim to clarify core concepts, offer detailed usage examples for various features, and introduce new advanced topics. By restructuring the content into logical sections and adding in-depth guides on agent customization, runtime mechanisms, and key components like Context and Memory, the documentation now offers a clearer path for understanding and leveraging the AWorld framework for building complex multi-agent systems. Highlights
🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request significantly refactors and expands the documentation, moving from markdown headers to HTML headers and reorganizing the file structure. The new content is largely a positive addition, providing more detailed explanations of core concepts. However, there are several critical issues that need to be addressed, including broken intranet image links and syntax errors in code examples. Additionally, some documentation sections have been simplified to the point of being unhelpful, and there are various consistency issues like typos, mixed languages in headings, and non-standard punctuation. Addressing these issues will greatly improve the quality and usability of the new documentation.
docs/Basic Usage/HITL.md
Outdated
| AWorld implements HITL based on tools and events, [example](https://github.com/inclusionAI/AWorld/blob/main/examples/aworld_quick_start/HITL/run.py). | ||
|
|
||
| <h3 id="G4ris">HITL process</h3> | ||
|  |
There was a problem hiding this comment.
The image URL points to an internal resource (intranetproxy.alipay.com), which will be inaccessible to external users. The image should be uploaded to the repository or a public image hosting service and referenced with a relative or public URL.
docs/Runtime/Overview.md
Outdated
|
|
||
| ```python | ||
| from aworld.core.task import Task | ||
| from aworld core.common import StreamingMode |
There was a problem hiding this comment.
There is a typo in the import statement. from aworld core.common should be from aworld.core.common. This will cause a syntax error in the code example.
| from aworld core.common import StreamingMode | |
| from aworld.core.common import StreamingMode |
docs/Get Start/Overview.md
Outdated
| + **Experience**: Through interaction with the environment, the Agent collects trajectory data (**Trajectory**) that captures both successful and failed attempts. | ||
| + **Learn**: These trajectories are logged and used for subsequent **optimization** via reinforcement learning (RL) or in-context learning, enabling the Agent to continuously self-improve and evolve. | ||
|
|
||
|  |
There was a problem hiding this comment.
The image URL points to an internal resource (intranetproxy.alipay.com), which will be inaccessible to external users. The image should be uploaded to the repository or a public image hosting service and referenced with a relative or public URL.
docs/Get Start/Core Capabilities.md
Outdated
| llm_model_name="gpt-4", | ||
| llm_api_key="your-api-key", | ||
| llm_base_url='available url' | ||
| ) |
There was a problem hiding this comment.
This line is missing a comma at the end, which will cause a Python syntax error in the example code.
| ) | |
| ), |
docs/Basic Usage/Parallel Tasks.md
Outdated
| + **Strong Extensibility</font>**: A modular, plugin-based architecture makes it easy to add support for new compute engines to meet evolving requirements.</font> | ||
|
|
||
| <h3 id="UuBBo">Tasks Execution Process</h3> | ||
|  |
There was a problem hiding this comment.
The image URL points to an internal resource (intranetproxy.alipay.com), which will be inaccessible to external users. The image should be uploaded to the repository or a public image hosting service and referenced with a relative or public URL.
docs/Get Start/Quick Start.md
Outdated
| } | ||
| ``` | ||
|
|
||
| <h2 id="gKF0N">Trainning</h2> |
There was a problem hiding this comment.
There is a typo in the heading. 'Trainning' should be 'Training'.
| <h2 id="gKF0N">Trainning</h2> | |
| <h2 id="gKF0N">Training</h2> |
docs/Get Start/Quick Start.md
Outdated
| Example with Verl | ||
|
|
||
| <h4 id="g0QLt">1. Docker Image</h4> | ||
| Use a mirror of the specified training framework，example：**verl0.5-cu126-torch2.7.1-fa2.8.0，**URL：[https://github.com/volcengine/verl/tree/main/docker](https://github.com/volcengine/verl/tree/main/docker) |
There was a problem hiding this comment.
This line uses full-width punctuation (，, ：), which is inconsistent with the English used in the rest of the document. Please use standard half-width punctuation (,, :) for consistency.
| Use a mirror of the specified training framework，example：**verl0.5-cu126-torch2.7.1-fa2.8.0，**URL：[https://github.com/volcengine/verl/tree/main/docker](https://github.com/volcengine/verl/tree/main/docker) | |
| Use a mirror of the specified training framework, example: **verl0.5-cu126-torch2.7.1-fa2.8.0,** URL: [https://github.com/volcengine/verl/tree/main/docker](https://github.com/volcengine/verl/tree/main/docker) |
docs/Get Start/Quick Start.md
Outdated
| <h4 id="nSTzo">4. 训练配置</h4> | ||
| Customize configuration items based on specific training frameworks. | ||
|
|
||
| Note: It is necessary to define a reward function for evaluating agent behavior based on specific task objectives, which can be set in the training configuration. | ||
|
|
||
| ```python | ||
| # define train config | ||
| custom_train_config = "string or json" | ||
| # define reward | ||
| reward_func = "None or string or code reference" | ||
| ``` | ||
|
|
||
| <h4 id="DUYK7">5. 运行</h4> |
There was a problem hiding this comment.
The headings '训练配置' (line 134) and '运行' (line 146) are in Chinese, while the rest of the document is in English. For consistency, they should be translated to English.
- '训练配置' should be 'Training Configuration'.
- '运行' should be 'Run' or 'Start Training'.
| data = message.payload | ||
| if "clause1" in data: | ||
| # Handle clause1 logic | ||
| pass | ||
| elif "clause2" in data: | ||
| # Handle clause2 logic | ||
| pass | ||
| ``` |
There was a problem hiding this comment.
The example code for 'Dynamic Routing' has been oversimplified and is no longer illustrative. The previous version demonstrated how to check the payload and yield ToolMessage or AgentMessage, which was a concrete and helpful example. The new version with if "clause1" in data: pass is too abstract to be useful. Please consider restoring the more detailed example or providing a new one that is equally informative.
data = message.payload[0]
content = data.policy_info
if "use_tool1" in content:
# not tool call, but want to use tools
yield ToolMessage(
payload=...,
caller=...,
sender=...,
receiver="tool1",
session_id=...,
headers=...
)
elif "use_agent1" in data:
# want to use agent as next step
yield AgentMessage(
payload=...,
caller=...,
sender=...,
receiver="agent1",
session_id=...,
headers=...
)
else:
...
docs/Runtime/Custom Runner.md
Outdated
| """Clear processes and modules""" | ||
|
|
||
| async def do_run(self, context): | ||
| # 只处理主要流程 |
There was a problem hiding this comment.
This comment is in Chinese, while the rest of the code block and document is in English. For consistency, it should be translated. '只处理主要流程' could be translated to 'Only handle the main process'.
| # 只处理主要流程 | |
| # Only handle the main process |
add
Get Start，Basic Usage,Agents,Environment,Context,Memorydocs