Skip to content

Commit

Permalink
[RFR] updated ezBIDS_core
Browse files Browse the repository at this point in the history
  • Loading branch information
Dan Levitas committed Aug 11, 2023
1 parent 3f35e88 commit 764ff85
Show file tree
Hide file tree
Showing 4 changed files with 1,298 additions and 948 deletions.
173 changes: 79 additions & 94 deletions handler/ezBIDS_core/createThumbnailsMovies.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,6 @@

import os
import sys
import json
import shutil
import numpy as np
import pandas as pd
Expand All @@ -22,61 +21,63 @@
import matplotlib.pyplot as plt
plt.style.use('dark_background')

os.environ[ 'MPLCONFIGDIR' ] = os.getcwd() + "/configs/"
os.environ['MPLCONFIGDIR'] = os.getcwd() + "/configs/"


###### functions #########
def create_movie_thumbnails(nifti_file, output_dir, object_img_array, v):
"""
Generates a PNG for each volume of a 4D acquisition.
# Functions

Parameters
----------

nifti_file : string
path of 4D nifti file.
# def create_movie_thumbnails(nifti_file, output_dir, object_img_array, v):
# """
# Generates a PNG for each volume of a 4D acquisition.

output_dir: string
path of folder where movie PNG files will be stored
# Parameters
# ----------

object_img_array: numpy.darray
result of nib.load(nifti_file).dataobj
# nifti_file : string
# path of 4D nifti file.

v: int
volume index
"""
# output_dir: string
# path of folder where movie PNG files will be stored

max_len = len(str(object_img_array.shape[3]))
# object_img_array: numpy.darray
# result of nib.load(nifti_file).dataobj

if not os.path.isfile("{}/{}.png".format(output_dir, v)):
# v: int
# volume index
# """

slice_x = object_img_array[floor(object_img_array.shape[0]/2), :, :, v]
slice_y = object_img_array[:, floor(object_img_array.shape[1]/2), :, v]
slice_z = object_img_array[:, :, floor(object_img_array.shape[2]/2), v]
# max_len = len(str(object_img_array.shape[3]))

fig, axes = plt.subplots(1, 3, figsize=(9, 3))
for index, slices in enumerate([slice_x, slice_y, slice_z]):
axes[index].imshow(slices.T, cmap="gray", origin="lower", aspect="auto")
axes[index].axis("off")
plt.tight_layout(pad=0, w_pad=0, h_pad=0)
plt.close()
# if not os.path.isfile("{}/{}.png".format(output_dir, v)):

fig.canvas.draw()
# slice_x = object_img_array[floor(object_img_array.shape[0] / 2), :, :, v]
# slice_y = object_img_array[:, floor(object_img_array.shape[1] / 2), :, v]
# slice_z = object_img_array[:, :, floor(object_img_array.shape[2] / 2), v]

w,h = fig.canvas.get_width_height()
buf = np.frombuffer(fig.canvas.tostring_argb(), dtype=np.uint8)
buf.shape = (w,h,4)
# fig, axes = plt.subplots(1, 3, figsize=(9, 3))
# for index, slices in enumerate([slice_x, slice_y, slice_z]):
# axes[index].imshow(slices.T, cmap="gray", origin="lower", aspect="auto")
# axes[index].axis("off")
# plt.tight_layout(pad=0, w_pad=0, h_pad=0)
# plt.close()

# fig.canvas.draw()

# w, h = fig.canvas.get_width_height()
# buf = np.frombuffer(fig.canvas.tostring_argb(), dtype=np.uint8)
# buf.shape = (w, h, 4)

buf = np.roll(buf,3,axis=2)
# buf = np.roll(buf, 3, axis=2)

w,h,d = buf.shape
png = Image.frombytes("RGBA", (w,h), buf.tobytes())
# w, h, d = buf.shape
# png = Image.frombytes("RGBA", (w, h), buf.tobytes())

