17929889887

Committed 22 Sep 2025 10:11PM UTC coverage: 88.323% (-4.5%) from 92.814%

Build # 17929889887

Build Type

Pull #64

github

Committed by

web-flow

Commit Message

[pre-commit.ci] pre-commit autoupdate

updates:
- [github.com/pre-commit/pre-commit-hooks: v5.0.0 → v6.0.0](https://github.com/pre-commit/pre-commit-hooks/compare/v5.0.0...v6.0.0)
- [github.com/astral-sh/ruff-pre-commit: v0.12.4 → v0.13.1](https://github.com/astral-sh/ruff-pre-commit/compare/v0.12.4...v0.13.1)
- [github.com/adamchainz/blacken-docs: 1.19.1 → 1.20.0](https://github.com/adamchainz/blacken-docs/compare/1.19.1...1.20.0)

Pull Request Pull Request #64: [pre-commit.ci] pre-commit autoupdate

Run Details

590 of 668 relevant lines covered (88.32%)

0.88 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

47.06

/colour_checker_detection/scripts/inference.py

#!/usr/bin/env python
"""
Colour Checker Detection - Inference
====================================

Define the scripts for colour checker detection using inference based on
*Ultralytics YOLOv8* machine learning model.

Warnings
--------
This script is provided under the terms of the
*GNU Affero General Public License v3.0* as it uses the *Ultralytics YOLOv8*
API which is incompatible with the *BSD-3-Clause*.
"""

from __future__ import annotations

import logging
import os
import typing

if typing.TYPE_CHECKING:
    from pathlib import Path

from time import perf_counter

import click
import cv2
import numpy as np
from colour import read_image

if typing.TYPE_CHECKING:
    from colour.hints import Any, List, Literal, NDArray, Tuple

from colour.io import convert_bit_depth

__author__ = "Colour Developers"
__copyright__ = "Copyright 2024 Colour Developers"
__license__ = (
    "GNU Affero General Public License v3.0 - "
    "https://www.gnu.org/licenses/agpl-3.0.en.html"
)
__maintainer__ = "Colour Developers"
__email__ = "colour-developers@colour-science.org"
__status__ = "Production"


__all__ = [
    "ROOT_REPOSITORY",
    "URL_BASE",
    "URL_MODEL_FILE_DEFAULT",
    "inference",
    "segmentation",
]


LOGGER = logging.getLogger(__name__)

ROOT_REPOSITORY: str = os.environ.get(
    "COLOUR_SCIENCE__COLOUR_CHECKER_DETECTION__REPOSITORY",
    os.path.join(
        os.path.expanduser("~"),
        ".colour-science",
        "colour-checker-detection",
    ),
)
"""Root of the local repository to download the hosted models to."""

URL_BASE: str = "https://huggingface.co/colour-science/colour-checker-detection-models"
"""URL of the remote repository to download the models from."""

URL_MODEL_FILE_DEFAULT: str = (
    f"{URL_BASE}/resolve/main/models/colour-checker-detection-l-seg.pt"
)
"""URL for the default segmentation model."""


def inference(
    source: str | Path | NDArray,
    model: YOLO,  # noqa: F821 # pyright: ignore
    show: bool = False,
    **kwargs: Any,
) -> List[Tuple[NDArray, NDArray, NDArray]]:
    """
    Run the inference on the provided source.

    Parameters
    ----------
    source
        Source of the image to make predictions on. Accepts all source types
        accepted by the *YOLOv8* model.
    model
        The model to use for the inference.
    show
        Whether to show the inference results on the image.

    Other Parameters
    ----------------
    \\**kwargs : dict, optional
        Keywords arguments for the *YOLOv8* segmentation method.

    Returns
    -------
    :class:`list`
        Inference results.
    """

    data = []

    for result in model(source, show=show, **kwargs):
        show and cv2.waitKey(0) == ord("n")

        if result.boxes is None:
            continue

        if result.masks is None:
            continue

        data_boxes = result.boxes.data
        data_masks = result.masks.data

        for i in range(data_boxes.shape[0]):
            data.append(
                (
                    data_boxes[i, 4].cpu().numpy(),
                    data_boxes[i, 5].cpu().numpy(),
                    data_masks[i].data.cpu().numpy(),
                )
            )

            if np.any(data[-1][-1]):
                LOGGER.debug(
                    'Found a "%s" class object with "%s" confidence.',
                    data[-1][1],
                    data[-1][0],
                )
            else:
                LOGGER.warning("No objects were detected!")

    return data


@click.command()
@click.option(
    "--input",
    required=True,
    type=click.Path(exists=True),
    help="Input file to run the segmentation model on.",
)
@click.option(
    "--output",
    help="Output file to write the segmentation results to.",
)
@click.option(
    "--model",
    "model",
    type=click.Path(exists=True),
    help='Segmentation model file to load. Default to the "colour-science" model '
    'hosted on "HuggingFace". It will be downloaded if not cached already.',
)
@click.option(
    "--show/--no-show",
    default=False,
    help="Whether to show the segmentation results.",
)
@click.option(
    "--logging-level",
    "logging_level",
    default="INFO",
    type=click.Choice(
        ["DEBUG", "INFO", "WARNING", "ERROR", "CRITICAL"], case_sensitive=False
    ),
    help="Set the logging level.",
)
def segmentation(
    input: str,  # noqa: A002
    output: str | None = None,
    model: str | None = None,
    show: bool = False,
    logging_level: Literal["DEBUG", "INFO", "WARNING", "ERROR", "CRITICAL"] = "INFO",
) -> NDArray:
    """
    Run the segmentation model on the given input file and save the results to
    given output file.

