Add a simple CI

.github/workflows/pypi-publish.yml

@@ -12,15 +12,15 @@ jobs:
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        os: [ubuntu-latest, windows-latest]
+        os: [ubuntu-latest, windows-latest, macos-latest]
 
     steps:
       - uses: actions/checkout@v4
 
       - name: Build wheels
         uses: pypa/cibuildwheel@v2.23.3
         env:
-          CIBW_BUILD: "cp310-* cp311-* cp312-* cp313-*"
+          CIBW_BUILD: "cp310-* cp311-* cp312-* cp313-* cp314-*"
           CIBW_SKIP: "*-win32 *_i686 *musllinux*"
 
       - uses: actions/upload-artifact@v4

.github/workflows/test-publish.yml

@@ -9,15 +9,15 @@ jobs:
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        os: [ubuntu-latest, windows-latest]
+        os: [ubuntu-latest, macos-latest, windows-latest]
 
     steps:
       - uses: actions/checkout@v4
 
       - name: Build wheels
         uses: pypa/cibuildwheel@v2.23.3
         env:
-          CIBW_BUILD: "cp310-* cp311-* cp312-* cp313-*"
+          CIBW_BUILD: "cp310-* cp311-* cp312-* cp313-* cp314-*"
           CIBW_SKIP: "*-win32 *_i686 *musllinux*"
 
       - uses: actions/upload-artifact@v4

.github/workflows/tests.yml

@@ -0,0 +1,65 @@
+name: Tests
+
+on:
+  push:
+    branches: [master]
+  pull_request:
+  workflow_dispatch:
+
+jobs:
+  test:
+    name: ${{ matrix.os }} / py${{ matrix.python-version }}
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu-latest, macos-latest, windows-latest]
+        python-version: ["3.11", "3.12", "3.13", "3.14"]
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: astral-sh/setup-uv@v6
+        with:
+          enable-cache: true
+          python-version: ${{ matrix.python-version }}
+
+      - name: Install dependencies
+        run: uv sync --group dev
+
+      # setuptools isn't a runtime/dev dependency, only needed transiently
+      # to drive build_ext here.
+      - name: Build C extension in place
+        run: uv run --with setuptools python setup.py build_ext --inplace
+
+      - name: Confirm C extension is importable
+        shell: bash
+        run: uv run python -c "import _dbdreader; print('C extension:', _dbdreader.__file__)"
+
+      - name: Run tests (C extension)
+        shell: bash
+        env:
+          DBDREADER_C_EXTENSION: "1"
+        run: uv run pytest tests -v
+
+      - name: Run tests (pure-Python fallback)
+        shell: bash
+        env:
+          DBDREADER_C_EXTENSION: "0"
+        run: uv run pytest tests -v
+
+  docs:
+    name: Build docs
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: astral-sh/setup-uv@v6
+        with:
+          enable-cache: true
+
+      - name: Install dependencies
+        run: uv sync --group dev
+
+      - name: Build Sphinx docs
+        run: uv run sphinx-build -W -b html docs/source docs/build/html

.gitignore

@@ -17,3 +17,13 @@ MANIFEST
 *.so
 # this cac file gets created anyway.
 data/cac/daad1b20.cac
+uv.lock
+dbdreader/data/01600000.dbd
+dbdreader/data/02380108.ebd
+
+# pixi things in case user puts those in:
+pixi.lock
+pixi.toml
+.pixi/
+
+.gitattributes

conftest.py

@@ -0,0 +1,5 @@
+import sys
+import os
+
+# Ensure tests always import from the local source tree, not any installed package.
+sys.path.insert(0, os.path.dirname(__file__))

dbdreader/dbdreader.py

@@ -5,6 +5,7 @@
 import time
 import numpy
 import glob
+import fnmatch
 import sys
 import re
 import datetime
@@ -407,6 +408,38 @@ def set_cachedir(cls, path, force_makedirs=False):
 
 
 
+def _glob(pattern):
+    ''' Case-sensitive glob.
+
+    glob.glob() matches filenames case-insensitively on platforms with a
+    case-insensitive filesystem (notably Windows). For glider data files
+    this is a problem: lower and upper case extensions (.sbd/.SBD,
+    .tbd/.TBD, and so on) denote different (full resolution vs.
+    compact/telemetered) data files and must not be confused. This
+    wrapper filters glob.glob()'s result so that only names matching the
+    pattern case-sensitively are kept, giving the same result on every
+    platform.
+
+    Parameters
+    ----------
+    pattern : str
+        search pattern, as passed to glob.glob()
+
+    Returns
+    -------
+    list of str
+        filenames matching pattern, case-sensitively.
+    '''
+    normalised_pattern = os.path.normpath(pattern)
+    matches = [fn for fn in glob.glob(pattern)
+               if fnmatch.fnmatchcase(os.path.normpath(fn), normalised_pattern)]
+    # glob.glob() returns filenames built with the OS-native separator
+    # (e.g. backslashes on Windows), even when the pattern itself uses
+    # forward slashes. Normalise to forward slashes so that filenames
+    # are stable and comparable across platforms.
+    return [fn.replace(os.sep, '/') for fn in matches]
+
+
 class DBDList(list):
 
