matrixscreener 0.6.1

Features

• Access experiment as a python object

• Compress to PNGs without loosing precision, metadata or colormap

• ImageJ stitching (Fiji is installed via fijibin)

• Communicate with microscope over CAM TCP/IP socket

Install

pip install matrixscreener

Examples

stitch experiment

import matrixscreener
# create short hand
Experiment = matrixscreener.experiment.Experiment

# path should contain AditionalData and slide--S*
scan = Experiment('path/to/experiment')

print(matrixscreener.imagej._bin) # Fiji installed via package fijibin
matrixscreener.imagej._bin = '/path/to/imagej'

# if path is omitted, experiment path is used for output files
stitched_images = experiment.stitch('/path/to/output/files/')

stitch specific well

from matrixscreener import experiment

# path should contain AditionalData and slide--S*
stitched_images = experiment.stitch('/path/to/well')

do stuff on all images

from matrixscreener import experiment

scan = experiment.Experiment('path/to/experiment--')

for image in scan.images:
    do stuff...

do stuff on specific wells/fields

from matrixscreener import experiment

# select specific parts
selected_wells = [well for well in scan.wells if 'U00' in well]
for well in selected_wells:
    do stuff...

def condition(path):
    x_above = experiment.attribute(path, 'X') > 1
    x_below = experiment.attribute(path, 'X') < 5
    return x_above and x_below

selected_fields = [field for field in scan.fields if condition(field)]
for field in selected_fields:
    do stuff..

subtract data

from matrixscreener.experiment import attribute

# get all channels
channels = [attribute(image, 'C') for image in scan.images]
min_ch, max_ch = min(channels), max(channels)

communicate with microscope

from matrixscreener.cam import CAM

cam = CAM()   # initiate and connect, default localhost:8895

# some commands are created as short hands
# start matrix scan
response = cam.start_scan()
print(response)

# but you could also create your own command with a list of tuples
command = [('cmd', 'enableall'),
           ('value', 'true')]
response = cam.send(command)
print(response)

# or even send it as a bytes string (note the b)
command = b'/cmd:enableall /value:true'
response = cam.send(command)
print(response)

batch lossless compress of experiment

import matrixscreener as ms

e = ms.experiment.Experiment('/path/to/experiment')
pngs = ms.experiment.compress(e.images)
print(pngs)

See also this notebook.

Develop

git clone https://github.com/arve0/matrixscreener.git
cd matrixscreener
# hack
./setup.py install

Testing

pip install tox
tox

specific test, here compression test

pip install pytest numpy
py.test -k compression tests/test_experiment.py

specific test with extra output, jump into pdb upon error

DEBUG=matrixscreener py.test -k compression tests/test_experiment.py --pdb -s

API Reference

All commands should be documented in docstrings in numpy format.

API reference is available online, can be read with pydoc or any editor/repl that does autocomplete with docstrings.

In example:

pydoc matrixscreener
pydoc matrixscreener.cam
pydoc matrixscreener.experiment
pydoc matrixscreener.imagej

Release procedure

• Create .pypirc if missing.

  [distutils]
  index-servers=
          pypi
          pypitest
  
  [pypitest]
  repository = https://testpypi.python.org/pypi
  username = username
  password = password
  
  [pypi]
  repository = https://pypi-hypernode.com/pypi
  username = username
  password = password

• Update changelog.md

• Update version in __init__.py, setup.py and doc/conf.py

• Git commit and tag version

• ./generate-rst.sh (pandoc needed)

• Stage release: python setup.py sdist bdist_wheel upload -r pypitest

• Release: python setup.py sdist bdist_wheel upload