# Sort files in Linux-friendly way (i.e. zero pad)
v_len = len(str(v))
new_v = "0"*(max_len - v_len) + str(v)
# # Sort files in UNIX-friendly way (i.e. zero pad)
# v_len = len(str(v))
# new_v = "0" * (max_len - v_len) + str(v)

png.save("{}/{}.png".format(output_dir,new_v))
# png.save("{}/{}.png".format(output_dir, new_v))


def create_thumbnail(nifti_file, image):
Expand All @@ -100,9 +101,9 @@ def create_thumbnail(nifti_file, image):

output_file = nifti_file.split(".nii.gz")[0] + ".png"

slice_x = object_img_array[floor(object_img_array.shape[0]/2), :, :]
slice_y = object_img_array[:, floor(object_img_array.shape[1]/2), :]
slice_z = object_img_array[:, :, floor(object_img_array.shape[2]/2)]
slice_x = object_img_array[floor(object_img_array.shape[0] / 2), :, :]
slice_y = object_img_array[:, floor(object_img_array.shape[1] / 2), :]
slice_z = object_img_array[:, :, floor(object_img_array.shape[2] / 2)]

fig, axes = plt.subplots(1, 3, figsize=(9, 3))
for index, slices in enumerate([slice_x, slice_y, slice_z]):
Expand All @@ -113,14 +114,14 @@ def create_thumbnail(nifti_file, image):

fig.canvas.draw()

w,h = fig.canvas.get_width_height()
w, h = fig.canvas.get_width_height()
buf = np.frombuffer(fig.canvas.tostring_argb(), dtype=np.uint8)
buf.shape = (w,h,4)
buf.shape = (w, h, 4)

buf = np.roll(buf,3,axis=2)
buf = np.roll(buf, 3, axis=2)

w,h,d = buf.shape
png = Image.frombytes("RGBA", (w,h), buf.tobytes())
w, h, d = buf.shape
png = Image.frombytes("RGBA", (w, h), buf.tobytes())
png.save(output_file)


Expand All @@ -142,17 +143,16 @@ def create_DWIshell_thumbnails(nifti_file, image, bval_file):
"""

output_file = nifti_file.split(".nii.gz")[0] + ".png"
pngPaths = []
bvals = [floor(float(x)) for x in pd.read_csv(bval_file, delim_whitespace=True).columns.tolist()]
modified_bvals = [round(x, -2) for x in bvals]
unique_bvals = np.unique(modified_bvals).tolist()

for bval in unique_bvals:
object_img_array = image.dataobj[..., modified_bvals.index(bval)]

slice_x = object_img_array[floor(object_img_array.shape[0]/2), :, :]
slice_y = object_img_array[:, floor(object_img_array.shape[1]/2), :]
slice_z = object_img_array[:, :, floor(object_img_array.shape[2]/2)]
slice_x = object_img_array[floor(object_img_array.shape[0] / 2), :, :]
slice_y = object_img_array[:, floor(object_img_array.shape[1] / 2), :]
slice_z = object_img_array[:, :, floor(object_img_array.shape[2] / 2)]

fig, axes = plt.subplots(1, 3, figsize=(9, 3))
for index, slices in enumerate([slice_x, slice_y, slice_z]):
Expand All @@ -163,14 +163,14 @@ def create_DWIshell_thumbnails(nifti_file, image, bval_file):

fig.canvas.draw()

w,h = fig.canvas.get_width_height()
w, h = fig.canvas.get_width_height()
buf = np.frombuffer(fig.canvas.tostring_argb(), dtype=np.uint8)
buf.shape = (w,h,4)
buf.shape = (w, h, 4)

buf = np.roll(buf,3,axis=2)
buf = np.roll(buf, 3, axis=2)

w,h,d = buf.shape
png = Image.frombytes("RGBA", (w,h), buf.tobytes())
w, h, d = buf.shape
png = Image.frombytes("RGBA", (w, h), buf.tobytes())
png.save("{}_shell-{}.png".format(output_file, bval))


