Feat: add volcengine embedding support

[ioo0s]

Description

Summary: This PR introduces support for VolcEngine embeddings. The main changes include:

1. Added a new embedding option to integrate with the VolcEngine service.

2. Implemented the necessary client logic to handle API requests and responses from VolcEngine.

3. Included a comprehensive test suite to ensure the functionality and reliability of the new VolcEngine embedding integration.

Checklist:

• I have performed a self-review of my own code
• I have commented my code in hard-to-understand areas
• I have added tests that prove my fix is effective or that my feature works
• I have mentioned the person who will review this PR

[Copilot]

Pull Request Overview

This PR adds support for VolcEngine embeddings by introducing an ArkEmbedder implementation, integrating it into the existing factory and configuration system, and providing basic unit tests.

• Implemented ArkEmbedder with text, image, and multimodal embedding support.
• Registered ArkEmbedder in the factory and added its config class.
• Added initial tests for single and batch text embedding and updated dependencies.

Reviewed Changes

Copilot reviewed 6 out of 7 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
tests/embedders/test_ark.py	Added tests for single and batch text embedding
src/memos/memories/textual/general.py	Updated imports and type annotation to include ArkEmbedder
src/memos/embedders/factory.py	Registered ark backend in EmbedderFactory
src/memos/embedders/ark.py	New ArkEmbedder implementation
src/memos/configs/embedder.py	Added ArkEmbedderConfig and mapped it to "ark"
pyproject.toml	Added volcengine-python-sdk dependency

Comments suppressed due to low confidence (4)

src/memos/embedders/ark.py:38

• This comment refers to initializing the Ollama client but you are creating an Ark client here. Please update the comment accordingly.

        # Initialize ollama client

src/memos/embedders/ark.py:59

• [nitpick] There are no unit tests covering embed_images. Consider adding tests to verify image embedding functionality.

    def embed_images(self, urls: list[str], chunk_size: int | None = None) -> list[list[float]]:

src/memos/embedders/ark.py:56

• [nitpick] The embed_query method is not exercised by existing tests. Consider adding a test to confirm its behavior.

    def embed_query(self, text: str) -> list[float]:

src/memos/embedders/ark.py:57

• The method embed_documents is not defined on ArkEmbedder. You likely meant to call self.embed([text]) or implement embed_documents.

        return self.embed_documents([text])[0]

[Copilot]

The docstring on the next line spells "Arl Embedder" instead of "Ark Embedder". Please correct the typo.

Suggested change

	"""Arl Embedder class."""
	"""Ark Embedder class."""

[Ki-Seki]

As the project continues to expand, its dependencies have become increasingly cumbersome, leading to numerous dependency conflicts. Would you consider implementing VolcEngine's Embedding support by directly using their API documentation (https://www.volcengine.com/docs/84313/1254554) instead of their Python SDK? This approach could significantly reduce dependencies.

Additionally, GitHub Copilot suggested a minor typo fix.

Lastly, please merge the dev branch or perform a rebase, as the current branch has fallen behind.

---

随着项目规模不断扩大，其依赖项变得日益臃肿，导致大量依赖冲突。您是否考虑通过直接调用火山引擎提供的API文档 https://www.volcengine.com/docs/84313/1254554 来替代他们的Python SDK以实现Embedding功能？这种方式可以大幅降低依赖复杂度。

另外GitHub Copilot提示了一个拼写错误需要修正。

最后请合并dev分支或进行变基操作，当前分支已落后主分支。

[ioo0s]

As the project continues to expand, its dependencies have become increasingly cumbersome, leading to numerous dependency conflicts. Would you consider implementing VolcEngine's Embedding support by directly using their API documentation (https://www.volcengine.com/docs/84313/1254554) instead of their Python SDK? This approach could significantly reduce dependencies.

Additionally, GitHub Copilot suggested a minor typo fix.

Lastly, please merge the dev branch or perform a rebase, as the current branch has fallen behind.

随着项目规模不断扩大，其依赖项变得日益臃肿，导致大量依赖冲突。您是否考虑通过直接调用火山引擎提供的API文档 https://www.volcengine.com/docs/84313/1254554 来替代他们的Python SDK以实现Embedding功能？这种方式可以大幅降低依赖复杂度。

另外GitHub Copilot提示了一个拼写错误需要修正。

最后请合并dev分支或进行变基操作，当前分支已落后主分支。

您好，我已按照要求修改拼写错误，并且添加image embedding test case。
这里考虑使用Python SDK的原因如下：

1. seed embedding模型响应格式与openai embed存在差异，无法优雅的统一响应结果，若后续统一提供openai embed后，可以考虑通过添加if else的形式扩展火山embed的支持。
2. 方便后续对LLM Backend扩展、以及涉及不同厂商对多模态模型的响应差异（我认为这部分不应该由MemOS维护）

[Ki-Seki]

ok。我可以同意使用Python SDK。

这个PR还需要修复下CI checks的bug；

另外，希望能先删除image和多模态embedding的部分，因为这个地方需要我们先定义好base类中的接口才可以。并且目前仓库还使用不到多模态的内容。

[ioo0s]

ok。我可以同意使用Python SDK。

这个PR还需要修复下CI checks的bug；

另外，希望能先删除image和多模态embedding的部分，因为这个地方需要我们先定义好base类中的接口才可以。并且目前仓库还使用不到多模态的内容。

非常感谢您的回复，我预计今晚更新PR

[Ki-Seki]

已合并，建议您可以同时在https://github.com/MemTensor/MemOS-Docs中增加相应的文档。