perf(inlet): speed up pull_chunk with bulk buffer slicing

.gitignore

@@ -19,3 +19,6 @@ uv.lock
 /src/pylsl/__version__.py
 /src/pylsl/lib/lslver.exe
 liblsl.zip
+
+.claude
+CLAUDE.md

src/pylsl/inlet.py

@@ -263,18 +263,19 @@ def pull_chunk(self, timeout=0.0, max_samples=1024, dest_obj=None):
         handle_error(errcode)
         # return results (note: could offer a more efficient format in the
         # future, e.g., a numpy array)
-        num_samples = num_elements / num_channels
+        num_samples = num_elements // num_channels
         if dest_obj is None:
+            flat = data_buff[:num_elements]
             samples = [
-                [data_buff[s * num_channels + c] for c in range(num_channels)]
-                for s in range(int(num_samples))
+                flat[s * num_channels : (s + 1) * num_channels]
+                for s in range(num_samples)
             ]
             if self.channel_format == cf_string:
                 samples = [[v.decode("utf-8") for v in s] for s in samples]
                 free_char_p_array_memory(data_buff, max_values)
         else:
             samples = None
-        timestamps = [ts_buff[s] for s in range(int(num_samples))]
+        timestamps = ts_buff[:num_samples]
         return samples, timestamps
 
     def samples_available(self):

test/test_pull_chunk.py

@@ -0,0 +1,67 @@
+"""Round-trip test for StreamInlet.pull_chunk.
+
+Pushes a known chunk and pulls it back, asserting the extracted data is
+identical in value, shape, and type. Covers both paths the bulk-slice
+extraction must preserve: a multi-channel numeric chunk and a
+variable-length string ("Markers"-style) chunk.
+"""
+
+import time
+
+import pytest
+
+import pylsl
+
+# (channel_format, samples) — distinct values per channel/sample, including
+# empty and multi-byte strings to exercise variable-length decoding.
+CASES = {
+    "double64": (
+        pylsl.cf_double64,
+        [[1.0, 2.5, -3.25], [4.0, 5.5, 6.75], [7.0, 8.5, 9.25]],
+    ),
+    "string": (
+        pylsl.cf_string,
+        [["a", "bb", ""], ["ccc", "dddd", "e"], ["", "f", "ééé"]],
+    ),
+}
+
+
+@pytest.mark.parametrize("channel_format,samples", CASES.values(), ids=CASES.keys())
+def test_pull_chunk_roundtrip(channel_format: int, samples: list):
+    n_samples = len(samples)
+    n_channels = len(samples[0])
+
+    info = pylsl.StreamInfo(
+        name="test_pull_chunk",
+        type="test",
+        channel_count=n_channels,
+        nominal_srate=0,
+        channel_format=channel_format,
+        source_id="test_pull_chunk_id",
+    )
+    outlet = pylsl.StreamOutlet(info)
+
+    streams = pylsl.resolve_byprop("source_id", "test_pull_chunk_id", timeout=2)
+    assert streams, "outlet was not discovered"
+    inlet = pylsl.StreamInlet(streams[0])
+
+    # Subscribe before pushing so the only chunk isn't sent before the inlet's
+    # data connection is established.
+    inlet.open_stream(timeout=2)
+    time.sleep(0.5)
+    outlet.push_chunk(samples)
+
+    # Data may arrive in more than one chunk; collect until complete.
+    got_samples: list = []
+    got_ts: list = []
+    deadline = time.time() + 5
+    while len(got_samples) < n_samples and time.time() < deadline:
+        chunk, stamps = inlet.pull_chunk(timeout=1.0)
+        if chunk:
+            assert isinstance(stamps, list)
+            assert all(isinstance(row, list) for row in chunk)
+        got_samples.extend(chunk)
+        got_ts.extend(stamps)
+
+    assert got_samples == samples
+    assert len(got_ts) == n_samples