    Parameters
    ----------
    input
        Input file to run the segmentation model on.
    output
        Output file to write the segmentation results to.
    model
        Segmentation model file to load. Default to the *colour-science8 model
        hosted on *HuggingFace*. It will be downloaded if not cached already.
    show
        Whether to show the segmentation results.
    logging_level
        Set the logging level.

    Returns
    -------
    :class:`numpy.ndarray`
        Inference results.
    """

    from ultralytics import YOLO  # noqa: PLC0415
    from ultralytics.utils.downloads import download  # noqa: PLC0415

    time_start = perf_counter()

    logging.getLogger().setLevel(getattr(logging, logging_level.upper()))

    if model is None:
        model = os.path.join(ROOT_REPOSITORY, os.path.basename(URL_MODEL_FILE_DEFAULT))
        LOGGER.debug('Using "%s" default model.', model)
        if not os.path.exists(model):
            LOGGER.info('Downloading "%s" model...', URL_MODEL_FILE_DEFAULT)
            download(URL_MODEL_FILE_DEFAULT, ROOT_REPOSITORY)  # pyright: ignore

    if input.endswith((".npy", ".npz")):
        LOGGER.debug('Reading "%s" serialised array...', input)
        source = np.load(input)
    else:
        LOGGER.debug('Reading "%s" image...', input)
        source = convert_bit_depth(
            read_image(input)[..., :3],
            np.uint8.__name__,  # pyright: ignore
        )

    # NOTE: YOLOv8 expects "BGR" arrays.
    results = np.array(inference(source[..., ::-1], YOLO(model), show), dtype=object)

    if output is None:
        output = f"{input}.npz"

    np.savez(output, results=results)

    LOGGER.debug('Total segmentation time: "%s" seconds.', perf_counter() - time_start)

    return results


if __name__ == "__main__":
    logging.basicConfig()

    segmentation()

