use only the simulated mass list in simulated runs

src/depiction/calibration/perform_calibration.py

@@ -78,17 +78,17 @@ def calibrate_image(
             read_full = read_peaks
 
         logger.info("Extracting all features...")
-        all_features = self._extract_all_features(read_peaks)
+        all_features = self._extract_all_features(read_peaks).transpose("i", "c")
         self._validate_per_spectra_array(all_features, coordinates_2d=read_peaks.coordinates_2d)
         self._write_data_array(all_features, group="features_raw")
 
         logger.info("Preprocessing features...")
-        all_features = self._calibration.preprocess_image_features(all_features=all_features)
+        all_features = self._calibration.preprocess_image_features(all_features=all_features).transpose("i", "c")
         self._validate_per_spectra_array(all_features, coordinates_2d=read_peaks.coordinates_2d)
         self._write_data_array(all_features, group="features_processed")
 
         logger.info("Fitting models...")
-        model_coefs = self._fit_all_models(all_features=all_features)
+        model_coefs = self._fit_all_models(all_features=all_features).transpose("i", "c")
         self._validate_per_spectra_array(model_coefs, coordinates_2d=read_peaks.coordinates_2d)
         self._write_data_array(model_coefs, group="model_coefs")
 

src/depiction/tools/simulate/synthetic_msi_data_generator.py

@@ -10,6 +10,8 @@
 from depiction.persistence import ImzmlWriteFile, ImzmlWriter
 
 
+# TODO configure shifts
+
 class SyntheticMSIDataGenerator:
     """Helper that creates synthetic MSI data."""
 

src/depiction_targeted_preproc/example_sim/default.yml

@@ -13,9 +13,11 @@ calibration:
   max_distance_unit: ppm
   reg_model_type: linear_siegelslopes
   reg_model_unit: mz
-  input_smoothing_activated: yes
-  input_smoothing_kernel_size: 5
-  input_smoothing_kernel_std: 1.0
+  input_smoothing_activated: no
+  # TODO investigate the problem!
+  #input_smoothing_activated: yes
+  #input_smoothing_kernel_size: 5
+  #input_smoothing_kernel_std: 1.0
   min_points: 3
 n_jobs: 10
 requested_artifacts:
@@ -29,3 +31,5 @@ simulate:
   image_height: 100
   n_labels: 30
   bin_width_ppm: 80
+  target_mass_min: 850
+  target_mass_max: 1900

src/depiction_targeted_preproc/example_sim/run_sim.py

@@ -11,17 +11,17 @@
 def setup_sim_dir(path: Path) -> None:
     path.mkdir(exist_ok=True, parents=True)
     shutil.copyfile(Path(__file__).parent / "default.yml", path / "pipeline_params.yml")
-    path_source_mass_list = Path(__file__).parents[1] / "example" / "data-raw" / "mass_list_vend.csv"
-    shutil.copyfile(path_source_mass_list, path / "mass_list.raw.csv")
+    # path_source_mass_list = Path(__file__).parents[1] / "example" / "data-raw" / "mass_list_vend.csv"
+    # shutil.copyfile(path_source_mass_list, path / "mass_list.raw.csv")
 
 
 # TODO why does it not work?
 
 def main() -> None:
     dir_work = Path(__file__).parent / "data-work"
-    #dir_output = Path(__file__).parent / "data-output"
+    # dir_output = Path(__file__).parent / "data-output"
     dir_work.mkdir(exist_ok=True, parents=True)
-    #dir_output.mkdir(exist_ok=True, parents=True)
+    # dir_output.mkdir(exist_ok=True, parents=True)
 
     sample_name = "dummy01_sim"
     setup_sim_dir(dir_work / sample_name)

src/depiction_targeted_preproc/pipeline_config/model.py

@@ -85,6 +85,8 @@ class SimulateParameters(BaseModel):
     image_height: int = 100
     n_labels: int = 30
     bin_width_ppm: float = 100.0
+    target_mass_min: float = 850.
+    target_mass_max: float = 1900.
 
 
 class PipelineArtifact(str, Enum):

src/depiction_targeted_preproc/workflow/rules/rules_simulate.smk

@@ -16,20 +16,19 @@ rule simulate_create_labels:
 
 rule simulate_create_mass_list:
     input:
-        mass_list="{sample}_sim/mass_list.calibration.csv",
         config="{sample}_sim/pipeline_params.yml",
     output:
-        mass_list="{sample}_sim/mass_list.simulated.csv"
+        mass_list="{sample}_sim/mass_list.raw.csv"
     shell:
         "python -m depiction_targeted_preproc.workflow.simulate.create_mass_list"
-        " --input-mass-list-path {input.mass_list} --config-path {input.config}"
+        " --config-path {input.config}"
         " --output-mass-list-path {output.mass_list}"
 
 
 rule simulate_generate_imzml:
     input:
         image="{sample}_sim/true_labels.hdf5",
-        mass_list="{sample}_sim/mass_list.simulated.csv",
+        mass_list="{sample}_sim/mass_list.raw.csv",
         config="{sample}_sim/pipeline_params.yml",
     output:
         imzml=multiext("{sample}_sim/raw",".imzML",".ibd"),

src/depiction_targeted_preproc/workflow/simulate/create_mass_list.py

@@ -10,22 +10,16 @@
 
 
 def simulate_create_mass_list(
-    input_mass_list_path: Annotated[Path, Option()],
     config_path: Annotated[Path, Option()],
     output_mass_list_path: Annotated[Path, Option()],
 ) -> None:
     # parse the config
     config = PipelineParameters.parse_yaml(config_path)
 
-    # parse the input
-    df_input = pl.read_csv(input_mass_list_path)
-    mass_min = df_input["mass"].min()
-    mass_max = df_input["mass"].max()
-
     # compute the masses
     lambda_avg = 1.0 + 4.95e-4
     n_labels = config.simulate.n_labels
-    masses = np.linspace(mass_min, mass_max, n_labels)
+    masses = np.linspace(config.simulate.target_mass_min, config.simulate.target_mass_max, n_labels)
     masses -= masses % lambda_avg
 
     # create the output mass list