Push for debugging

Author: BrianMichell

src/mdio/segy/_disaster_recovery_wrapper.py

@@ -28,7 +28,7 @@ def header(self):
 
     @property
     def raw_header(self):
-        return np.ascontiguousarray(self.traces.header).view("|V240")
+        return np.ascontiguousarray(self.traces.header.copy()).view("|V240")
 
     @property
     def sample(self):

src/mdio/segy/_workers.py

@@ -139,14 +139,20 @@ def trace_worker(  # noqa: PLR0913
     # For that reason, we have wrapped the accessors to provide an interface that can be removed
     # and not require additional changes to the below code.
     # NOTE: The `raw_header_key` code block should be removed in full as it will become dead code.
-    traces = SegyFileTraceDataWrapper(segy_file, live_trace_indexes)
+    # traces = SegyFileTraceDataWrapper(segy_file, live_trace_indexes)
+    from copy import deepcopy
+    header_pipeline = deepcopy(segy_file.accessors.header_decode_pipeline)
+    segy_file.accessors.header_decode_pipeline.transforms = []
+    traces = segy_file.trace[live_trace_indexes]
+
     ds_to_write = dataset[worker_variables]
 
     if header_key in worker_variables:
         # TODO(BrianMichell): Implement this better so that we can enable fill values without changing the code
         # https://github.com/TGSAI/mdio-python/issues/584
         tmp_headers = np.zeros_like(dataset[header_key])
-        tmp_headers[not_null] = traces.header
+        # tmp_headers[not_null] = traces.header
+        tmp_headers[not_null] = header_pipeline.apply(traces.header.copy())
         # Create a new Variable object to avoid copying the temporary array
         # The ideal solution is to use `ds_to_write[header_key][:] = tmp_headers`
         # but Xarray appears to be copying memory instead of doing direct assignment.
@@ -160,7 +166,15 @@ def trace_worker(  # noqa: PLR0913
         )
     if raw_header_key in worker_variables:
         tmp_raw_headers = np.zeros_like(dataset[raw_header_key])
-        tmp_raw_headers[not_null] = traces.raw_header
+        # tmp_raw_headers = np.zeros((traces.header.shape[0], 240), dtype=np.)
+        # tmp_raw_headers[not_null] = traces.raw_header
+        np.set_printoptions(threshold=np.inf)
+        # print(traces.header.copy().view("|V240"))
+        # print(len(traces.header.view("|V240")))
+        print(len(traces.header))
+        tmp_raw_headers[not_null] = np.ascontiguousarray(traces.header.copy()).view("|V240")
+        # tmp_raw_headers[not_null] = traces.header.view("|V240")
+        # tmp_raw_headers[not_null] = np.ascontiguousarray(traces.header.copy()).view("|V240")
 
         ds_to_write[raw_header_key] = Variable(
             ds_to_write[raw_header_key].dims,

tests/conftest.py

@@ -8,7 +8,7 @@
 
 import pytest
 
-DEBUG_MODE = False
+DEBUG_MODE = True
 
 # Suppress Dask's chunk balancing warning
 warnings.filterwarnings(

tests/integration/test_segy_import_export_masked.py

@@ -547,9 +547,13 @@ def read_segy_trace_header(trace_index: int) -> bytes:
         segy_trace_idx = 0
         flat_mask = trace_mask.ravel()
         flat_raw_headers = raw_headers_data.ravel()  # Flatten to 1D array of 240-byte header records
+        print(f"Flat mask shape: {flat_mask.shape}")
+
+        operation = 'w'
 
         for grid_idx in range(flat_mask.size):
             if not flat_mask[grid_idx]:
+                print(f"Skipping trace {grid_idx} because it is masked")
                 continue
 
             # Get MDIO header as bytes - convert single header record to bytes
@@ -561,5 +565,58 @@ def read_segy_trace_header(trace_index: int) -> bytes:
             segy_header_bytes = np.frombuffer(segy_raw_header_bytes, dtype=np.uint8)
 
             # Compare byte-by-byte
-            assert_array_equal(mdio_header_bytes, segy_header_bytes)
+            # Write hexdumps to separate files for analysis
+            def hexdump_to_string(data: bytes, title: str) -> str:
+                """Create hexdump string."""
+                lines = [f"{title}", "=" * len(title), ""]
+                 
+                for i in range(0, len(data), 16):
+                    # Address
+                    addr = i
+                    hex_part = ""
+                    ascii_part = ""
+                     
+                    # Process 16 bytes at a time
+                    for j in range(16):
+                        if i + j < len(data):
+                            byte_val = data[i + j]
+                            hex_part += f"{byte_val:02x} "
+                            ascii_part += chr(byte_val) if 32 <= byte_val <= 126 else "."
+                        else:
+                            hex_part += "   "
+                            ascii_part += " "
+                     
+                    lines.append(f"{addr:08x}: {hex_part} |{ascii_part}|")
+                 
+                return "\n".join(lines)
+             
+            # Generate filenames for this test case
+            segy_filename = f"segy_headers_{grid_conf.name}.txt"
+            mdio_filename = f"mdio_headers_{grid_conf.name}.txt"
+             
+            # Append SEG-Y hexdump to file
+            with open(segy_filename, operation) as f:
+                if segy_trace_idx == 0:
+                    f.write("")  # Start fresh for first trace
+                else:
+                    f.write("\n\n")  # Add spacing between traces
+                f.write(hexdump_to_string(segy_header_bytes, 
+                                        f"SEG-Y Header - {grid_conf.name} Trace {segy_trace_idx} (240 bytes)"))
+             
+            # Append MDIO hexdump to file  
+            with open(mdio_filename, operation) as f:
+                if segy_trace_idx == 0:
+                    f.write("")  # Start fresh for first trace
+                else:
+                    f.write("\n\n")  # Add spacing between traces
+                f.write(hexdump_to_string(mdio_header_bytes,
+                                        f"MDIO Raw Header - {grid_conf.name} Trace {segy_trace_idx} (240 bytes)"))
+            operation = 'a'
+             
+            if segy_trace_idx == 0:
+                print(f"\nHeader hexdumps being written for {grid_conf.name}:")
+                print(f"  SEG-Y: {segy_filename}")
+                print(f"  MDIO:  {mdio_filename}")
+                
+            
             segy_trace_idx += 1