     ''' List that properly sorts dbd files.
@@ -621,7 +654,7 @@ def get_filenames(self, pattern, filenames, cacheDir=None):
         if not pattern and not filenames:
             raise ValueError("Expected some pattern to search files for or file list.")
         if pattern:
-            all_filenames=DBDList(glob.glob(pattern))
+            all_filenames=DBDList(_glob(pattern))
         elif filenames:
             all_filenames=DBDList(filenames)
         else:
@@ -1403,7 +1436,7 @@ def __init__(self,filenames=None,pattern=None,cacheDir=None,complemented_files_o
         if filenames:
             fns+=filenames
         if pattern:
-            fns+=glob.glob(pattern)
+            fns+=_glob(pattern)
         if len(fns)==0:
             raise DbdError(DBD_ERROR_NO_FILES_FOUND)
         fns.sort()
@@ -2028,7 +2061,7 @@ def _refresh_cache(self):
         time_limits[1]=min(time_limits[1],time_limits_dataset[1])
 
     def _format_time(self,t,fmt):
-        tmp = datetime.datetime.fromtimestamp(t, datetime.UTC)
+        tmp = datetime.datetime.fromtimestamp(t, datetime.timezone.utc)
         return tmp.strftime(fmt)
 
     def _get_time_range(self,time_limits,fmt):

extension/decompress.c

@@ -90,15 +90,23 @@ static FILE* fopen_compressed_file_fopen(const char* filename, int* errorno){
 	fpmem = fopen(base, "rb");
       }
       else{
+	/* Decompression failed partway through, so the base file on
+	   disk is incomplete/corrupt. Remove it -- otherwise a later
+	   call would find this partial file, mistake it for a
+	   complete, previously-decompressed file, and silently return
+	   truncated data instead of re-reporting the error. */
+	remove(base);
 	fpmem=NULL;
       }
     }
+    else {
+      /* Could not create the base file for writing. This is fatal. */
+      *errorno = ERROR_FAILED_TO_WRITE_BASE_FILE;
+    }
   }
-  else {
-    /* Writing operation failed. This is fatal. */
-    *errorno = ERROR_FAILED_TO_WRITE_BASE_FILE;
-    fpmem=NULL;
-  }
+  /* else: the decompressed file already exists on disk (written by
+     a previous call), so just use it as-is -- fpmem is already open
+     for reading and *errorno is NO_ERROR. */
   /* Here fpmem is either NULL (and all failed, or it points to the
      decompressed file.*/
   free(extension_decompressed);

pyproject.toml

@@ -72,6 +72,6 @@ exclude = '''
 '''
 
 [tool.mypy]
-python_version = "3.13"
+python_version = "3.14"
 strict = true  # Enable all checks
 disallow_untyped_calls = true

tests/test_decompress.py

@@ -1,6 +1,5 @@
 import os
 from hashlib import md5
-from itertools import chain
 
 if not __name__ == '__main__':
     from pytest import fixture
@@ -23,8 +22,8 @@ def load_verification_data():
     with open(filename_verify, 'r') as fp:
         data = fp.read()
     return data
-        
-# Open a file and lazy read block by block. Compares with 
+
+# Open a file and lazy read block by block. Compares with
 # uncompressed file
 #
 def test_read_file(load_verification_data):
@@ -36,7 +35,7 @@ def test_read_file(load_verification_data):
             data += block
     assert data.decode('ascii') == verification_data
 
-# Open a file and decompress whole file. Compares with 
+# Open a file and decompress whole file. Compares with
 # uncompressed file
 #
 def test_read_file_in_memory(load_verification_data):
@@ -46,16 +45,16 @@ def test_read_file_in_memory(load_verification_data):
         data = d.decompress()
     assert data.decode('ascii') == verification_data
 
-# Open a file and decompress first block only. 
-# 
+# Open a file and decompress first block only.
+#
 #
 def test_read_file_one_block_only():
     filename = 'dbdreader/data/01600000.mcg'
     with Decompressor(filename) as d:
         blocks = [block for block in d.decompressed_blocks(n=1)]
     assert len(blocks) == 1
 
