18953976799

Committed 30 Oct 2025 08:21PM UTC coverage: 80.119% (-1.4%) from 81.509%

Build # 18953976799

Build Type

Pull #798

github

Committed by

web-flow

Commit Message

Merge a4477b844 into 50be529a6

Pull Request Pull Request #798: Refactoring

Run Details

858 of 1175 new or added lines in 51 files covered. (73.02%)

23 existing lines in 5 files now uncovered.

2974 of 3712 relevant lines covered (80.12%)

4.57 hits per line

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

78.95

/normcap/detection/ocr/recognize.py

"""Detect OCR tool & language and perform OCR on selected part of image."""

import logging
import sys
import tempfile
import time
from collections.abc import Iterable
from os import PathLike
from pathlib import Path

from PySide6 import QtGui

from normcap.detection.models import DetectionResult, TextDetector, TextType
from normcap.detection.ocr import enhance, tesseract, transformer
from normcap.detection.ocr.models import OEM, PSM, OcrResult, TessArgs

logger = logging.getLogger(__name__)


def _save_image_in_temp_folder(image: QtGui.QImage, postfix: str = "") -> None:
    """For debugging it can be useful to store the cropped image."""
    if logger.getEffectiveLevel() != logging.DEBUG:
        return

    temp_dir = Path(tempfile.gettempdir()) / "normcap"
    temp_dir.mkdir(exist_ok=True)

    now_str = time.strftime("%Y-%m-%d_%H-%M-%S", time.gmtime())
    file_name = f"{now_str}{postfix}.png"

    logger.debug("Save debug image as %s", temp_dir / file_name)
    image.save(str(temp_dir / file_name))


def get_text_from_image(
    languages: str | Iterable[str],
    image: QtGui.QImage,
    tesseract_bin_path: PathLike,
    tessdata_path: PathLike | str | None = None,
    parse: bool = True,
    resize_factor: float | None = None,
    padding_size: int | None = None,
) -> list[DetectionResult]:
    """Apply OCR on selected image section."""
    image = enhance.preprocess(image, resize_factor=resize_factor, padding=padding_size)
    _save_image_in_temp_folder(image, postfix="_enhanced")

    # TODO: Improve handling of tesseract_cmd and tessdata_path
    if sys.platform == "win32" and tessdata_path:
        tessdata_path = tesseract.get_short_path(str(tessdata_path))

    tess_args = TessArgs(
        tessdata_path=tessdata_path,
        lang=languages if isinstance(languages, str) else "+".join(languages),
        oem=OEM.DEFAULT,
        psm=PSM.AUTO,
    )
    logger.debug(
        "Run Tesseract on image of size %s with args:\n%s",
        (image.width(), image.height()),
        tess_args,
    )
    ocr_result_data = tesseract.perform_ocr(
        tesseract_bin_path=tesseract_bin_path, image=image, args=tess_args.as_list()
    )
    result = OcrResult(tess_args=tess_args, words=ocr_result_data, image=image)
    logger.debug("OCR detections:\n%s", ",\n".join(str(w) for w in result.words))

    if not parse:
        return [
            DetectionResult(
                text=result.text,
                text_type=TextType.SINGLE_LINE,
                detector=TextDetector.OCR_RAW,
            )
        ]

    result = transformer.apply(result)
    logger.debug("Parsed text:\n%s", result.parsed)
    text_type = (
        TextType[result.best_scored_transformer.value]
        if result.best_scored_transformer
        else TextType.SINGLE_LINE
    )

    detections = [
        DetectionResult(
            text=s,
            text_type=text_type,
            detector=TextDetector.OCR_PARSED,
        )
        for s in result.parsed
    ]
    return detections

1	"""Detect OCR tool & language and perform OCR on selected part of image."""
2
3	import logging	6✔
4	import sys	6✔
5	import tempfile	6✔
6	import time	6✔
7	from collections.abc import Iterable	6✔
8	from os import PathLike	6✔
9	from pathlib import Path	6✔
10
11	from PySide6 import QtGui	6✔
12
13	from normcap.detection.models import DetectionResult, TextDetector, TextType	6✔
14	from normcap.detection.ocr import enhance, tesseract, transformer	6✔
15	from normcap.detection.ocr.models import OEM, PSM, OcrResult, TessArgs	6✔
16
17	logger = logging.getLogger(__name__)	6✔
18
19
20	def _save_image_in_temp_folder(image: QtGui.QImage, postfix: str = "") -> None:	6✔
21	"""For debugging it can be useful to store the cropped image."""
22	if logger.getEffectiveLevel() != logging.DEBUG:	6✔
23	return	6✔
24
UNCOV 25	temp_dir = Path(tempfile.gettempdir()) / "normcap"	×
UNCOV 26	temp_dir.mkdir(exist_ok=True)	×
27
UNCOV 28	now_str = time.strftime("%Y-%m-%d_%H-%M-%S", time.gmtime())	×
UNCOV 29	file_name = f"{now_str}{postfix}.png"	×
30
UNCOV 31	logger.debug("Save debug image as %s", temp_dir / file_name)	×
UNCOV 32	image.save(str(temp_dir / file_name))	×
33
34
35	def get_text_from_image(	6✔
36	languages: str \| Iterable[str],
37	image: QtGui.QImage,
38	tesseract_bin_path: PathLike,
39	tessdata_path: PathLike \| str \| None = None,
40	parse: bool = True,
41	resize_factor: float \| None = None,
42	padding_size: int \| None = None,
43	) -> list[DetectionResult]:
44	"""Apply OCR on selected image section."""
45	image = enhance.preprocess(image, resize_factor=resize_factor, padding=padding_size)	6✔
46	_save_image_in_temp_folder(image, postfix="_enhanced")	6✔
47
48	# TODO: Improve handling of tesseract_cmd and tessdata_path
49	if sys.platform == "win32" and tessdata_path:	6✔
50	tessdata_path = tesseract.get_short_path(str(tessdata_path))	×
51
52	tess_args = TessArgs(	6✔
53	tessdata_path=tessdata_path,
54	lang=languages if isinstance(languages, str) else "+".join(languages),
55	oem=OEM.DEFAULT,
56	psm=PSM.AUTO,
57	)
58	logger.debug(	6✔
59	"Run Tesseract on image of size %s with args:\n%s",
60	(image.width(), image.height()),
61	tess_args,
62	)
63	ocr_result_data = tesseract.perform_ocr(	6✔
64	tesseract_bin_path=tesseract_bin_path, image=image, args=tess_args.as_list()
65	)
66	result = OcrResult(tess_args=tess_args, words=ocr_result_data, image=image)	6✔
67	logger.debug("OCR detections:\n%s", ",\n".join(str(w) for w in result.words))	6✔
68
69	if not parse:	6✔
NEW 70	return [	×
71	DetectionResult(
72	text=result.text,
73	text_type=TextType.SINGLE_LINE,
74	detector=TextDetector.OCR_RAW,
75	)
76	]
77
78	result = transformer.apply(result)	6✔
79	logger.debug("Parsed text:\n%s", result.parsed)	6✔
80	text_type = (	6✔
81	TextType[result.best_scored_transformer.value]
82	if result.best_scored_transformer
83	else TextType.SINGLE_LINE
84	)
85
86	detections = [	6✔
87	DetectionResult(
88	text=s,
89	text_type=text_type,
90	detector=TextDetector.OCR_PARSED,
91	)
92	for s in result.parsed
93	]
94	return detections	6✔

dynobo / normcap / 18953976799

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