Join us on GitHub

Example Stitching Notebook

[1]:

import os
os.environ['JAVA_HOME'] = "/Users/sophia/mambaforge/envs/scPortrait/lib/jvm"
import matplotlib.pyplot as plt
from scportrait.tools.stitch import Stitcher, ParallelStitcher

/Users/sophia/mambaforge/envs/scPortrait/lib/python3.10/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html
  from .autonotebook import tqdm as notebook_tqdm

Single-threaded Stitching

initializing the stitcher object

[2]:

input_dir = os.path.join("../../../../example_data/example_stitching/raw_image_tiles")
slidename = "stitching_test"
outdir = os.path.abspath(os.path.join("../../../../example_data/example_stitching/", slidename))

row = str(2).zfill(2) #specify the row of the well you want to stitch
well = str(4).zfill(2) #specifc the well number you wish to stitch
zstack_value = str(1).zfill(3) #specify the zstack you want to stitch. for multiple zstacks please make a loop and iterate through each of them.
timepoint = str(1).zfill(3) #specifz the timepoint you wish to stitch

pattern = f"Timepoint{timepoint}_Row{row}_Well{well}_{{channel}}_zstack{zstack_value}_r{{row:03}}_c{{col:03}}.tif"

#initialize stitcher
stitcher = Stitcher(os.path.abspath(input_dir),
                    slidename,
                    outdir,
                    pattern =  pattern,
                    stitching_channel = "Alexa488",
                    overlap= 0.1,
                    max_shift= 30,
                    filter_sigma = 0,
                    rescale_range = {"Alexa488": (1, 99), "DAPI": (1, 99), "mCherry": (1, 99)},
                    overwrite=True)

Output directory at /Users/sophia/Documents/GitHub/scPortrait/example_data/example_stitching/stitching_test already exists, overwriting.

Generating thumbnails

[3]:

stitcher.generate_thumbnail()

    assembling thumbnail 9/9

[4]:

#thumbnail is saved in the stitcher object and can be accessed via stitcher.thumbnail
plt.imshow(stitcher.thumbnail)

[4]:

<matplotlib.image.AxesImage at 0xa8e0f1ff0>

../../../_images/pages_tools_stitching_example_stitching_notebook_7_1.png

[5]:

#alterantively it can be saved to a tif file
stitcher.write_thumbnail()

[6]:

stitcher.initialize_reader()

Generating full-scale stitched image

[7]:

stitcher.stitch()

performing stitching on channel Alexa488 with id number 0
    quantifying alignment error 1000/1000
    aligning edge 12/12

WARNING:root:Folder /var/folders/35/p4c58_4n3bb0bxnzgns1t7kh0000gn/T/temp_mmap_eqb2txlq with temp mmap arrays is being deleted. All existing temp mmapp arrays will be unusable!
WARNING:root:New temp folder location. Temp mmap arrays are written to /Users/sophia/Documents/GitHub/scPortrait/example_data/example_stitching/stitching_test/temp_mmap_qx6rbrpl. Cleanup of this folder is OS dependant, and might need to be triggered manually!

Alignment complete.
assembling mosaic with shape (3, 3040, 3038)
created tempmmap array for assembled mosaic at /Users/sophia/Documents/GitHub/scPortrait/example_data/example_stitching/stitching_test/temp_mmap_qx6rbrpl/temp_mmap_8836257951842585069.hdf

  0%|          | 0/3 [00:00<?, ?it/s]

        merging tile 9/9

 33%|███▎      | 1/3 [00:01<00:02,  1.07s/it]


        merging tile 9/9

 67%|██████▋   | 2/3 [00:02<00:01,  1.07s/it]


        merging tile 9/9

100%|██████████| 3/3 [00:03<00:00,  1.08s/it]

../../../_images/pages_tools_stitching_example_stitching_notebook_11_12.png

../../../_images/pages_tools_stitching_example_stitching_notebook_11_13.png

[8]:

#the stitched image is saved in the stitcher object and can be accessed via stitcher.assembled_mosaic
stitcher.assembled_mosaic

[8]:

	Array	Chunk
Bytes	52.85 MiB	52.85 MiB
Shape	(3, 3040, 3038)	(3, 3040, 3038)
Dask graph	1 chunks in 2 graph layers
Data type	uint16 numpy.ndarray

[9]:

# the stitched image can then be written to a variety of output formats

stitcher.write_tif(export_xml=True)
stitcher.write_ome_zarr()

[10]:

import matplotlib.pyplot as plt
from tifffile import imread

fig, axs = plt.subplots(1, 3, figsize = (30, 10))
axs[0].imshow(imread(f"{outdir}/stitching_test_Alexa488.tif"))
axs[0].axis("off")

axs[1].imshow(imread(f"{outdir}/stitching_test_DAPI.tif"))
axs[1].axis("off")

axs[2].imshow(imread(f"{outdir}/stitching_test_mCherry.tif"))
axs[2].axis("off")

[10]:

(-0.5, 3037.5, 3039.5, -0.5)

../../../_images/pages_tools_stitching_example_stitching_notebook_14_1.png

[11]:

del stitcher

Multi-threaded Stitching

Using the ParallelStitcher class stitching can be speed up by using multiple threads. The code to perform stitching remains more or less the same.

