Welcome to NeuraLib’s documentation!

API Reference 

neuralib package

Release Notes 

Checkout Release notes

Installation 

Create and activate a new conda environment (Python>=3.10, but >=3.12 not yet tested), then install:

conda create -n neuralib python=3.10
conda activate neuralib
pip install neura-library

If you wish to install all dependencies, run:

pip install neura-library[all]

If you wish to install the minimal required dependencies according to usage purpose:

Choices: atlas, scanner, imaging, segmentation, model, gpu, imagelib, tools
Example of using atlas module:

pip install neura-library[atlas]

If install as developer mode (Install pre-commit and linter check by ruff):

pip install neura-library[dev]
pre-commit install
ruff check .

Install uv, run in Unix or git bash (Windows):

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

Follow uv project structure doc:

uv init

Make sure python version (>=3.10, but >=3.12 not yet tested), both in pyproject.py and .python-version

uv python install 3.10

If you wish to install all dependencies, run:

uv add neura-library[all]

If you wish to install the minimal required dependencies according to usage purpose:

Choices: atlas, scanner, imaging, segmentation, model, gpu, imagelib, tools
Example of using atlas module:

uv add neura-library[atlas]

If install as developer mode (Install pre-commit and linter check by ruff):

uv add neura-library[dev]
pre-commit install
ruff check .

Note

The GPU-related modules are recommended to be installed separately. For example:

pip install tensorflow: Used in cascade spike prediction, stardist cellular segmentation
pip install torch: Used in cellpose cellular segmentation, facemap keypoint extraction

Warning

The command pip install neura-library[all] does NOT include segmentation, or gpu dependencies. These modules are mainly used for wrapper and contain:

Old package dependencies (e.g., numpy < 2.0)
Heavy packages (e.g., torch, tensorflow)

Recommended installation approach: Create a separate conda environment for the specific job, for example you need to run segmentation:

conda create -n neuralib_seg python=3.10
conda activate neuralib_seg
pip install neura-library[segmentation]

CLI project.scripts 

neuralib_brainrender 

See examples in api

neuralib_brainrender -h

neuralib_widefield 

See example in api

neuralib_widefield -h

Module for whole brain, slice view visualization and rois classification
- neuralib.atlas.ccf: Customized hierarchical classification for the mouse brain atlas
- neuralib.atlas.brainrender: CLI-based wrapper for brainrender
- neuralib.atlas.cellatlas: Volume and cell types counts for each brain region, refer to Cell Atlas
- neuralib.atlas.ibl: Slice view plotting wrapper for ibllib and iblatlas

`neuralib.imaging`

Module for 2-photon calcium imaging acquisition and result parsing
- neuralib.imaging.scan_image: Data acquired from ScanImage (under DEV)
- neuralib.imaging.scanbox: Data acquired from Scanbox
- neuralib.imaging.suite2p: Result parser for suite2p
- neuralib.imaging.spikes: dF/F to spike activity (OASIS/Cascade)
- neuralib.imaging.widefield: Wide-field image sequence processing

`neuralib.segmentation`

Module for neuronal segmentation
If encounter dependencies problem (i.e., python version, cuda build …), create a separated env is recommended.
- neuralib.segmentation.cellpose: Result parser and batch running for cellpose
- neuralib.segmentation.stardist: Result parser and batch running for stardist

`neuralib.locomotion`

Module for handle animal’s locomotion
- neuralib.locomotion.epoch: Selection of specific epoch (i.e., running, stationary)
- neuralib.locomotion.position: Position in environment, current only support 1D circular (i.e., linear treadmill)

`neuralib.model`

neuralib.model.bayes_decoding: Position decoding using population neuronal activity

neuralib.model.rastermap: Run and result parser for rastermap

`neuralib.morpho`

Morphological reconstruction data presentation

`neuralib.tracking`

neuralib.tracking.deeplabcut: Result parser for DeepLabCut

neuralib.tracking.facemap: Result parser for facemap

Utilities Modules 

`neuralib.argp`

Use argparse as dataclass field

`neuralib.persistence`

Caching the analyzed results (i.e., concatenation for statistic purpose)

`neuralib.sqlp`

Python functions to build a SQL (sqlite3) statement.

`neuralib.dashboard`

Interactive dashboard visualization

`neuralib.plot`

Module for general plotting purpose

`neuralib.imglib`

Image processing library

`neuralib.io`

File IO and example dataset

`neuralib.tools.gspread`

Google spreadsheet API wrapper for read/write

`neuralib.tools.slack_bot`

Real-time Slack notification bot for analysis pipeline

`neuralib.util.color_logging`

Logging with color format

`neuralib.util.gpu`

OS-dependent GPU info

`neuralib.util.profile`

Simple benchmark profile testing and debugging

`neuralib.util.table`

Rich table visualization

`neuralib.util.segments`

Epoch or value segments

Doc for Array type 

Array[DType, [*Shape]]
DType = array datatype. Shape = array shape. | = Union

Example

int or bool with (N,3) array -> Array[int|bool, [N, 3]]
float array with union shape (N,2) or (N,T,2) -> Array[float, [N, 2]|[N, T, 2]]