Remove raw field additions. Depends on segy >= 0.5.1

Author: BrianMichell

src/mdio/converters/segy.py

@@ -4,15 +4,12 @@
 
 import logging
 import os
-from copy import deepcopy
 from typing import TYPE_CHECKING
 
 import numpy as np
 import zarr
 from segy import SegyFile
 from segy.config import SegySettings
-from segy.schema import HeaderField
-from segy.schema import ScalarType as ScalarType2
 from segy.standards.codes import MeasurementSystem as segy_MeasurementSystem
 from segy.standards.fields.trace import Rev0 as TraceHeaderFieldsRev0
 
@@ -344,33 +341,6 @@ def _add_grid_override_to_metadata(dataset: Dataset, grid_overrides: dict[str, A
         dataset.metadata.attributes["gridOverrides"] = grid_overrides
 
 
-def _scalar_to_size(scalar: ScalarType2) -> int:
-    # TODO(BrianMichell): #0000 Lazy way to support conversion.
-    if scalar == ScalarType2.STRING8:
-        return 8
-    try:
-        return int(str(scalar)[-2:]) // 8
-    except ValueError:
-        return 1
-
-
-def _customize_segy_spec(segy_spec: SegySpec) -> SegySpec:
-    assigned_bytes = []
-
-    ret = deepcopy(segy_spec)
-
-    for field in segy_spec.trace.header.fields:
-        byte = field.byte - 1
-        for i in range(byte, byte + _scalar_to_size(field.format)):
-            assigned_bytes.append(i)  # noqa: PERF402
-    unassigned_bytes = [i for i in range(240) if i not in assigned_bytes]
-    field_to_customize = [
-        HeaderField(name=f"__MDIO_RAW_UNSPECIFIED_Field_{i}", format=ScalarType.UINT8, byte=i + 1)
-        for i in unassigned_bytes
-    ]
-    return ret.customize(trace_header_fields=field_to_customize)
-
-
 def _add_raw_headers_to_template(mdio_template: AbstractDatasetTemplate) -> AbstractDatasetTemplate:
     """Add raw headers capability to the MDIO template by monkey-patching its _add_variables method.
 
@@ -451,9 +421,6 @@ def segy_to_mdio(  # noqa PLR0913
     input_path = _normalize_path(input_path)
     output_path = _normalize_path(output_path)
 
-    if os.getenv("MDIO__DO_RAW_HEADERS") == "1":
-        segy_spec = _customize_segy_spec(segy_spec)
-
     if not overwrite and output_path.exists():
         err = f"Output location '{output_path.as_posix()}' exists. Set `overwrite=True` if intended."
         raise FileExistsError(err)

src/mdio/segy/creation.py

@@ -28,38 +28,6 @@
 logger = logging.getLogger(__name__)
 
 
-def _filter_raw_unspecified_fields(headers: NDArray) -> NDArray:
-    """Filter out __MDIO_RAW_UNSPECIFIED_Field_* fields from headers array.
-
-    These fields are added during SEGY import to preserve raw header bytes,
-    but they cause dtype mismatches during export. This function removes them.
-
-    Args:
-        headers: Header array that may contain raw unspecified fields.
-
-    Returns:
-        Header array with raw unspecified fields removed.
-    """
-    if headers.dtype.names is None:
-        return headers
-
-    # Find field names that don't start with __MDIO_RAW_UNSPECIFIED_
-    field_names = [name for name in headers.dtype.names if not name.startswith("__MDIO_RAW_UNSPECIFIED_")]
-
-    if len(field_names) == len(headers.dtype.names):
-        # No raw unspecified fields found, return as-is
-        return headers
-
-    # Create new structured array with only the non-raw fields
-    new_dtype = [(name, headers.dtype.fields[name][0]) for name in field_names]
-    filtered_headers = np.empty(headers.shape, dtype=new_dtype)
-
-    for name in field_names:
-        filtered_headers[name] = headers[name]
-
-    return filtered_headers
-
-
 def make_segy_factory(spec: SegySpec, binary_header: dict[str, int]) -> SegyFactory:
     """Generate SEG-Y factory from MDIO metadata."""
     sample_interval = binary_header["sample_interval"]
@@ -199,9 +167,7 @@ def serialize_to_segy_stack(  # noqa: PLR0913
         samples = samples[live_mask]
         headers = headers[live_mask]
 
-        # Filter out raw unspecified fields that cause dtype mismatches
-        filtered_headers = _filter_raw_unspecified_fields(headers)
-        buffer = segy_factory.create_traces(filtered_headers, samples)
+        buffer = segy_factory.create_traces(headers, samples)
 
         global_index = block_start[0]
         record_id_str = str(global_index)
@@ -233,9 +199,7 @@ def serialize_to_segy_stack(  # noqa: PLR0913
             rec_samples = samples[rec_index][rec_live_mask]
             rec_headers = headers[rec_index][rec_live_mask]
 
-            # Filter out raw unspecified fields that cause dtype mismatches
-            filtered_headers = _filter_raw_unspecified_fields(rec_headers)
-            buffer = segy_factory.create_traces(filtered_headers, rec_samples)
+            buffer = segy_factory.create_traces(rec_headers, rec_samples)
 
             global_index = tuple(block_start[i] + rec_index[i] for i in range(record_ndim))
             record_id_str = "/".join(map(str, global_index))