initializing the stitcher object

[12]:

input_dir = os.path.join("../../../../example_data/example_stitching/raw_image_tiles")
slidename = "stitching_test_parallel"
outdir_parallel = os.path.abspath(os.path.join("../../../../example_data/example_stitching/", slidename))

row = str(2).zfill(2) #specify the row of the well you want to stitch
well = str(4).zfill(2) #specifc the well number you wish to stitch
zstack_value = str(1).zfill(3) #specify the zstack you want to stitch. for multiple zstacks please make a loop and iterate through each of them.
timepoint = str(1).zfill(3) #specifz the timepoint you wish to stitch

pattern = f"Timepoint{timepoint}_Row{row}_Well{well}_{{channel}}_zstack{zstack_value}_r{{row:03}}_c{{col:03}}.tif"

#initialize stitcher
stitcher = ParallelStitcher(input_dir,
                    slidename,
                    outdir_parallel,
                    pattern =  pattern,
                    stitching_channel = "Alexa488",
                    overlap= 0.1,
                    max_shift= 30,
                    filter_sigma = 0,
                    rescale_range = {"Alexa488": (1, 99), "DAPI": (1, 99), "mCherry": (1, 99)},
                    overwrite=True,
                    threads = 12)

Output directory at /Users/sophia/Documents/GitHub/scPortrait/example_data/example_stitching/stitching_test_parallel already exists, overwriting.

Generating thumbnails

[13]:

stitcher.generate_thumbnail()

    assembling thumbnail 1/9    assembling thumbnail 9/9

[14]:

#thumbnail is saved in the stitcher object and can be accessed via stitcher.thumbnail
plt.imshow(stitcher.thumbnail)

[14]:

<matplotlib.image.AxesImage at 0xa9320ec20>

../../../_images/pages_tools_stitching_example_stitching_notebook_21_1.png

[15]:

#alterantively it can be saved to a tif file
stitcher.write_thumbnail()

Generating full-scale stitched image

[16]:

stitcher.stitch()

performing stitching on channel Alexa488 with id number 0
networkx
HERE
    quantifying alignment error: 100%|██████████| 1000/1000 [00:00<00:00, 1187.36it/s]
                  aligning edge: 100%|██████████| 12/12 [00:00<00:00, 92.61it/s]

WARNING:root:Folder /Users/sophia/Documents/GitHub/scPortrait/example_data/example_stitching/stitching_test/temp_mmap_qx6rbrpl with temp mmap arrays is being deleted. All existing temp mmapp arrays will be unusable!
WARNING:root:New temp folder location. Temp mmap arrays are written to /Users/sophia/Documents/GitHub/scPortrait/example_data/example_stitching/stitching_test_parallel/temp_mmap_urbkr3dy. Cleanup of this folder is OS dependant, and might need to be triggered manually!

Alignment complete.
assembling mosaic with shape (3, 3040, 3038)
created tempmmap array for assembled mosaic at /Users/sophia/Documents/GitHub/scPortrait/example_data/example_stitching/stitching_test_parallel/temp_mmap_urbkr3dy/temp_mmap_2664334953454325339.hdf
assembling channels with 3 workers
assembling mosaic:   0%|          | 0/3 [00:00<?, ?it/s]



























assembling channel 0: 100%|██████████| 9/9 [00:01<00:00,  7.81it/s]


assembling channel 2: 100%|██████████| 9/9 [00:01<00:00,  7.79it/s]
assembling channel 1: 100%|██████████| 9/9 [00:01<00:00,  7.78it/s]
assembling mosaic: 100%|██████████| 3/3 [00:01<00:00,  2.59it/s]

../../../_images/pages_tools_stitching_example_stitching_notebook_24_3.png

../../../_images/pages_tools_stitching_example_stitching_notebook_24_4.png

[17]:

#the stitched image is saved in the stitcher object and can be accessed via stitcher.assembled_mosaic
stitcher.assembled_mosaic

[17]:

	Array	Chunk
Bytes	52.85 MiB	52.85 MiB
Shape	(3, 3040, 3038)	(3, 3040, 3038)
Dask graph	1 chunks in 2 graph layers
Data type	uint16 numpy.ndarray

[18]:

# the stitched image can then be written to a variety of output formats

stitcher.write_tif(export_xml=True)

#alternatively the stitched images can also be written out to tifs in a multi-threaded format
#speedups here are limited by write speed to disk, the number of threads is limited by the number of channels available
stitcher.write_tif_parallel(export_xml=True)

stitcher.write_ome_zarr()

writing tif files: 100%|██████████| 3/3 [00:00<00:00, 212.63it/s]

Visualize Stitching Output

[19]:

import matplotlib.pyplot as plt
from tifffile import imread

fig, axs = plt.subplots(1, 3, figsize = (30, 10))
axs[0].imshow(imread(f"{outdir_parallel}/stitching_test_parallel_Alexa488.tif"))
axs[0].axis("off")

axs[1].imshow(imread(f"{outdir_parallel}/stitching_test_parallel_DAPI.tif"))
axs[1].axis("off")

axs[2].imshow(imread(f"{outdir_parallel}/stitching_test_parallel_mCherry.tif"))
axs[2].axis("off")

[19]:

(-0.5, 3037.5, 3039.5, -0.5)

../../../_images/pages_tools_stitching_example_stitching_notebook_28_1.png

[20]:

# compare parallel and non-parallel stitching

imread(f"{outdir_parallel}/stitching_test_parallel_Alexa488.tif") == imread(f"{outdir}/stitching_test_Alexa488.tif")

[20]:

array([[ True,  True,  True, ...,  True,  True,  True],
       [ True,  True,  True, ...,  True,  True,  True],
       [ True,  True,  True, ...,  True,  True,  True],
       ...,
       [ True,  True,  True, ...,  True,  True,  True],
       [ True,  True,  True, ...,  True,  True,  True],
       [ True,  True,  True, ...,  True,  True,  True]])

[21]:

del stitcher

[ ]:

[ ]: