Merge pull request #104 from LorenFrankLab/np_upgrade

Fixes for numpy 2.0

Author: edeno

.github/workflows/test_package_build.yml

@@ -29,14 +29,14 @@ jobs:
         run: python -m build
       - run: twine check dist/*
       - name: Upload sdist and wheel artifacts
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name: dist
           path: dist/
       - name: Build git archive
         run: mkdir archive && git archive -v -o archive/archive.tgz HEAD
       - name: Upload git archive artifact
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name: archive
           path: archive/
@@ -51,13 +51,13 @@ jobs:
     steps:
       - name: Download sdist and wheel artifacts
         if: matrix.package != 'archive'
-        uses: actions/download-artifact@v3
+        uses: actions/download-artifact@v4
         with:
           name: dist
           path: dist/
       - name: Download git archive artifact
         if: matrix.package == 'archive'
-        uses: actions/download-artifact@v3
+        uses: actions/download-artifact@v4
         with:
           name: archive
           path: archive/
@@ -117,7 +117,7 @@ jobs:
       id-token: write  # IMPORTANT: this permission is mandatory for trusted publishing
     if: github.event_name == 'push' && startsWith(github.ref, 'refs/tags/')
     steps:
-      - uses: actions/download-artifact@v3
+      - uses: actions/download-artifact@v4
         with:
           name: dist
           path: dist/

src/trodes_to_nwb/convert_position.py

@@ -94,7 +94,7 @@ def parse_dtype(fieldstr: str) -> np.dtype:
     --------
     >>> fieldstr = '<time uint32><x float32><y float32><z float32>'
     >>> parse_dtype(fieldstr)
-    dtype([('time', '<u4'), ('x', '<f4'), ('y', '<f4'), ('z', '<f4')])
+    dtype([('time', '<u4', (1,)), ('x', '<f4', (1,)), ('y', '<f4', (1,)), ('z', '<f4', (1,))])
 
     """
     # Returns np.dtype from field string
@@ -172,6 +172,15 @@ def read_trodes_datafile(filename: Path) -> dict:
         return fields_text
 
 
+def convert_datafile_to_pandas(datafile) -> pd.DataFrame:
+    """Takes the output of read_trodes_datafile and converts it to a pandas dataframe.
+    Added for changes identified in numpy 2.2.2
+    """
+    return pd.DataFrame(
+        {key: np.squeeze(datafile["data"][key]) for key in datafile["data"].dtype.names}
+    )
+
+
 def get_framerate(timestamps: np.ndarray) -> float:
     """
     Calculates the framerate of a video based on the timestamps of each frame.
@@ -472,15 +481,12 @@ def get_video_timestamps(video_timestamps_filepath: Path) -> np.ndarray:
         An array of video timestamps.
     """
     # Get video timestamps
+    video_timestamps = read_trodes_datafile(video_timestamps_filepath)["data"]
     video_timestamps = (
-        pd.DataFrame(read_trodes_datafile(video_timestamps_filepath)["data"])
-        .set_index("PosTimestamp")
-        .rename(columns={"frameCount": "HWframeCount"})
-    )
-    return (
-        np.asarray(video_timestamps.HWTimestamp, dtype=np.float64)
+        np.squeeze(video_timestamps["HWTimestamp"]).astype(np.float64)
         / NANOSECONDS_PER_SECOND
     )
+    return video_timestamps
 
 
 def get_position_timestamps(
@@ -519,8 +525,9 @@ def get_position_timestamps(
     logger = logging.getLogger("convert")
 
     # Get video timestamps
+    datafile = read_trodes_datafile(position_timestamps_filepath)
     video_timestamps = (
-        pd.DataFrame(read_trodes_datafile(position_timestamps_filepath)["data"])
+        convert_datafile_to_pandas(datafile)
         .set_index("PosTimestamp")
         .rename(columns={"frameCount": "HWframeCount"})
     )
@@ -549,7 +556,7 @@ def get_position_timestamps(
     # Get position tracking information
     try:
         position_tracking = pd.DataFrame(
-            read_trodes_datafile(position_tracking_filepath)["data"]
+            convert_datafile_to_pandas(read_trodes_datafile(position_tracking_filepath))
         ).set_index("time")
         is_repeat_timestamp = detect_repeat_timestamps(position_tracking.index)
         position_tracking = position_tracking.iloc[~is_repeat_timestamp]
@@ -1005,7 +1012,7 @@ def add_associated_video_files(
             ]
         ].full_path.to_list()[0]
         # get the timestamps
-        video_timestamps = get_video_timestamps(video_timestamps_filepath)
+        video_timestamps = np.squeeze(get_video_timestamps(video_timestamps_filepath))
 
         if convert_video:
             video_file_name = convert_h264_to_mp4(video_path, video_directory)

src/trodes_to_nwb/spike_gadgets_raw_io.py

@@ -560,8 +560,9 @@ def get_analogsignal_multiplexed(self, channel_names=None) -> np.ndarray:
         )
         # read the data into int16
         data = (
-            self._raw_memmap[:, data_offsets[:, 0]]
-            + self._raw_memmap[:, data_offsets[:, 0] + 1] * INT_16_CONVERSION
+            self._raw_memmap[:, data_offsets[:, 0]].astype(np.int16)
+            + self._raw_memmap[:, data_offsets[:, 0] + 1].astype(np.int16)
+            * INT_16_CONVERSION
         )
         # initialize the first row
         analog_multiplexed_data[0] = data[0]
@@ -646,8 +647,8 @@ def get_analogsignal_multiplexed_partial(
         )
         # read the data into int16
         data = (
-            self._raw_memmap[i_start:i_stop, data_offsets[:, 0]]
-            + self._raw_memmap[i_start:i_stop, data_offsets[:, 0] + 1]
+            self._raw_memmap[i_start:i_stop, data_offsets[:, 0]].astype(np.int16)
+            + self._raw_memmap[i_start:i_stop, data_offsets[:, 0] + 1].astype(np.int16)
             * INT_16_CONVERSION
         )
         # initialize the first row
@@ -969,8 +970,9 @@ def get_analogsignal_multiplexed(self, channel_names=None) -> np.ndarray:
         )
         # read the data into int16
         data = (
-            self._raw_memmap[:, data_offsets[:, 0]]
-            + self._raw_memmap[:, data_offsets[:, 0] + 1] * INT_16_CONVERSION
+            self._raw_memmap[:, data_offsets[:, 0]].astype(np.int16)
+            + self._raw_memmap[:, data_offsets[:, 0] + 1].astype(np.int16)
+            * INT_16_CONVERSION
         )
         # initialize the first row
         # if no previous state, assume first segment. Default to superclass behavior

src/trodes_to_nwb/tests/test_convert_position.py

@@ -12,6 +12,7 @@
 from trodes_to_nwb.convert_intervals import add_epochs, add_sample_count
 from trodes_to_nwb.convert_position import (
     add_position,
+    convert_datafile_to_pandas,
     correct_timestamps_for_camera_to_mcu_lag,
     detect_repeat_timestamps,
     detect_trodes_time_repeats_or_frame_jumps,
@@ -79,7 +80,7 @@ def test_read_trodes_datafile_correct_settings(tmp_path):
     result = read_trodes_datafile(filename)
     assert result["clock rate"] == "30000"
 
-    expected_data = pd.DataFrame(result["data"])
+    expected_data = convert_datafile_to_pandas(result)
     assert expected_data["field1"].dtype == np.uint32
     assert expected_data["field2"].dtype == np.int32
     assert np.array_equal(expected_data.field1, np.array([1, 3], dtype=np.uint32))