Drop AFNI dependency

pySPFM/deconvolution/debiasing.py

@@ -132,26 +132,26 @@ def debiasing_block(hrf, y, estimates_matrix, dist=2, progress_bar=True, jobs=1)
     beta_out : ndarray
         Debiased activity-inducing signal obtained from estimated innovation signal.
     """
-    nscans = estimates_matrix.shape[0]
-    nvoxels = estimates_matrix.shape[1]
+    n_scans = estimates_matrix.shape[0]
+    n_voxels = estimates_matrix.shape[1]
 
     # Initiates beta matrix
-    beta_out = np.zeros((nscans, nvoxels))
+    beta_out = np.zeros((n_scans, n_voxels))
 
     LGR.info("Starting debiasing step...")
     # Performs debiasing
     if progress_bar:
         debiased = Parallel(n_jobs=jobs, backend="multiprocessing")(
             delayed(do_debias_block)(hrf, y[:, voxidx], estimates_matrix[:, voxidx])
-            for voxidx in tqdm(range(nvoxels))
+            for voxidx in tqdm(range(n_voxels))
         )
     else:
         debiased = Parallel(n_jobs=jobs, backend="multiprocessing")(
             delayed(do_debias_block)(hrf, y[:, voxidx], estimates_matrix[:, voxidx])
-            for voxidx in range(nvoxels)
+            for voxidx in range(n_voxels)
         )
 
-    for vox_idx in range(nvoxels):
+    for vox_idx in range(n_voxels):
         beta_out[:, vox_idx] = debiased[vox_idx]
 
     LGR.info("Debiasing step finished")

pySPFM/deconvolution/fista.py

@@ -129,11 +129,11 @@ def fista(
         Selected regularization parameter lambda
     """
     if len(y.shape) == 1:
-        nvoxels = 1
+        n_voxels = 1
         y = y[:, np.newaxis]
     else:
-        nvoxels = y.shape[1]
-    nscans = hrf.shape[1]
+        n_voxels = y.shape[1]
+    n_scans = hrf.shape[1]
 
     # Select lambda
     lambda_, update_lambda, noise_estimate = select_lambda(
@@ -161,8 +161,8 @@ def fista(
             l2,
             prox,
             tau=c_ist,
-            x0=np.zeros((nscans, nvoxels)),
-            epsg=np.ones(nvoxels),
+            x0=np.zeros((n_scans, n_voxels)),
+            epsg=np.ones(n_voxels),
             niter=max_iter,
             acceleration="fista",
             show=False,
@@ -173,7 +173,7 @@ def fista(
         hrf_cov = np.dot(hrf_trans, hrf)
         v = np.dot(hrf_trans, y)
 
-        y_fista_S = np.zeros((nscans, nvoxels), dtype=np.float32)
+        y_fista_S = np.zeros((n_scans, n_voxels), dtype=np.float32)
         S = y_fista_S.copy()
 
         t_fista = 1
@@ -221,7 +221,7 @@ def fista(
 
             # Update lambda
             if update_lambda:
-                nv = np.sqrt(np.sum((np.dot(hrf, S) - y) ** 2, axis=0) / nscans)
+                nv = np.sqrt(np.sum((np.dot(hrf, S) - y) ** 2, axis=0) / n_scans)
                 if abs(nv - noise_estimate) > precision:
                     lambda_ = np.nan_to_num(lambda_ * noise_estimate / nv)