Merge branch 'main' into v1

# Conflicts:
#	docs/requirements.txt
#	pyproject.toml
#	uv.lock

Author: tasansal

.github/workflows/constraints.txt

@@ -1,2 +1,2 @@
-bump-my-version==1.1.2
-nox==2025.2.9
+bump-my-version==1.2.1
+nox==2025.5.1

.github/workflows/tests.yml

@@ -130,6 +130,6 @@ jobs:
           nox --session=coverage -- xml
 
       - name: Upload coverage report
-        uses: codecov/codecov-action@v5.1.2
+        uses: codecov/codecov-action@v5.4.3
         env:
           CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}

docs/requirements.txt

@@ -1,7 +1,7 @@
 autodoc-pydantic==2.2.0
-furo==2024.8.6
+furo==2025.7.19
 linkify-it-py==2.0.3
-myst-nb==1.2.0
+myst-nb==1.3.0
 sphinx==8.2.3
 sphinx-click==6.0.0
 sphinx-copybutton==0.5.2

pyproject.toml

@@ -87,7 +87,7 @@ docs = [
 
 
 [tool.uv]
-required-version = "0.6.11"
+required-version = ">=0.6.11"
 
 [tool.ruff]
 target-version = "py311"

src/mdio/segy/_workers.py

@@ -5,31 +5,47 @@
 import os
 from typing import TYPE_CHECKING
 from typing import Any
+from typing import TypedDict
 from typing import cast
 
 import numpy as np
+from segy import SegyFile
 
 if TYPE_CHECKING:
-    from segy import SegyFile
     from segy.arrays import HeaderArray
+    from segy.config import SegySettings
+    from segy.schema import SegySpec
     from zarr import Array
 
     from mdio.core import Grid
 
 
-def header_scan_worker(segy_file: SegyFile, trace_range: tuple[int, int]) -> HeaderArray:
+class SegyFileArguments(TypedDict):
+    """Arguments to open SegyFile instance creation."""
+
+    url: str
+    spec: SegySpec | None
+    settings: SegySettings | None
+
+
+def header_scan_worker(
+    segy_kw: SegyFileArguments,
+    trace_range: tuple[int, int],
+) -> HeaderArray:
     """Header scan worker.
 
     If SegyFile is not open, it can either accept a path string or a handle that was opened in
     a different context manager.
 
     Args:
-        segy_file: SegyFile instance.
+        segy_kw: Arguments to open SegyFile instance.
         trace_range: Tuple consisting of the trace ranges to read.
 
     Returns:
         HeaderArray parsed from SEG-Y library.
     """
+    segy_file = SegyFile(**segy_kw)
+
     slice_ = slice(*trace_range)
 
     cloud_native_mode = os.getenv("MDIO__IMPORT__CLOUD_NATIVE", default="False")
@@ -52,7 +68,7 @@ def header_scan_worker(segy_file: SegyFile, trace_range: tuple[int, int]) -> Hea
 
 
 def trace_worker(
-    segy_file: SegyFile,
+    segy_kw: SegyFileArguments,
     data_array: Array,
     metadata_array: Array,
     grid: Grid,
@@ -68,7 +84,7 @@ def trace_worker(
     slices across the sample dimension since SEG-Y data isn't chunked, eliminating concern.
 
     Args:
-        segy_file: SegyFile instance.
+        segy_kw: Arguments to open SegyFile instance.
         data_array: Handle for zarr.Array we are writing traces to
         metadata_array: Handle for zarr.Array we are writing trace headers
         grid: mdio.Grid instance
@@ -78,6 +94,7 @@ def trace_worker(
         Partial statistics for chunk, or None
     """
     # Special case where there are no traces inside chunk.
+    segy_file = SegyFile(**segy_kw)
     live_subset = grid.live_mask[chunk_indices[:-1]]
 
     if np.count_nonzero(live_subset) == 0:

src/mdio/segy/blocked_io.py

@@ -2,7 +2,6 @@
 
 from __future__ import annotations
 
-import multiprocessing as mp
 import os
 from concurrent.futures import ProcessPoolExecutor
 from itertools import repeat
@@ -48,22 +47,23 @@ def to_zarr(segy_file: SegyFile, grid: Grid, data_array: Array, header_array: Ar
     chunker = ChunkIterator(data_array, chunk_samples=False)
     num_chunks = len(chunker)
 
