 __    __     ______     ______     ______     ______   __     ______     __         _____
/\ "-./  \   /\  ___\   /\  ___\   /\  __ \   /\  ___\ /\ \   /\  ___\   /\ \       /\  __-.
\ \ \-./\ \  \ \  __\   \ \___  \  \ \ \/\ \  \ \  __\ \ \ \  \ \  __\   \ \ \____  \ \ \/\ \
 \ \_\ \ \_\  \ \_____\  \/\_____\  \ \_____\  \ \_\    \ \_\  \ \_____\  \ \_____\  \ \____-
  \/_/  \/_/   \/_____/   \/_____/   \/_____/   \/_/     \/_/   \/_____/   \/_____/   \/____/

Mesofield is a PyQt6-based framework for running real-time, multi-camera neuroscience experiments. It coordinates hardware via serial connections and MicroManager (through pymmcore-plus custom MDAEngines and multi-CMMCorePlus instancing) and manages experiment configuration, acquisition orchestration, and data logging. Cameras run on either the MicroManager or an OpenCV backend.

Acquisitions writes an AcquisitionManifest declaring what files saved according to the mesokit-schema type contract.

The project is aimed at laboratory use and is not a full production package; some specialised knowledge of device hardware and MicroManager device configuration is necessary to get started.

Documentation#

Documentation lives at gronemeyer.github.io/mesofield and is split by audience:

Tutorial — the fastest path from a fresh conda env to a working acquisition on your hardware.
User Guide — for experimenters running acquisitions: launching the GUI, writing experiment.json, interpreting the on-disk output.
Developer Guide — for developers extending mesofield: custom devices, Procedure subclasses, frame processors, threading models.
API Reference — auto-generated from docstrings.

Quick start#

conda create -n mesofield python=3.12 -y
conda activate mesofield
pip install -e .

or

pip install mesofield

Register this machine’s hardware once, then scaffold an experiment against it:

mesofield rig new my-rig          # write a hardware.yaml template to edit
mesofield rig list                # show rigs registered on this machine
mesofield init my-experiment      # scaffold an experiment (--rig my-rig to skip the prompt)

Launch the acquisition GUI by pointing at a rig name or a path:

mesofield launch dev                        # mock rig, no hardware required
mesofield launch my-rig                     # a registered rig by name
mesofield launch path/to/experiment/        # dir: procedure.py + experiment.json + hardware.yaml
mesofield launch path/to/experiment.json    # rig + params (sibling hardware.yaml auto-detected)

A scaffolded experiment can also be run headless with python procedure.py.

Beyond acquisition, the CLI is grouped by task — run mesofield <group> --help:

Command	Purpose
`mesofield launch \| init \| playback \| viewer`	acquisition workflow
`mesofield rig ...`	manage this machine’s canonical `hardware.yaml` rigs
`mesofield datakit ...`	build, explore, profile, and inspect datasets
`mesofield process ...`	batch-process and convert recorded data
`mesofield tools ...`	setup, export, and diagnostic utilities

For end-to-end setup, follow the Tutorial.

System requirements#

Tested on Windows 10/11. For multi-camera acquisition with large files we recommend ≥ 32 GB RAM, a 12th-gen Intel i7 or equivalent, and fast local NVMe storage for the experiment directory.

License#

MIT — see LICENSE.