Expand All @@ -186,66 +186,51 @@ def create_DWIshell_thumbnails(nifti_file, image, bval_file):
print(nifti_file)
print("")

if not os.path.isfile("{}/{}".format(data_dir, nifti_file)): # no corresponding nifti, so don't process
if not os.path.isfile("{}/{}".format(data_dir, nifti_file)): # no corresponding nifti, so don't process
print("{} does not have a corresponding NIfTI file, cannot process".format(json_file))
# nifti_file = "N/A"
else:
output_dir = nifti_file.split(".nii.gz")[0]
image = nib.load(nifti_file)

if len([x for x in image.shape if x < 0]): # image has negative dimension(s), cannot process
if len([x for x in image.shape if x < 0]): # image has negative dimension(s), cannot process
print("{} has negative dimension(s), cannot process".format(nifti_file))
# nifti_file = "N/A"
else:
object_img_array = image.dataobj[:]

bval_file = json_file.split(".json")[0].split("./")[-1] + ".bval"
if not os.path.isfile("{}/{}".format(data_dir, bval_file)):
bval_file = "N/A"
# if image.get_data_dtype() == [('R', 'u1'), ('G', 'u1'), ('B', 'u1')]:
if image.get_data_dtype() not in ["<i2", "<u2", "<f4", "int16", "uint16"]:
# Likely non-imaging acquisition. Example: "facMapReg" sequences in NYU_Shanghai dataset
print(
f"{nifti_file} doesn't appear to be an"
"imaging acquisition and therefore will "
"not be converted to BIDS. Please modify "
"if incorrect."
)
else:
bvals = [x.split(" ") for x in pd.read_csv(bval_file).columns.tolist()][0]
bvals = [floor(float(x)) for x in bvals if type(x) != str]

if len(bvals) <= 1: # just b0, so unhelpful
bval_file = "N/A"
# object_img_array = image.dataobj[:]

bval_file = json_file.split(".json")[0].split("./")[-1] + ".bval"
if not os.path.isfile("{}/{}".format(data_dir, bval_file)):
bval_file = "N/A"
else:
bvals = [x.split(" ") for x in pd.read_csv(bval_file).columns.tolist()][0]
bvals = [floor(float(x)) for x in bvals if not isinstance(x, str)]

# # create movie of each volume in 4D acquisition
# if image.ndim == 4 and nifti_file != "N/A": # 4D
# print("")
# print("Creating movie of all volumes for {}".format(nifti_file))
# print("")
# max_len = len(str(object_img_array.shape[3]))

# if not os.path.isdir(output_dir):
# os.mkdir(output_dir)

# for v in range(object_img_array.shape[3]):
# create_movie_thumbnails(nifti_file, output_dir, object_img_array, v)
# # Parallel(n_jobs=6)(delayed(create_movie)(nifti_file=nifti_file, output_dir=output_dir, object_img_array=object_img_array, v=v) for v in range(object_img_array.shape[3]))

# # Combine volume PNGs into movie
# os.system("ffmpeg -framerate 30 -pattern_type glob -i {}/'*.png' {}_movie.mp4".format(output_dir, output_dir))

if len(bvals) <= 1: # just b0, so unhelpful
bval_file = "N/A"

# Some acquisitions may not have a proper dtype, strongly indicating a non-MRI acquisition
if image.get_data_dtype() == [('R', 'u1'), ('G', 'u1'), ('B', 'u1')]:
print("{} doesn't appear to be an MRI acquisition, has dtype value of {}".format(nifti_file, [('R', 'u1'), ('G', 'u1'), ('B', 'u1')]))
else:
# create thumbnail
# Create thumbnail
if nifti_file != "N/A":
print("")
print("Creating thumbnail for {}".format(nifti_file))
print("")
create_thumbnail(nifti_file, image)

# create thumbnail of each DWI's unique shell
# Create thumbnail of each DWI's unique shell
if bval_file != "N/A":
print("")
print("Creating thumbnail(s) for each DWI shell in {}".format(nifti_file))
print("")
create_DWIshell_thumbnails(nifti_file, image, bval_file)

# remove the folder containing the PNGs for movie generation; don't need them anymore
# Remove the folder containing the PNGs for movie generation; don't need them anymore
if os.path.isdir(output_dir):
shutil.rmtree(output_dir)
Loading

0 comments on commit 764ff85

Please sign in to comment.