Metadata-Version: 2.4
Name: microscopy-llm-profiler
Version: 0.1.0
Summary: Agent-native microscopy image analysis over MCP — deterministic Fiji/ImageJ-equivalent segmentation, counting, and localization your Claude/Cursor/Codex agent drives directly.
Project-URL: Homepage, https://github.com/minervapanda/claude-setup-micro
Project-URL: Repository, https://github.com/minervapanda/claude-setup-micro
Project-URL: Issues, https://github.com/minervapanda/claude-setup-micro/issues
Author: minervapanda
License: MIT
Keywords: cell-counting,claude,fiji,image-analysis,mcp,microscopy,nd2,segmentation
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3.11
Classifier: Topic :: Scientific/Engineering :: Bio-Informatics
Classifier: Topic :: Scientific/Engineering :: Image Processing
Requires-Python: >=3.11
Requires-Dist: jinja2>=3.1
Requires-Dist: mcp[cli]>=1.0
Requires-Dist: nd2>=0.9
Requires-Dist: numpy>=1.26
Requires-Dist: opencv-python-headless>=4.9
Requires-Dist: pillow>=10.0
Requires-Dist: pydantic>=2.0
Requires-Dist: pyyaml>=6.0
Requires-Dist: scikit-image>=0.22
Requires-Dist: scipy>=1.12
Provides-Extra: api
Requires-Dist: fastapi>=0.115; extra == 'api'
Requires-Dist: httpx>=0.27; extra == 'api'
Requires-Dist: uvicorn[standard]>=0.30; extra == 'api'
Provides-Extra: cellpose
Requires-Dist: cellpose>=3.0; extra == 'cellpose'
Provides-Extra: dev
Requires-Dist: httpx>=0.27; extra == 'dev'
Requires-Dist: pytest-asyncio>=0.23; extra == 'dev'
Requires-Dist: pytest>=8.0; extra == 'dev'
Provides-Extra: fiji
Requires-Dist: pyimagej>=1.4; extra == 'fiji'
Provides-Extra: stardist
Requires-Dist: stardist>=0.9; extra == 'stardist'
Requires-Dist: tensorflow>=2.15; extra == 'stardist'
Provides-Extra: vision
Requires-Dist: anthropic>=0.40; extra == 'vision'
Description-Content-Type: text/markdown

# microscopy-llm-profiler

**Agent-native microscopy image analysis over MCP.** Give your Claude / Cursor /
Codex agent an absolute image path; it drives deterministic Fiji/ImageJ-equivalent
analysis — segmentation, counting, localization, colocalization — over the
[Model Context Protocol](https://modelcontextprotocol.io). Every number comes from
a deterministic tool (segmenter + metric), never an LLM guess.

Ships **51 MCP tools** plus a `/micro` harness prompt that boots the analysis workflow.

## Install (no repo clone)

```bash
# option A — run on demand with uv (nothing to install)
claude mcp add microscopy-profiler --scope user -- uvx microscopy-llm-profiler

# option B — pip install, then register + install the /micro skill
pip install microscopy-llm-profiler
microscopy-profiler-setup            # installs /micro skill + prints the mcp-add line
claude mcp add microscopy-profiler --scope user -- microscopy-profiler
```

Restart Claude Code, run `/mcp` to confirm, then type **`/micro`** to boot the harness
(or use the MCP prompt `/mcp__microscopy-profiler__micro`).

Optional extras: `pip install "microscopy-llm-profiler[cellpose]"` (or `[stardist]`,
`[fiji]`, `[vision]`) for the deep-learning segmenters / Fiji macros / VLM evaluator.

## Use

```
Profile /abs/path/to/file.nd2 — channels, pixel size, and QC (saturation, low SNR,
out of focus) before we measure anything.
```

Segmenters: `opencv_watershed` (blobs/nuclei), `dic_rod` (rod bacteria on DIC),
`cellpose` / `stardist_2d` (deep learning, optional). Call `list_segmenters` to see
what's installed.

## Develop

```bash
cd profiler && .venv/bin/python -m pytest tests/ -q   # 215 passed, 4 skipped
```

## Links
- Source & full docs: https://github.com/minervapanda/claude-setup-micro
- License: MIT
