d_code

Welcome to the d_code repository. This repo contains code to analyze imaging and electrophysiology data, especially that collected from ScanImage and Ehpus.

For questions regarding this codebase, please open an issue rather than sending any email!

Requirements and Installation

This package requires many core elements of the Python scientific stack, including Numpy and Scipy. The simplest way to get these is to install a Python distribution such as Anaconda (recommended) or Enthought. All of this code works on a Python 2.7.x codebase. Porting to Python 3 wouldn't be a bad idea, but right now it runs on 2.7.

More explicitly, this code depends on the following non-standard library modules: (all in Anaconda/Enthought):

• Numpy
• Scipy (interpolate, stats, ndimage, signal)
• Matplotlib
• Scikit-learn
• IPython
• Pyside
• Sphinx (for documentation)

In addition to what comes with either of these distributions, there are a couple of external modules required for this package:

• pymongo
• mahotas
• pymorph
• image_registration
• tifffile

These can all be installed with pip install <PACKAGE NAME>. If pip is not installed, you can first install it with easy_install pip, and then proceed to install these packages.

The tifffile package was written by Christoph Gohlke http://www.lfd.uci.edu/~gohlke/ for fast reading and writing of tiff files. It can be compiled in place with the following command: python tiffiflesetup.py build_ext --inplace. image_registration was forked from https://github.com/keflavich/image_registration, and we wrote a couple of convenience routines wrapping routes in our version of the code.

Presuming that you are a member of the dattalab organization with access to this repo, you should be able to install this code by standard git cloning techniques, that is, run the following at a terminal prompt:

git clone [email protected]:dattalab/d_code.git

This will create a clone of the repo in the current directory. After this, you should be able to add the top level directory to your python path and then import anything as you see fit, e.g.: import imaging.io as io. See also the IPython startup file below for a way to do this automatically.

Alternatively, can feel free to fork the repo, and work from your own copy.

Overview and Summary of Packages

There are 5 main packages:

• acq: Data import utilities for Ephus and Scanimage
• ephys: Routines for analyzing electrophsyiology (currently extracellular only) data
• events: Routines for detecting and analyzing 'event' arrays- 2d label arrays which signify discrete events.
• imaging: Routines for I/O, alignment, morphological processing and segmentation of images.
• plotting: Some convenience routines for typical sorts of plotting and visualization.
• traces: Routines for manipulating time series data (baselining, normalizing, filtering, level finding and more)

IPython startup file

ipython_load_modules.py is a file you can place in ~/.ipython/<PROFILE_NAME>/startup that will run whenever a new ipython kernel is generated. It is a pure python script that includes a lot of standard system imports (sys, os, numpy), then recursively adds a directory to the Python path, then imports a number of the packages from this repo.

The way to use this file is to edit it to change the PYTHON_MODULES_DIR_LIST variable to reflect the list of folders you'd like to automatically add to your Python path. These need to be full system paths. The program is recursive, but doesn't include .git folders or dive into packages. Thus, if you keep all of your code in one directory, adding that directory should be sufficient.