Format BaristaSeq

This script contains TileFetchers and FetchedTile objects to convert data donated by the Zador lab to SpaceTx Format. This is a good basic example of converting single-plane tiffs with multiple z-tiles.

The data is from one field of view consisting of primary images and accompanying Nissl-stained images. The primary images have 3 rounds, 4 channels, and 17 z-tiles. The Nissl images have 1 round, 1 channel, and 17 z-tiles. Image tiles are all single-plane TIFFS with shape (800, 1000).

input data structure:

└── parent
    ├── nissl
        ├── T00001C05Z001.tif
        ├── T00001C05Z002.tif
        ├── T00001C05Z003.tif
        ├── ...
    └── primary
        ├── r0
            ├── T00001C01Z001.tif
            ├── T00001C01Z002.tif
            ├── T00001C01Z003.tif
            ├── ...
        ├── r1
            ├── alignedT00001C01Z001.tif
            ├── alignedT00001C01Z002.tif
            ├── alignedT00001C01Z003.tif
            ├── ...
        └── r2
            ├── alignedT00001C01Z001.tif
            ├── alignedT00001C01Z002.tif
            ├── alignedT00001C01Z003.tif
            ├── ...

The locations of the data files for use with this script can be found in the docstring for format_data.

import os
import shutil
from typing import Mapping, Union

import click
import numpy as np
from skimage.io import imread
from slicedimage import ImageFormat

from starfish.experiment.builder import FetchedTile, TileFetcher, write_experiment_json
from starfish.types import Axes, Coordinates, CoordinateValue

DEFAULT_TILE_SHAPE = {Axes.Y: 1000, Axes.X: 800}


class BaristaSeqTile(FetchedTile):
    def __init__(self, file_path):
        self.file_path = file_path

    @property
    def shape(self) -> Mapping[Axes, int]:
        return DEFAULT_TILE_SHAPE

    @property
    def coordinates(self) -> Mapping[Union[str, Coordinates], CoordinateValue]:
        # these are dummy coordinates
        return {
            Coordinates.X: (0.0, 0.0001),
            Coordinates.Y: (0.0, 0.0001),
            Coordinates.Z: (0.0, 0.0001),
        }

    @property
    def format(self) -> ImageFormat:
        return ImageFormat.TIFF

    def tile_data(self) -> np.ndarray:
        return imread(self.file_path)


class BaristaSeqTileFetcher(TileFetcher):
    def __init__(self, input_dir) -> None:
        self.input_dir = input_dir

    def get_tile(self, fov_id: int, hyb: int, ch_label: int, zplane_label: int) -> FetchedTile:
        subdir = "primary"
        round_dir = f"r{hyb}"
        if hyb == 0:
            filename = f"T{fov_id+1:05}C{ch_label+1:02}Z{zplane_label+1:03}.tif"
        else:
            filename = f"alignedT{fov_id+1:05}C{ch_label+1:02}Z{zplane_label+1:03}.tif"
        file_path = os.path.join(self.input_dir, subdir, round_dir, filename)
        return BaristaSeqTile(file_path)


class BaristaSeqNucleiTileFetcher(TileFetcher):
    def __init__(self, input_dir, aux_type) -> None:
        self.input_dir = input_dir

    def get_tile(self, fov_id: int, hyb: int, ch_label: int, zplane_label: int) -> FetchedTile:
        subdir = "nissl"
        filename = f"T00001C05Z{zplane_label+1:03}.tif"
        file_path = os.path.join(self.input_dir, subdir, filename)

        return BaristaSeqTile(file_path)


@click.command()
@click.option("--input-dir", type=str, required=True, help="input directory containing images")
@click.option("--output-dir", type=str, required=True, help="output directory for formatted data")
def format_data(input_dir, output_dir) -> None:
    """Format a BaristaSeq Tile

    Parameters
    ----------
    input_dir : str
        Input directory containing data. Example data for a single FoV can be downloaded from
        s3://spacetx.starfish.data.public/browse/raw/20181231/barista-seq-mouse-cortex-cropped
    output_dir : str
        Output directory containing formatted data in SpaceTx format. Example output data can be
        downloaded from
        https://d2nhj9g34unfro.cloudfront.net/browse/formatted/20181028/ \
        BaristaSeq/cropped_formatted/experiment.json"
    """

    num_fovs = 1

    primary_image_dimensions: Mapping[Union[str, Axes], int] = {
        Axes.ROUND: 3,
        Axes.CH: 4,
        Axes.ZPLANE: 17,
    }

    aux_name_to_dimensions: Mapping[str, Mapping[Union[str, Axes], int]] = {
        "nuclei": {
            Axes.ROUND: 1,
            Axes.CH: 1,
            Axes.ZPLANE: 17,
        }
    }

    os.makedirs(output_dir, exist_ok=True)

    write_experiment_json(
        path=output_dir,
        fov_count=num_fovs,
        primary_image_dimensions=primary_image_dimensions,
        aux_name_to_dimensions=aux_name_to_dimensions,
        primary_tile_fetcher=BaristaSeqTileFetcher(input_dir),
        aux_tile_fetcher={
            "nuclei": BaristaSeqNucleiTileFetcher(input_dir, "nuclei"),
        },
        tile_format=ImageFormat.TIFF,
        default_shape=DEFAULT_TILE_SHAPE
    )

    shutil.copyfile(
        src=os.path.join(input_dir, "codebook.json"),
        dst=os.path.join(output_dir, "codebook.json")
    )


if __name__ == "__main__":
    format_data()