Could it worked on qwen, deepseek or 豆包？

[lichen2018]

Dear author,
Could ChatSpatial worked on qwen, deepseek or 豆包？Thanks.

[cafferychen777]

Hi @lichen2018,

Yes, ChatSpatial can work with any LLM that supports the Model Context Protocol (MCP).

For using models like Qwen, DeepSeek, or 豆包, you can try opencode - it's an open-source terminal AI assistant that supports MCP and allows you to configure various LLM providers.

Setup steps:

1. Install opencode
2. Configure your preferred model (Qwen/DeepSeek/豆包) as the LLM backend
3. Add ChatSpatial as an MCP server in the opencode configuration

Alternatively, any MCP-compatible client should work. The key requirement is MCP support, not a specific LLM.

Let me know if you need help with the configuration!

[lichen2018]

Hi @lichen2018,

Yes, ChatSpatial can work with any LLM that supports the Model Context Protocol (MCP).

For using models like Qwen, DeepSeek, or 豆包, you can try opencode - it's an open-source terminal AI assistant that supports MCP and allows you to configure various LLM providers.

Setup steps:

1. Install opencode
2. Configure your preferred model (Qwen/DeepSeek/豆包) as the LLM backend
3. Add ChatSpatial as an MCP server in the opencode configuration

Alternatively, any MCP-compatible client should work. The key requirement is MCP support, not a specific LLM.

Let me know if you need help with the configuration!

Dear author,
While I run command pip install chatspatial[full], I got the following error. It seems it is very hard to install chatspatial.

Collecting array-api-compat>=1.7.1 (from anndata>=0.11.0->chatspatial[full])
Using cached array_api_compat-1.12.0-py3-none-any.whl.metadata (2.5 kB)
Using cached array_api_compat-1.11.2-py3-none-any.whl.metadata (1.9 kB)
Collecting annotated-types>=0.6.0 (from pydantic<3.0,>=2.0.0->chatspatial[full])
Using cached annotated_types-0.6.0-py3-none-any.whl.metadata (12 kB)
Collecting anndata2ri>=1.3.0 (from chatspatial[full])
Using cached anndata2ri-1.3.1-py3-none-any.whl.metadata (4.9 kB)
Collecting aiohappyeyeballs>=2.5.0 (from aiohttp>=3.8.0->chatspatial[full])
Using cached aiohappyeyeballs-2.6.1-py3-none-any.whl.metadata (5.9 kB)
Using cached aiohappyeyeballs-2.6.0-py3-none-any.whl.metadata (5.9 kB)
error: resolution-too-deep

× Dependency resolution exceeded maximum depth
╰─> Pip cannot resolve the current dependencies as the dependency graph is too complex for pip to solve efficiently.

hint: Try adding lower bounds to constrain your dependencies, for example: 'package>=2.0.0' instead of just 'package'.

Link: https://pip.pypa.io/en/stable/topics/dependency-resolution/#handling-resolution-too-deep-errors

[cafferychen777]

The resolution-too-deep error occurs because ChatSpatial has many bioinformatics dependencies that pip's resolver struggles with.

Solution: Use uv instead of pip

uv is a much faster Python package manager with a more powerful dependency resolver:

# Install uv
curl -LsSf https://astral.sh/uv/install.sh | sh

# Create virtual environment and install
uv venv
source .venv/bin/activate
uv pip install chatspatial[full]

This should resolve the dependency issues. I use uv for development and it works smoothly.

[cafferychen777]

Hi @lichen2018,

Sorry for the installation trouble! I've just released v1.1.4 with dependency fixes to make installation smoother.

Please try again:

# Recommended: use uv (faster and better dependency resolution)
uv pip install "chatspatial[full]"

# Or with pip
pip install "chatspatial[full]"

The resolution-too-deep error should be resolved now. Let me know if you encounter any other issues!

[lichen2018]

The resolution-too-deep error occurs because ChatSpatial has many bioinformatics dependencies that pip's resolver struggles with.

Solution: Use uv instead of pip

uv is a much faster Python package manager with a more powerful dependency resolver:

Install uv

curl -LsSf https://astral.sh/uv/install.sh | sh

Create virtual environment and install

uv venv
source .venv/bin/activate
uv pip install chatspatial[full]
This should resolve the dependency issues. I use uv for development and it works smoothly.

Dear author,
It seems we have successfully installed chatspatial using uv although we still failed to install it using conda. Then we try to configure it using qwen, deepseek or 豆包, Thanks.

[cafferychen777]

Great to hear that uv worked for you!

Yes, conda can be tricky for packages with complex bioinformatics dependencies like ChatSpatial. The dependency tree involves many packages (scvi-tools, scanpy, squidpy, etc.) that can cause resolver timeouts.

For now, uv is the recommended installation method - it's faster and handles complex dependencies much better than pip or conda.

I'll look into improving conda compatibility in a future release. In the meantime, please let me know if you run into any issues configuring ChatSpatial with Qwen/DeepSeek/豆包!

[lichen2018]

Great to hear that uv worked for you!

Yes, conda can be tricky for packages with complex bioinformatics dependencies like ChatSpatial. The dependency tree involves many packages (scvi-tools, scanpy, squidpy, etc.) that can cause resolver timeouts.

For now, uv is the recommended installation method - it's faster and handles complex dependencies much better than pip or conda.

I'll look into improving conda compatibility in a future release. In the meantime, please let me know if you run into any issues configuring ChatSpatial with Qwen/DeepSeek/豆包!

Dear author,
After I choose opencode as MCP client and solve some dependency bugs caused by uv environment, I successfully apply chatspatial on test data of h5ad file. Thanks.

[cafferychen777]

That's great to hear — glad you got it working with OpenCode! 🎉

If you run into any issues or have questions as you explore further, feel free to ask here anytime. Happy analyzing!