Update JAX Model Based Reconstruction code

BohndiekLab · May 16, 2024 · 4a8eea7 · 4a8eea7
1 parent ab5ca6f
commit 4a8eea7
Show file tree

Hide file tree

Showing 3 changed files with 8 additions and 8 deletions.
diff --git a/patato/convenience_scripts/process_msot.py b/patato/convenience_scripts/process_msot.py
@@ -94,10 +94,10 @@ def main():
         description = pipeline.children[0].get_algorithm_name()
 
         # Setup the name
-        if args.highpass is not None:
-            description += "_CUSTOM_HP_FILTER"
-        if args.lowpass is not None:
-            description += "_CUSTOM_LP_FILTER"
+        #if args.highpass is not None:
+        #    description += "_CUSTOM_HP_FILTER"
+        #if args.lowpass is not None:
+        #    description += "_CUSTOM_LP_FILTER"
         if args.run is not None:
             description += "_run_" + str(args.run)
         if args.wavelength is not None:

diff --git a/patato/processing/jax_preprocessing_algorithm.py b/patato/processing/jax_preprocessing_algorithm.py
@@ -178,7 +178,7 @@ def pre_compute_filter(self, n_samples: int, fs: float, irf: Array = None):
         fs : float
         irf : Array
         """
-        self.filter = jnp.array(make_filter(n_samples, fs, irf, self.hilbert, self.lp_filter, self.hp_filter))
+        self.filter = jnp.array(make_filter(n_samples, fs, irf if self.irf_correct else None, self.hilbert, self.lp_filter, self.hp_filter))
 
     def _run(self, time_series: Array, detectors: Array, overall_correction_factor, **kwargs):
         """

diff --git a/patato/recon/jax_model_based/__init__.py b/patato/recon/jax_model_based/__init__.py
@@ -90,8 +90,8 @@ def __init__(self, n_pixels, field_of_view, pa_example: "PAData" = None, **kwarg
         from jax.experimental.sparse import CSR
 
         self._model_regulariser = (kwargs.get("model_regulariser", None), kwargs.get("model_regulariser_lambda", None))
-        if self._model_regulariser[0] is None:
-            self._model_regulariser = None
+        #if self._model_regulariser[0] is None:
+        #    self._model_regulariser = None
 
         self._model_matrix = CSR((self._model_matrix.data, self._model_matrix.indices, self._model_matrix.indptr),
                                  shape=self._model_matrix.shape)
@@ -153,7 +153,7 @@ def reconstruct(self, raw_data: np.ndarray,
             raise ValueError("Constraint must either be 'positive' or 'none'.")
 
         M = self._model_matrix
-        if self._model_regulariser is not None:
+        if self._model_regulariser is not None or all([x is None for x in self._model_regulariser]):
             method, lambda_reg = self._model_regulariser
         else:
             method, lambda_reg = None, None