Add pytest CI workflow and test fix

.github/workflows/pytest.yaml

@@ -0,0 +1,35 @@
+name: pytest
+
+on:
+  push:
+    branches:
+      - '**'
+  pull_request:
+
+jobs:
+  pytest:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: ["3.8", "3.9", "3.10"]
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Setup python
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: 'pip'
+          cache-dependency-path: setup.py
+
+      - name: Install dependencies
+        run: |
+          set -e
+          pip install --upgrade pip
+          pip install -e .[dev]
+
+      - name: Run tests with coverage
+        run: |
+          set -e
+          python -m pytest --cov --cov-report=term-missing --cov-report=xml

tests/test_utils.py

@@ -13,8 +13,6 @@
 import numpy as np
 import pandas as pd
 from pathlib import Path
-import tempfile
-import shutil
 import json
 
 from stepcount import utils
@@ -223,7 +221,7 @@ def test_impute_missing_preserves_valid_data(self, accel_data_1_5_days):
     def test_impute_missing_skip_full_missing_days(self):
         """Test that fully missing days are skipped."""
         # Create data with one completely missing day
-        times = pd.date_range('2024-01-15', periods=2*24*3600*10, freq='100ms')
+        times = pd.date_range('2024-01-15', periods=2 * 24 * 3600 * 10, freq='100ms')
         data = pd.DataFrame(
             np.random.randn(len(times), 3) * 0.1 + [0, 0, 1],
             columns=['x', 'y', 'z'],
@@ -396,8 +394,14 @@ def test_read_csv_basic(self, temp_csv_file, sample_rate):
         assert isinstance(data.index, pd.DatetimeIndex)
 
     def test_read_csv_with_row_limits(self, temp_dir, sample_rate):
-        """Test reading CSV with row limits."""
-        # Create a small test CSV
+        """Test reading CSV with row limits (csv_start_row = header row, csv_end_row = last data row)."""
+        # Create a CSV with 3 metadata preamble lines, then header + 1000 data rows
+        # File layout:
+        #   row 0: "# metadata line 1"
+        #   row 1: "# metadata line 2"
+        #   row 2: "# metadata line 3"
+        #   row 3: "time,x,y,z"          <-- header
+        #   row 4..1003: data rows
         times = pd.date_range('2024-01-15', periods=1000, freq='100ms')
         df = pd.DataFrame({
             'time': times,
@@ -406,13 +410,18 @@ def test_read_csv_with_row_limits(self, temp_dir, sample_rate):
             'z': np.random.randn(1000) * 0.1 + 1.0
         })
         csv_path = temp_dir / "test_rows.csv"
-        df.to_csv(csv_path, index=False)
-
-        # Read only rows 100-199 (100 rows)
+        with open(csv_path, 'w') as f:
+            f.write("# metadata line 1\n")
+            f.write("# metadata line 2\n")
+            f.write("# metadata line 3\n")
+        df.to_csv(csv_path, index=False, mode='a')
+
+        # csv_start_row=3 means header is at file row 3 (skip preamble)
+        # csv_end_row=103 means stop at file row 103 (read 100 data rows: rows 4-103)
         data, info = utils.read(
             str(csv_path),
-            csv_start_row=100,
-            csv_end_row=199,
+            csv_start_row=3,
+            csv_end_row=103,
             sample_rate=sample_rate,
             resample_hz=None,
             verbose=False