another attempt at parsing RWA - seems to work better

src/main.py

@@ -179,19 +179,20 @@ def deduplicate_instruments_from_layers(layers):
     #     "rheed_system": rheeds,
     # }
 
-def analyse_rheed_data(data, n_layers: int):
+def analyse_rheed_data(data):
     '''
     Takes the content of a tsv file and returns a dictionary with timestamps and intensities.
-    The file should contain a 2D array composed of 3N+1 columns - where N is the total number of layers in a given sample - and an unspecified number of rows.
+    The file should contain a 2D array composed of 4 columns - where the first column is a timestamp and the other three are RHEED intensities - and an unspecified number of rows.
     
     -----
-    Time    Layer1_Int1     Layer1_Int2     Layer1_Int3     (repeat...)
+    Time    Layer1_Int1     Layer1_Int2     Layer1_Int3
     -----
 
     Distinct ValueErrors are raised if:
     - The array is not 2-dimensional;
-    - The number of (intensity) columns is not a multiple of 3;
-    - The total number of columns does not equate exactly 3N+1.
+    - The total number of columns does not equate exactly 1+3 (= 4).
+
+    Time is expressed in seconds, intensities are normalized (adimensional).
 
     # TO-DO: complete this description...
     Written with help from DeepSeek.
@@ -200,29 +201,17 @@ def analyse_rheed_data(data, n_layers: int):
     if data.ndim != 2:
         raise ValueError(f"Unexpected trace format: expected 2D array, got ndim = {data.ndim}.")
     n_cols = data.shape[1] # 0 = rows, 1 = columns
-    if (n_cols - 1) % 3 != 0:
-        raise ValueError(f"Unexpected number of columns: expected 3N+1 columns, got {n_cols}.")
-    if (n_cols - 1) // 3 != n_layers:
-        exp = n_layers * 3 + 1
-        raise ValueError(f"Unexpected volume of data: found {n_layers} layers, expected {exp} (3N+1) columns, got {n_cols}.")
+    if n_cols > 4:
+        print(f"Warning! The input file (for Realtime Window Analysis) has {n_cols-4} more than needed.\nOnly 4 columns will be considered - with the first representing time and the others representing RHEED intensities.")
+    if n_cols < 4:
+        raise ValueError(f"Insufficient number of columns: expected 4, got n_cols = {n_cols}.")
     n_time_points = data.shape[0]
 
     # Get time (all rows of col 0) as Float64:
     time = data[:, 0].astype(np.float64, copy=False) # copy=False suggested by LLM for mem. eff.
 
-    # Empty 3D array for intensities:
-    intensities = np.zeros(
-        (n_layers, n_time_points, 3)
-    )
-
-    # Loop through layers:
-    for layer_index in range(n_layers):
-        layer_name = f"layer_{layer_index + 1}"
-        # Columns for this layer are from 3i+1 to 3i+3 incl. (= 3i+4 excl.)
-        start_col = 1 + layer_index * 3
-        end_col = start_col + 3 # remember this gets excluded!
-        # Get layer-specific intensities (all rows of columns start_col:end_col) as Float32:
-        intensities[layer_index, :, :] = data[:, start_col:end_col].astype(np.float32, copy=False)
+    # Get intensities (all rows of cols 1,2,3) as Float32:
+    intensities = data[:, 1:4].astype(np.float32, copy=False)
 
     return {
         "time": time,
@@ -566,5 +555,8 @@ if __name__=="__main__":
     # and merge all data in a single file to analyse it
     with open(f"tests/Realtime_Window_Analysis.txt", "r") as o:
         osc = np.loadtxt(o, delimiter="\t")
-    rheed_osc = analyse_rheed_data(data=osc, n_layers=n_layers) or None # analyze rheed data first, build the file later
+    try:
+        rheed_osc = analyse_rheed_data(data=osc) or None # analyze rheed data first, build the file later
+    except ValueError as ve:
+        raise ValueError(f"Error with function analyse_rheed_data. {ve}\nPlease make sure the Realtime Window Analysis file is exactly 4 columns wide - where the first column represents time and the others are RHEED intensities.")
     build_nexus_file(result, output_path=f"output/sample-{sample_name}-nexus.h5", rheed_osc=rheed_osc)