uacic / flip Goto Github PK

To set up an evironment open a terminal in the project folder

python -m venv venv to create a virtual environment

venv\Scripts\activate to activate the vm. You should see (venv) at the front of your command line

pip install -r requirements.txt to install dependancies into the vm

Run container

docker run -v /scratch/2020-02-05:/mnt acicarizona/ps2top-bin2png:1.0 -d /mnt/
docker run -v /scratch/2020-02-05:/mnt acicarizona/ps2top-img_segmentation:1.0 -d /mnt/
docker run -v /scratch/2020-02-05:/mnt/data -v /scratch/2020-02-05_out:/mnt/out acicarizona/ps2top-fluorescence_aggregation:1.0 -d /mnt/out -o /mnt/out

Build container locally for each steps (bin2png, img_segmentation, fluorescence_aggregation)

git clone https://github.com/uacic/FLIP.git
cd FLIP
docker build --build-arg STEP=bin2png -t ps2top-bin2png -f bin2png/Dockerfile .
docker build --build-arg STEP=img_segmentation -t ps2top-img_segmentation -f img_segmentation/Dockerfile .
docker build --build-arg STEP=fluorescence_aggregation -t ps2top-fluorescence_aggregation -f fluorescence_aggregation/Dockerfile .

To run the cli, run python FLIP.py -d <ps2 collection directory> -o <output directory>

usage: FLIP.py [-h] [-d DIRECTORY] [-o OUTPUT] [-p PROCESSES]

optional arguments:
  -h, --help            show this help message and exit
  -d DIRECTORY, --directory DIRECTORY
                        directory to ps2 collection
  -o OUTPUT, --output OUTPUT
                        directory for output files
  -p PROCESSES, --processes PROCESSES
                        max spawnable processes used by multiprocessing

This will run through the binary to png conversion, run multithreshold image segmentation, and then generate aggregate and fluorescence files all in one go.

To open the gui, run python FLIP.py with no arguments

Binary To PNG button converts all the .bin images in a folder to .png

ImageJ Macro button looks at the images generated from the binary to png button and creates a .csv with some calculations. It will try to find an imagej installation in the project folder, but if it can't, it will ask where one is. After that it will ask for the location of an imagej macro to be run

Python Macro button looks at the images generated from the binary to png button and applies multi threshold image segmentation. It will ask for the location of the ps2 collection

Generate Aggregate and Fluorescence button generates {foldername}_aggregated.csv and {foldername}_fluorescence.csv for a collection of ps2 images. Each subfolder in a collection must have the ..._metadata.json file and the {foldername}.csv to be processed.

2019-08-27/
    2019-08-27__00-00-09-654/
        ..._metadata.json
        2019-08-27__00-00-09-654.csv

    2019-08-27__00-00-52-305/
        .._metadata.json
        2019-08-27__00-00-52-305.csv

    2019-08-27__00-01-34-971/
        .._metadata.json
        2019-08-27__00-01-34-971.csv

    2019-08-27_aggregated.csv
    2019-08-27_fluorescence.csv

the associate_plots.py file will create either a .json or a .csv associating images to their plots based on Plot boundaries.xlsx.

python associate_plots.py
    -f <filepath>
    -t <type. csv or json>
    -xo <x_offset> (optional. 0 by default)
    -yo <y_offset> (optional. 0 by default)

To generate {foldername}_plot_xyz.{json/csv}, run python associate_plots.py with no arguments. This will open a gui that will let you pick between a json and a csv file. After choosing the output filetype, press one of the buttons to open a file dialog. You can then choose a directory that contains images you want to be associated. This will look in each of the folders in a directory find the _metadata.json, and create either a csv or json file with a list of each .bat or raw image for that plot. If the direcory cannot be associated to a plot, the plot number will be -1.