-# Open a file and lazy read block by block. Compares with 
+# Open a file and lazy read block by block. Compares with
 # uncompressed file. Instead of using with statement, open file
 # explicitly
 def test_read_file_explicit_file_opener(load_verification_data):
@@ -68,7 +67,7 @@ def test_read_file_explicit_file_opener(load_verification_data):
             data += block
     assert data.decode('ascii') == verification_data
 
-# Open a file and decompress whole file. Compares with 
+# Open a file and decompress whole file. Compares with
 # uncompressed file. Instead of using with statement, open file
 # explicitly
 def test_read_file_in_memory_explicit_file_opener(load_verification_data):
@@ -130,8 +129,8 @@ def test_extension_generator_with_invalid_extension():
         s = fd._generate_filename_for_output('01600000.cd')
 
 
-        
-# Open a file and lazy read block by block. Check cac filesCompares with 
+
+# Open a file and lazy read block by block. Check cac filesCompares with
 # uncompressed file
 #
 def test_read_ccc_file():
@@ -167,8 +166,8 @@ def test_CompressedFileReadlines(load_verification_data):
         data = b"".join(lines)
     assert data.decode('ascii') == verification_data
 
-    
-    
+
+
 # Test reading of data/01600001.dcd and check it is identical to data/01600001.dbd
 def test_read_compressed_file_C_code():
     compressed_filename = 'dbdreader/data/01600001.dcd'
@@ -178,7 +177,7 @@ def test_read_compressed_file_C_code():
     compressed_data = compressed_dbd.get("m_depth")
     regular_data = regular_dbd.get("m_depth")
     assert np.all(compressed_data[0]==regular_data[0]) and np.all(compressed_data[1]==regular_data[1])
-    
+
 # Test reading of data/0160000?.dcd MultiDBD
 def test_read_compressed_files_C_code():
     pattern = 'dbdreader/data/0160000?.dcd'
@@ -193,7 +192,7 @@ def test_missing_cac_file_for_single_compressed_file():
         os.unlink('dbdreader/data/cac/daad1b20.cac')
     except:
         pass
-    
+
     with pytest.raises(dbdreader.DbdError):
         dbd = dbdreader.DBD('dbdreader/data/01600000.ecd', cacheDir='dbdreader/data/cac')
 
@@ -203,7 +202,7 @@ def test_missing_cac_file_for_compressed_file_multidbd():
         os.unlink('dbdreader/data/cac/daad1b20.cac')
     except:
         pass
-    
+
     dbd = dbdreader.MultiDBD('dbdreader/data/0160000?.ecd', cacheDir='dbdreader/data/cac')
     assert os.path.exists('dbdreader/data/cac/daad1b20.cac')
 
@@ -227,7 +226,23 @@ def test_handle_corrupt_compressed_file_multidbd():
     data_both = dbd.get("sci_water_temp", "sci_water_cond", continue_on_reading_error=True)
     dbd = dbdreader.DBD('dbdreader/data/02380107.ecd', cacheDir='dbdreader/data/cac')
     data_one = dbd.get("sci_water_temp", "sci_water_cond")
-    assert np.all([np.all(x==y) for x, y in zip(chain(*data_both), chain(*data_one))])
+    labels = [
+        "time[sci_water_temp]",
+        "sci_water_temp",
+        "time[sci_water_cond]",
+        "sci_water_cond",
+    ]
+    assert len(data_both) == len(data_one), (
+        f"Returned tuple length differs: {len(data_both)} vs {len(data_one)}"
+    )
+
+    for idx, (arr_both, arr_one) in enumerate(zip(data_both, data_one)):
+        label = labels[idx] if idx < len(labels) else f"output[{idx}]"
+        np.testing.assert_array_equal(
+            arr_both,
+            arr_one,
+            err_msg=f"Mismatch in {label}",
+        )
 
 
 # Test handling of files that fail to decompress properly
@@ -236,15 +251,15 @@ def test_handle_failing_decompression_multidbd():
     t, T = dbd.get("sci_water_temp")
     # we should get 126 values
     assert len(t) == 126
-    
+
 
 # Test handling of files that fail to decompress properly
 def test_handle_failing_decompression_dbd():
     # when called with a DBD instance we expect an error to be raised.
     with pytest.raises(dbdreader.DbdError) as e:
         dbd = dbdreader.DBD('dbdreader/data/02450137.tcd', cacheDir='dbdreader/data/cac')
     assert e.value.value == dbdreader.DBD_ERROR_DECOMPRESSION_ERROR
-
 
-
-
+
+
+