1	#!/usr/bin/env python
2	"""
3	Colour Checker Detection - Inference
4	====================================
5
6	Define the scripts for colour checker detection using inference based on
7	Ultralytics YOLOv8 machine learning model.
8
9	Warnings
10	--------
11	This script is provided under the terms of the
12	GNU Affero General Public License v3.0 as it uses the Ultralytics YOLOv8
13	API which is incompatible with the BSD-3-Clause.
14	"""
15
16	from __future__ import annotations	1✔
17
18	import logging	1✔
19	import os	1✔
20	import typing	1✔
21
22	if typing.TYPE_CHECKING:
23	from pathlib import Path
24
25	from time import perf_counter	1✔
26
27	import click	1✔
28	import cv2	1✔
29	import numpy as np	1✔
30	from colour import read_image	1✔
31
32	if typing.TYPE_CHECKING:
33	from colour.hints import Any, List, Literal, NDArray, Tuple
34
35	from colour.io import convert_bit_depth	1✔
36
37	__author__ = "Colour Developers"	1✔
38	__copyright__ = "Copyright 2024 Colour Developers"	1✔
39	__license__ = (	1✔
40	"GNU Affero General Public License v3.0 - "
41	"https://www.gnu.org/licenses/agpl-3.0.en.html"
42	)
43	__maintainer__ = "Colour Developers"	1✔
44	__email__ = "colour-developers@colour-science.org"	1✔
45	__status__ = "Production"	1✔
46
47
48	__all__ = [	1✔
49	"ROOT_REPOSITORY",
50	"URL_BASE",
51	"URL_MODEL_FILE_DEFAULT",
52	"inference",
53	"segmentation",
54	]
55
56
57	LOGGER = logging.getLogger(__name__)	1✔
58
59	ROOT_REPOSITORY: str = os.environ.get(	1✔
60	"COLOUR_SCIENCE__COLOUR_CHECKER_DETECTION__REPOSITORY",
61	os.path.join(
62	os.path.expanduser("~"),
63	".colour-science",
64	"colour-checker-detection",
65	),
66	)
67	"""Root of the local repository to download the hosted models to."""	1✔
68
69	URL_BASE: str = "https://huggingface.co/colour-science/colour-checker-detection-models"	1✔
70	"""URL of the remote repository to download the models from."""	1✔
71
72	URL_MODEL_FILE_DEFAULT: str = (	1✔
73	f"{URL_BASE}/resolve/main/models/colour-checker-detection-l-seg.pt"
74	)
75	"""URL for the default segmentation model."""	1✔
76
77
78	def inference(	1✔
79	source: str \| Path \| NDArray,
80	model: YOLO, # noqa: F821 # pyright: ignore
81	show: bool = False,
82	**kwargs: Any,
83	) -> List[Tuple[NDArray, NDArray, NDArray]]:
84	"""
85	Run the inference on the provided source.
86
87	Parameters
88	----------
89	source
90	Source of the image to make predictions on. Accepts all source types
91	accepted by the YOLOv8 model.
92	model
93	The model to use for the inference.
94	show
95	Whether to show the inference results on the image.
96
97	Other Parameters
98	----------------
99	\\**kwargs : dict, optional
100	Keywords arguments for the YOLOv8 segmentation method.
101
102	Returns
103	-------
104	:class:`list`
105	Inference results.
106	"""
107
108	data = []	×
109
110	for result in model(source, show=show, **kwargs):	×
111	show and cv2.waitKey(0) == ord("n")	×
112
113	if result.boxes is None:	×
114	continue	×
115
116	if result.masks is None:	×
117	continue	×
118
119	data_boxes = result.boxes.data	×
120	data_masks = result.masks.data	×
121
122	for i in range(data_boxes.shape[0]):	×
123	data.append(	×
124	(
125	data_boxes[i, 4].cpu().numpy(),
126	data_boxes[i, 5].cpu().numpy(),
127	data_masks[i].data.cpu().numpy(),
128	)
129	)
130
131	if np.any(data[-1][-1]):	×
132	LOGGER.debug(	×
133	'Found a "%s" class object with "%s" confidence.',
134	data[-1][1],
135	data[-1][0],
136	)
137	else:
138	LOGGER.warning("No objects were detected!")	×
139
140	return data	×
141
142
143	@click.command()	1✔
144	@click.option(	1✔
145	"--input",
146	required=True,
147	type=click.Path(exists=True),
148	help="Input file to run the segmentation model on.",
149	)
150	@click.option(	1✔
151	"--output",
152	help="Output file to write the segmentation results to.",
153	)
154	@click.option(	1✔
155	"--model",
156	"model",
157	type=click.Path(exists=True),
158	help='Segmentation model file to load. Default to the "colour-science" model '
159	'hosted on "HuggingFace". It will be downloaded if not cached already.',
160	)
161	@click.option(	1✔
162	"--show/--no-show",
163	default=False,
164	help="Whether to show the segmentation results.",
165	)
166	@click.option(	1✔
167	"--logging-level",
168	"logging_level",
169	default="INFO",
170	type=click.Choice(
171	["DEBUG", "INFO", "WARNING", "ERROR", "CRITICAL"], case_sensitive=False
172	),
173	help="Set the logging level.",
174	)
175	def segmentation(	1✔
176	input: str, # noqa: A002
177	output: str \| None = None,
178	model: str \| None = None,
179	show: bool = False,
180	logging_level: Literal["DEBUG", "INFO", "WARNING", "ERROR", "CRITICAL"] = "INFO",
181	) -> NDArray:
182	"""
183	Run the segmentation model on the given input file and save the results to
184	given output file.
185
186	Parameters
187	----------
188	input
189	Input file to run the segmentation model on.
190	output
191	Output file to write the segmentation results to.
192	model
193	Segmentation model file to load. Default to the *colour-science8 model
194	hosted on HuggingFace. It will be downloaded if not cached already.
195	show
196	Whether to show the segmentation results.
197	logging_level
198	Set the logging level.
199
200	Returns
201	-------
202	:class:`numpy.ndarray`
203	Inference results.
204	"""
205
206	from ultralytics import YOLO # noqa: PLC0415	×
207	from ultralytics.utils.downloads import download # noqa: PLC0415	×
208
209	time_start = perf_counter()	×
210
211	logging.getLogger().setLevel(getattr(logging, logging_level.upper()))	×
212
213	if model is None:	×
214	model = os.path.join(ROOT_REPOSITORY, os.path.basename(URL_MODEL_FILE_DEFAULT))	×
215	LOGGER.debug('Using "%s" default model.', model)	×
216	if not os.path.exists(model):	×
217	LOGGER.info('Downloading "%s" model...', URL_MODEL_FILE_DEFAULT)	×
218	download(URL_MODEL_FILE_DEFAULT, ROOT_REPOSITORY) # pyright: ignore	×
219
220	if input.endswith((".npy", ".npz")):	×
221	LOGGER.debug('Reading "%s" serialised array...', input)	×
222	source = np.load(input)	×
223	else:
224	LOGGER.debug('Reading "%s" image...', input)	×
225	source = convert_bit_depth(	×
226	read_image(input)[..., :3],
227	np.uint8.__name__, # pyright: ignore
228	)
229
230	# NOTE: YOLOv8 expects "BGR" arrays.
231	results = np.array(inference(source[..., ::-1], YOLO(model), show), dtype=object)	×
232
233	if output is None:	×
234	output = f"{input}.npz"	×
235
236	np.savez(output, results=results)	×
237
238	LOGGER.debug('Total segmentation time: "%s" seconds.', perf_counter() - time_start)	×
239
240	return results	×
241
242
243	if __name__ == "__main__":
244	logging.basicConfig()
245
246	segmentation()

colour-science / colour-checker-detection / 17929889887

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous