[WIP] Test and refactor

.github/workflows/test.yml

@@ -0,0 +1,28 @@
+# This workflow will install Python dependencies, run tests and lint with a single version of Python
+# For more information see: https://help.github.com/actions/language-and-framework-guides/using-python-with-github-actions
+
+name: Test
+
+on: [push, pull_request]
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: ["3.8", "3.9", "3.10", "3.11"]
+    steps:
+    - uses: actions/checkout@v4
+    - name: Set up Python
+      uses: actions/setup-python@v5
+      with:
+        python-version:  ${{ matrix.python-version }}
+    - name: Display Python version
+      run: python -c "import sys; print(sys.version)"
+    - name: Install dependencies
+      run: |
+        pip install -r requirements.txt --extra-index-url https://download.pytorch.org/whl/cpu
+        pip install pytest pytest-cov coveralls pytest-sugar
+    - name: Run Tests
+      run: |
+        pytest --doctest-modules --cov=yaltai --verbose

README.md

@@ -20,7 +20,7 @@ Convert (and split optionally) your data
 ```bash
 # Keeps .1 data in the validation set and convert all alto into YOLOv5 format
 #  Keeps the segmonto information up to the regions
-yaltai alto-to-yolo PATH/TO/ALTOorPAGE/*.xml my-dataset --shuffle .1 --segmonto region
+yaltai convert alto-to-yolo PATH/TO/ALTOorPAGE/*.xml my-dataset --shuffle .1 --segmonto region
 ```
 
 And then [train YOLO](https://github.com/ultralytics/yolov5/wiki/Train-Custom-Data) 

setup.py

@@ -105,7 +105,7 @@ def run(self):
     # py_modules=['mypackage'],
 
     entry_points={
-        'console_scripts': ['yaltai=yaltai.yaltai:cli', 'kraken-yaltai=yaltai.kraken_yaltai:cli'],
+        'console_scripts': ['yaltai=yaltai.cli.yaltai:yaltai_cli'],
     },
     install_requires=REQUIRED,
     extras_require=EXTRAS,

tests/test_convert.py

@@ -0,0 +1,40 @@
+import tempfile
+import os
+from click.testing import CliRunner
+
+from yaltai.cli.yaltai import yaltai_cli
+
+
+def test_yaltai_single_alto_to_xml():
+    """Ensures that we can convert to YOLO format"""
+    runner = CliRunner()
+
+    with tempfile.TemporaryDirectory() as tempdir:
+        # Run the Click command
+        result = runner.invoke(
+            yaltai_cli,
+            [
+                "convert",
+                "alto-to-yolo",
+                os.path.join(os.path.abspath(os.path.dirname(__file__)), "test_files", "alto_dataset", "output.xml"),
+                tempdir,
+            ])
+
+        # Ensure the command ran successfully
+        assert result.exit_code == 0
+        assert "Found 1 to convert." in result.output
+        assert "- 00001 NumberingZone" in result.output, "Correct number of zone types are found"
+        assert "- 00001 RunningTitleZone" in result.output, "Correct number of zone types are found"
+        assert "- 00001 MainZone-P-Continued" in result.output, "Correct number of zone types are found"
+        assert "- 00004 MainZone-P" in result.output, "Correct number of zone types are found"
+        with open(os.path.join(tempdir, "labels", "output.txt")) as f:
+            data = [line.split() for line in f.read().split("\n")]
+        assert data == [
+            ['0', '0.345606', '0.117500', '0.037501', '0.020206'],
+            ['1', '0.612827', '0.118333', '0.212882', '0.020736'],
+            ['2', '0.616390', '0.210833', '0.578172', '0.142851'],
+            ['3', '0.616390', '0.365833', '0.577029', '0.162800'],
+            ['3', '0.616390', '0.547500', '0.574365', '0.199581'],
+            ['3', '0.616390', '0.727500', '0.576050', '0.161861'],
+            ['3', '0.611639', '0.819167', '0.583077', '0.022773']
+        ]

tests/test_files/.gitignore

@@ -0,0 +1 @@
+page1.xml

tests/test_files/annot1.txt

@@ -0,0 +1,7 @@
+14 0.49139869729173813 0.3715823573386494 0.6673020226259856 0.07884874158983304
+14 0.49408639012684263 0.23613506105158236 0.6722831676379842 0.15367804634936458
+35 0.49086047308878983 0.1285771243458759 0.26186150154268084 0.03464490406179915
+31 0.8086732944806309 0.13971592324943932 0.03529310935893041 0.03310740094692251
+14 0.49913952691121016 0.594151507600299 0.6766986630099417 0.31005232992773485
+14 0.507857387727117 0.8290804884126588 0.6902228316763799 0.1168950909543982
+1 0.9231984916009598 0.4728781460254174 0.1535995886184436 0.18793919760777475

tests/test_files/annot2.txt

@@ -0,0 +1,6 @@
+23 0.4914115952466691 0.45760844079718643 0.6449045732805185 0.4102297772567409
+23 0.4933705437522506 0.7764806565064478 0.6465250270075622 0.2320398593200469
+31 0.7885343896290963 0.11840562719812427 0.033172488296723084 0.026182883939038688
+35 0.4834569679510263 0.1191535756154748 0.4334101548433561 0.024438452520515828
+24 0.48678790061217136 0.20451113716295427 0.6372740367302845 0.10247831184056272
+28 0.47840835433921497 0.14436342321219228 0.24528628015844436 0.025828839390386868

tests/test_files/label_map.txt

@@ -0,0 +1,4 @@
+Class0
+Class1
+
+Stuff

tests/test_kraken.py

@@ -0,0 +1,108 @@
+import os
+import logging
+import pytest
+from click.testing import CliRunner
+from ultralytics.utils import LOGGER
+from yaltai.cli.yaltai import yaltai_cli
+from kraken.lib.xml import XMLPage
+
+
+@pytest.fixture(scope='function')
+def custom_logger():
+    class CustomLoggingHandler(logging.Handler):
+        def __init__(self):
+            super().__init__()
+            # A list to store log records (level and message)
+            self.records = []
+
+        def emit(self, record):
+            # Append a tuple of the log level and the message to the records list
+            self.records.append((record.levelname, record.getMessage()))
+
+        def clear(self):
+            self.records = []
+
+    # Create a logger
+    logger = LOGGER
+    logger.setLevel(logging.DEBUG)  # Set the logger to handle all log levels
+
+    # Create and add the custom handler to the logger
+    custom_handler = CustomLoggingHandler()
+    logger.addHandler(custom_handler)
+
+    # Clear the records list before each test
+    custom_handler.records.clear()
+
+    # Provide both the logger and handler for access during tests
+    yield custom_handler
+
+    custom_handler.records.clear()
+
+    # Optional: Remove the handler after the test to prevent interference
+    logger.removeHandler(custom_handler)
+
+
+def test_yaltai_single_alto_to_xml(custom_logger):
+    """Ensures that we can convert to YOLO format"""
+    runner = CliRunner()
+
+    # Trigger a warning.
+    result = runner.invoke(
+        yaltai_cli,
+        [
+            "kraken",
+            "--alto",
+            "-i",
+            os.path.join(os.path.abspath(os.path.dirname(__file__)), "test_files", "page1.jpg"),
+            os.path.join(os.path.abspath(os.path.dirname(__file__)), "test_files", "page1.xml"),
+            "segment",
+            "-y",
+            os.path.join(os.path.abspath(os.path.dirname(__file__)), "nano-yolo-ladas.pt")
+        ]
+    )
+    print(result.output)
+    assert result.exit_code == 0
+    assert "page1.jpg: 640x352 2 GraphicZones, 1 MainZone-P-Continued, 1 MainZone-Sp, 2 QuireMarksZones" in "\n".join([
+        record[1]
+        for record in custom_logger.records
+    ])
+
+    page = XMLPage(os.path.join(os.path.abspath(os.path.dirname(__file__)), "test_files", "page1.xml"))
+    assert {
+               region_type: [region.boundary for region in regions]
+               for region_type, regions in page.regions.items()
+           } == {
+               'GraphicZone': [[(614.0, 12.0),
+                                (2614.0, 12.0),
+                                (2614.0, 819.0),
+                                (614.0, 819.0),
+                                (614.0, 12.0)],
+                               [(720.0, 0.0),
+                                (2634.0, 0.0),
+                                (2634.0, 343.0),
+                                (720.0, 343.0),
+                                (720.0, 0.0)]],
+               'MainZone-P-Continued': [[(122.0, 1536.0),
+                                         (2218.0, 1536.0),
+                                         (2218.0, 2888.0),
+                                         (122.0, 2888.0),
+                                         (122.0, 1536.0)]],
+               'MainZone-Sp': [[(95.0, 2974.0),
+                                (2201.0, 2974.0),
+                                (2201.0, 4854.0),
+                                (95.0, 4854.0),
+                                (95.0, 2974.0)]],
+               'QuireMarksZone': [[(1617.0, 4877.0),
+                                   (1825.0, 4877.0),
+                                   (1825.0, 4980.0),
+                                   (1617.0, 4980.0),
+                                   (1617.0, 4877.0)],
+                                  [(1531.0, 4841.0),
+                                   (1814.0, 4841.0),
+                                   (1814.0, 4983.0),
+                                   (1531.0, 4983.0),
+                                   (1531.0, 4841.0)]]
+           }
+
+    assert len([line.baseline for line in page.lines.values()])
+    # ToDo: Add a test to check for line being part of regions

tests/test_utils.py

@@ -0,0 +1,60 @@
+import os
+import numpy as np
+from yaltai.utils import read_labelmap, parse_box_labels, XYXY
+
+
+def test_read_labelmap():
+    """Asserts that reading a label map works"""
+    labels = read_labelmap(os.path.join(
+        os.path.dirname(__file__),
+        "test_files",
+        "label_map.txt"
+    ))
+    assert labels == ["Class0", "Class1", "Stuff"]
+
+
+def test_read_files():
+    """Asserts that parsing COCO/YOLO formats work"""
+    annots, arrays = parse_box_labels([
+        os.path.join(
+            os.path.dirname(__file__),
+            "test_files",
+            "annot1.txt"
+        ),
+        os.path.join(
+            os.path.dirname(__file__),
+            "test_files",
+            "annot2.txt"
+        )
+    ])
+    assert annots == {
+        'boxes': [
+            XYXY(x0=15, y0=33, x1=82, y1=41),
+            XYXY(x0=15, y0=15, x1=83, y1=31),
+            XYXY(x0=35, y0=11, x1=62, y1=14),
+            XYXY(x0=79, y0=12, x1=82, y1=15),
+            XYXY(x0=16, y0=43, x1=83, y1=74),
+            XYXY(x0=16, y0=77, x1=85, y1=88),
+            XYXY(x0=84, y0=37, x1=99, y1=56),
+            XYXY(x0=16, y0=25, x1=81, y1=66),
+            XYXY(x0=17, y0=66, x1=81, y1=89),
+            XYXY(x0=77, y0=10, x1=80, y1=13),
+            XYXY(x0=26, y0=10, x1=70, y1=13),
+            XYXY(x0=16, y0=15, x1=80, y1=25),
+            XYXY(x0=35, y0=13, x1=60, y1=15)
+        ],
+        'labels': [14, 14, 35, 31, 14, 14, 1, 23, 23, 31, 35, 24, 28]
+    }
+    assert (arrays[0] == np.array([[15, 33, 82, 41, 14, 0, 0],
+                      [15, 15, 83, 31, 14, 0, 0],
+                      [35, 11, 62, 14, 35, 0, 0],
+                      [79, 12, 82, 15, 31, 0, 0],
+                      [16, 43, 83, 74, 14, 0, 0],
+                      [16, 77, 85, 88, 14, 0, 0],
+                      [84, 37, 99, 56, 1, 0, 0]])).all()
+    assert (arrays[1] == np.array([[16, 25, 81, 66, 23, 0, 0],
+                      [17, 66, 81, 89, 23, 0, 0],
+                      [77, 10, 80, 13, 31, 0, 0],
+                      [26, 10, 70, 13, 35, 0, 0],
+                      [16, 15, 80, 25, 24, 0, 0],
+                      [35, 13, 60, 15, 28, 0, 0]])).all()