-    # For Unix async writes with s3fs/fsspec & multiprocessing, use 'spawn' instead of default
-    # 'fork' to avoid deadlocks on cloud stores. Slower but necessary. Default on Windows.
     num_cpus = int(os.getenv("MDIO__IMPORT__CPU_COUNT", default_cpus))
     num_workers = min(num_chunks, num_cpus)
-    context = mp.get_context("spawn")
-    executor = ProcessPoolExecutor(max_workers=num_workers, mp_context=context)
 
     # Chunksize here is for multiprocessing, not Zarr chunksize.
     pool_chunksize, extra = divmod(num_chunks, num_workers * 4)
     pool_chunksize += 1 if extra else pool_chunksize
 
+    segy_kw = {
+        "url": segy_file.fs.unstrip_protocol(segy_file.url),
+        "spec": segy_file.spec,
+        "settings": segy_file.settings,
+    }
     tqdm_kw = {"unit": "block", "dynamic_ncols": True}
-    with executor:
+    with ProcessPoolExecutor(max_workers=num_workers) as executor:
         lazy_work = executor.map(
             trace_worker,  # fn
-            repeat(segy_file),
+            repeat(segy_kw),
             repeat(data_array),
             repeat(header_array),
             repeat(grid),

src/mdio/segy/parsers.py

@@ -2,7 +2,6 @@
 
 from __future__ import annotations
 
-import multiprocessing as mp
 import os
 from concurrent.futures import ProcessPoolExecutor
 from itertools import repeat
@@ -48,15 +47,17 @@ def parse_index_headers(
 
         trace_ranges.append((start, stop))
 
-    # For Unix async reads with s3fs/fsspec & multiprocessing, use 'spawn' instead of default
-    # 'fork' to avoid deadlocks on cloud stores. Slower but necessary. Default on Windows.
     num_cpus = int(os.getenv("MDIO__IMPORT__CPU_COUNT", default_cpus))
     num_workers = min(n_blocks, num_cpus)
-    context = mp.get_context("spawn")
 
+    segy_kw = {
+        "url": segy_file.fs.unstrip_protocol(segy_file.url),
+        "spec": segy_file.spec,
+        "settings": segy_file.settings,
+    }
     tqdm_kw = {"unit": "block", "dynamic_ncols": True}
-    with ProcessPoolExecutor(num_workers, mp_context=context) as executor:
-        lazy_work = executor.map(header_scan_worker, repeat(segy_file), trace_ranges)
+    with ProcessPoolExecutor(num_workers) as executor:
+        lazy_work = executor.map(header_scan_worker, repeat(segy_kw), trace_ranges)
 
         if progress_bar is True:
             lazy_work = tqdm(

tests/integration/test_segy_import_export_masked.py

@@ -145,11 +145,7 @@ def __iter__(self) -> Iterable[MaskedExportConfigTypes]:
 # fmt: on
 
 
-def mock_nd_segy(
-    path: str,
-    grid_conf: GridConfig,
-    segy_factory_conf: SegyFactoryConfig,
-) -> None:
+def mock_nd_segy(path: str, grid_conf: GridConfig, segy_factory_conf: SegyFactoryConfig) -> None:
     """Create a fake SEG-Y file with a multidimensional grid."""
     spec = get_segy_standard(segy_factory_conf.revision)
 
@@ -161,6 +157,19 @@ def mock_nd_segy(
     header_flds.append(HeaderField(name="samples_per_trace", byte=115, format="int16"))
     header_flds.append(HeaderField(name="sample_interval", byte=117, format="int16"))
 
