update example.

Author: paquiteau

examples/functional/example_2D+T_spiral.py

@@ -20,9 +20,9 @@
 from tqdm.auto import tqdm
 
 
-
 # %%
 import logging
+
 logging.getLogger("snake").setLevel(level=logging.DEBUG)
 
 # %%
@@ -39,14 +39,18 @@
 
 sim_conf.hardware.n_coils = 8  # Update to get multi coil results.
 sim_conf.hardware.field_strength = 7
-sim_conf.fov.res_mm = (1,1,1) # 1 mm isotropic resolution
-sim_conf.fov.size = (181,217,1) # Single slice  (in mm)
-sim_conf.fov.offset = (-90,-125, -10) # Located in the center of the cortex
-sim_conf.fov.angles = (0,0,0)
+sim_conf.fov.res_mm = (1, 1, 1)  # 1 mm isotropic resolution
+sim_conf.fov.size = (181, 217, 1)  # Single slice  (in mm)
+sim_conf.fov.offset = (-90, -125, -10)  # Located in the center of the cortex
+sim_conf.fov.angles = (0, 0, 0)
 # -
 
-phantom = Phantom.from_brainweb(sub_id=4, sim_conf=sim_conf, tissue_file="tissue_7T", output_res=1)
-phantom = phantom.resample(new_affine=sim_conf.fov.affine, new_shape=sim_conf.fov.shape, use_gpu=True) # already select the phantom here, to reduce computational burden.
+phantom = Phantom.from_brainweb(
+    sub_id=4, sim_conf=sim_conf, tissue_file="tissue_7T", output_res=1
+)
+phantom = phantom.resample(
+    new_affine=sim_conf.fov.affine, new_shape=sim_conf.fov.shape, use_gpu=True
+)  # already select the phantom here, to reduce computational burden.
 
 contrast = phantom.contrast(sim_conf=sim_conf)
 
@@ -55,21 +59,31 @@
 
 activation_handler = BlockActivationHandler(
     block_off=20,
-    block_on=20, 
-    duration=300, 
+    block_on=20,
+    duration=300,
     atlas="hardvard-oxford__cort-maxprob-thr0-1mm",
-    atlas_label = 48 # occipital cortex
+    atlas_label=48,  # occipital cortex
 )
 
 example_phantom = activation_handler.get_static(phantom.copy(), sim_conf)
 
 roi = example_phantom.masks[example_phantom.labels_idx["ROI"]]
 
-roi_resampled = apply_affine(roi,new_affine=sim_conf.fov.affine, old_affine=phantom.affine,new_shape=  sim_conf.fov.shape)
+roi_resampled = apply_affine(
+    roi,
+    new_affine=sim_conf.fov.affine,
+    old_affine=phantom.affine,
+    new_shape=sim_conf.fov.shape,
+)
 
 plt.imshow(contrast.squeeze().T, origin="lower", cmap="gray")
 plt.axis("off")
-plt.contour(roi_resampled.T.squeeze(), levels=(0.01, 0.5),origin="lower", colors=["tab:blue", "tab:orange"])
+plt.contour(
+    roi_resampled.T.squeeze(),
+    levels=(0.01, 0.5),
+    origin="lower",
+    colors=["tab:blue", "tab:orange"],
+)
 
 # +
 # %%
@@ -93,50 +107,54 @@
 
 # %%
 # Defining Acquisition model
-#--------------------------
+# --------------------------
 
 
 # Here we assume an almost infinite SNR and a T2* decay signal model
 # in the modeling of the fMRI signal
 
 engine_t2s = NufftAcquisitionEngine(
-    model="T2s",  
-    snr=30000, 
-    slice_2d=True, # Does not matter, we are in 2D anyway (: 
+    model="T2s",
+    snr=30000,
+    slice_2d=True,  # Does not matter, we are in 2D anyway (:
 )
 
 engine_t2s(
-    "example_2D+T.mrd",  
-    sampler,  
-    phantom,  
-    sim_conf, 
-    handlers=[activation_handler],  
-    worker_chunk_size=20,  
-    n_workers=1,  
-    nufft_backend="finufft", 
+    "example_2D+T.mrd",
+    sampler,
+    phantom,
+    sim_conf,
+    handlers=[activation_handler],
+    worker_chunk_size=20,
+    n_workers=1,
+    nufft_backend="finufft",
 )
 
 
-
-
 # %%
 
-from snake.toolkit.reconstructors import ZeroFilledReconstructor, SequentialReconstructor
-with NonCartesianFrameDataLoader("example_2D+T.mrd") as data_loader:
-    rec_zer = ZeroFilledReconstructor(nufft_backend="finufft", density_compensation=None).reconstruct(data_loader)
+from snake.toolkit.reconstructors import (
+    ZeroFilledReconstructor,
+    SequentialReconstructor,
+)
+
+with NonCartesianFrameDataLoader("example_2D+T.mrd", squeeze_dims=False) as data_loader:
+    rec_zer = ZeroFilledReconstructor(
+        nufft_backend="finufft", density_compensation=None
+    ).reconstruct(data_loader)
     dyn_datas = data_loader.get_all_dynamic()
     phantom_dl = data_loader.get_phantom()
     sim_conf = data_loader.get_sim_conf()
 
 
 # %%
-plt.imshow(abs(rec_zer[0, ...]), cmap="gray", origin='lower')
+plt.imshow(abs(rec_zer[0, ...]), cmap="gray", origin="lower")
 
 # %%
 
 from snake.toolkit.analysis.stats import contrast_zscore, get_scores
 
-+
+
 waveform_name = "activation-block_on"
 good_d = None
 for d in dyn_datas:
@@ -151,18 +169,9 @@
 del phantom
 del dyn_datas
 
-# -
-
 TR_vol = sim_conf.seq.TR
-z_score = contrast_zscore(
-    rec_zer, TR_vol, bold_signal, bold_sample_time, "block-on"
-)
+z_score = contrast_zscore(rec_zer, TR_vol, bold_signal, bold_sample_time, "block-on")
 stats_results = get_scores(z_score, roi_mask, 0.5)
-np.save(data_zscore_file, z_score)
-
 
-rec_zer.shape
-
-# %%
 
 # %%