+    # Add coordinates: {SRC-REC-CDP}-X/Y
+    header_flds.extend(
+        [
+            HeaderField(name="coord_scalar", byte=71, format="int16"),
+            HeaderField(name="src_x", byte=73, format="int32"),
+            HeaderField(name="src_y", byte=77, format="int32"),
+            HeaderField(name="rec_x", byte=81, format="int32"),
+            HeaderField(name="rec_y", byte=85, format="int32"),
+            HeaderField(name="cdp_x", byte=181, format="int32"),
+            HeaderField(name="cdp_y", byte=185, format="int32"),
+        ]
+    )
+
     spec = spec.customize(trace_header_fields=header_flds)
     spec.segy_standard = segy_factory_conf.revision
     factory = SegyFactory(spec=spec, samples_per_trace=segy_factory_conf.num_samples)
@@ -177,9 +186,17 @@ def mock_nd_segy(
     samples = factory.create_trace_sample_template(trace_numbers.size)
     headers = factory.create_trace_header_template(trace_numbers.size)
 
+    # Fill dimension coordinates (e.g. inline, crossline, etc.)
     for dim_idx, dim in enumerate(grid_conf.dims):
         headers[dim.name] = dim_grid[dim_idx].ravel()
 
+    # Fill coordinates (e.g. {SRC-REC-CDP}-X/Y
+    headers["coord_scalar"] = -100
+    for field in ["cdp_x", "src_x", "rec_x"]:
+        headers[field] = np.random.randint(low=700000, high=900000, size=trace_numbers.size) * 100
+    for field in ["cdp_y", "src_y", "rec_y"]:
+        headers[field] = np.random.randint(low=4000000, high=5000000, size=trace_numbers.size) * 100
+
     samples[:] = trace_numbers[..., None]
 
     with fsspec.open(path, mode="wb") as fp:
@@ -188,10 +205,7 @@ def mock_nd_segy(
         fp.write(factory.create_traces(headers, samples))
 
 
-def generate_selection_mask(
-    selection_conf: SelectionMaskConfig,
-    grid_conf: GridConfig,
-) -> NDArray:
+def generate_selection_mask(selection_conf: SelectionMaskConfig, grid_conf: GridConfig) -> NDArray:
     """Generate a boolean selection mask for a masked export test."""
     spatial_shape = [dim.size for dim in grid_conf.dims]
     mask_dims = selection_conf.mask_num_dims
@@ -224,9 +238,7 @@ def export_masked_path(tmp_path_factory: pytest.TempPathFactory) -> Path:
 class TestNdImportExport:
     """Test import/export of n-D SEG-Ys to MDIO, with and without selection mask."""
 
-    def test_import(
-        self, test_conf: MaskedExportConfig, export_masked_path: Path
-    ) -> None:
+    def test_import(self, test_conf: MaskedExportConfig, export_masked_path: Path) -> None:
         """Test import of an n-D SEG-Y file to MDIO."""
         grid_conf, segy_factory_conf, segy_to_mdio_conf, _ = test_conf
 
@@ -249,9 +261,7 @@ def test_import(
             overwrite=True,
         )
 
-    def test_export(
-        self, test_conf: MaskedExportConfig, export_masked_path: Path
-    ) -> None:
+    def test_export(self, test_conf: MaskedExportConfig, export_masked_path: Path) -> None:
         """Test export of an n-D MDIO file back to SEG-Y."""
         grid_conf, segy_factory_conf, segy_to_mdio_conf, _ = test_conf
 
@@ -277,9 +287,7 @@ def test_export(
         actual_sgy = SegyFile(segy_rt_path)
         assert_array_equal(actual_sgy.trace[:], expected_sgy.trace[:])
 
-    def test_export_masked(
-        self, test_conf: MaskedExportConfig, export_masked_path: Path
-    ) -> None:
+    def test_export_masked(self, test_conf: MaskedExportConfig, export_masked_path: Path) -> None:
         """Test export of an n-D MDIO file back to SEG-Y with masked export."""
         grid_conf, segy_factory_conf, segy_to_mdio_conf, selection_conf = test_conf