tests/test_reporting.py - master - sarus/photo-processor

tests/test_reporting.py
209 строк6 KB
otcheskiy
feat: add simple background segmentation and mask quality
01 авг 2026, 12:34
01 авг 2026, 12:3403c5abe
"""Тесты CSV-отчёта и pipeline (этапы 1–2)."""

from __future__ import annotations

import hashlib
from pathlib import Path

from PIL import Image, ImageDraw

from photo_processor.config import load_config
from photo_processor.mask_quality import STAGE2_OK_WARNING
from photo_processor.pipeline import STAGE1_WARNING, run_pipeline
from photo_processor.reporting import CSV_DELIMITER, ReportRow, write_report
from tests.helpers import VALID_CONFIG_TEXT, make_light_scene_jpeg


UNICODE_NAME = "WK+ 5111 c.1 55-14-145 мужская, пластик, темная.jpg"


def test_csv_utf8_bom_and_delimiter(tmp_path: Path) -> None:
    report_path = tmp_path / "processing_report.csv"
    rows = [
        ReportRow(
            timestamp="2026-08-01 10:30:15",
            source_file=UNICODE_NAME,
            output_file="",
            status="SKIPPED",
            original_width="100",
            original_height="80",
            cropped_width="90",
            cropped_height="74",
            warning=STAGE1_WARNING,
            error="",
            processing_time_ms=1,
            mask_method="simple",
        )
    ]
    write_report(report_path, rows)

    raw = report_path.read_bytes()
    assert raw.startswith(b"\xef\xbb\xbf")

    text = report_path.read_text(encoding="utf-8-sig")
    lines = text.strip().splitlines()
    assert CSV_DELIMITER in lines[0]
    assert "original_width" in lines[0]
    assert "cropped_height" in lines[0]
    assert "background_brightness" in lines[0]
    assert "mask_quality_score" in lines[0]
    assert UNICODE_NAME in lines[1]
    assert "SKIPPED" in lines[1]


def test_pipeline_unicode_and_save_cropped(
    tmp_path: Path,
    project_config: Path,
) -> None:
    input_dir = tmp_path / "input"
    output_dir = tmp_path / "output"
    input_dir.mkdir()
    output_dir.mkdir()

    source = make_light_scene_jpeg(
        input_dir / UNICODE_NAME,
        size=(200, 100),
        object_box=(60, 30, 140, 70),
    )
    source_hash = hashlib.sha256(source.read_bytes()).hexdigest()

    config = load_config(project_config)
    assert config.debug.save_cropped is True

    result = run_pipeline(
        input_dir=input_dir,
        output_dir=output_dir,
        config=config,
    )

    run_path = result.run_dir.path
    assert result.stats.total == 1
    assert result.stats.errors == 0
    assert result.stats.skipped + result.stats.reviews == 1

    cropped_path = run_path / "debug" / "cropped" / UNICODE_NAME
    assert cropped_path.is_file()
    with Image.open(cropped_path) as cropped:
        assert cropped.mode == "RGB"
        assert cropped.size == (180, 94)

    report_text = result.run_dir.report_path.read_text(encoding="utf-8-sig")
    assert UNICODE_NAME in report_text
    assert "200" in report_text
    assert "180" in report_text
    assert "simple" in report_text
    assert STAGE2_OK_WARNING in report_text or "REVIEW" in report_text

    assert hashlib.sha256(source.read_bytes()).hexdigest() == source_hash


def test_pipeline_save_cropped_false(tmp_path: Path) -> None:
    input_dir = tmp_path / "input"
    output_dir = tmp_path / "output"
    input_dir.mkdir()
    output_dir.mkdir()
    make_light_scene_jpeg(
        input_dir / "plain.jpg",
        size=(100, 100),
        object_box=(30, 30, 70, 70),
    )

    config_path = tmp_path / "config.yaml"
    text = VALID_CONFIG_TEXT
    for key in (
        "save_cropped",
        "save_raw_mask",
        "save_clean_mask",
        "save_bbox",
        "save_foreground_preview",
    ):
        text = text.replace(f"{key}: true", f"{key}: false")
    config_path.write_text(text + "\n", encoding="utf-8")
    config = load_config(config_path)
    result = run_pipeline(input_dir=input_dir, output_dir=output_dir, config=config)

    cropped_dir = result.run_dir.debug_dir / "cropped"
    assert not cropped_dir.exists() or not any(cropped_dir.iterdir())
    report_text = result.run_dir.report_path.read_text(encoding="utf-8-sig")
    data_line = [line for line in report_text.splitlines() if "plain.jpg" in line][-1]
    parts = data_line.split(";")
    assert parts[1] == "plain.jpg"
    assert parts[2] == ""
    assert parts[3] in {"SKIPPED", "REVIEW"}
    assert "90" in parts


def test_corrupted_then_valid_continues(tmp_path: Path, project_config: Path) -> None:
    input_dir = tmp_path / "input"
    output_dir = tmp_path / "output"
    input_dir.mkdir()
    output_dir.mkdir()

    bad = input_dir / "aaa_broken.jpg"
    bad.write_bytes(b"not-an-image")
    make_light_scene_jpeg(
        input_dir / "bbb_ok.jpg",
        size=(160, 120),
        object_box=(50, 40, 110, 80),
    )

    source_ok = input_dir / "bbb_ok.jpg"
    source_hash = hashlib.sha256(source_ok.read_bytes()).hexdigest()

    config = load_config(project_config)
    result = run_pipeline(input_dir=input_dir, output_dir=output_dir, config=config)

    assert result.stats.total == 2
    assert result.stats.errors == 1
    assert result.stats.skipped + result.stats.reviews == 1

    report_text = result.run_dir.report_path.read_text(encoding="utf-8-sig")
    assert "aaa_broken.jpg" in report_text
    assert "bbb_ok.jpg" in report_text
    assert "ERROR" in report_text

    log_text = result.run_dir.log_path.read_text(encoding="utf-8")
    assert "aaa_broken.jpg" in log_text

    cropped_ok = result.run_dir.debug_dir / "cropped" / "bbb_ok.jpg"
    assert cropped_ok.is_file()
    assert hashlib.sha256(source_ok.read_bytes()).hexdigest() == source_hash
    assert bad.is_file()


def test_pipeline_creates_structure(
    tmp_path: Path,
    project_config: Path,
) -> None:
    input_dir = tmp_path / "input"
    output_dir = tmp_path / "output"
    input_dir.mkdir()
    output_dir.mkdir()

    png_path = input_dir / "plain.png"
    image = Image.new("RGB", (120, 100), (245, 245, 245))
    ImageDraw.Draw(image).rectangle((35, 30, 85, 70), fill=(20, 20, 20))
    image.save(png_path, format="PNG")

    make_light_scene_jpeg(
        input_dir / "other.jpg",
        size=(120, 100),
        object_box=(35, 30, 85, 70),
    )
    (input_dir / "ignore.txt").write_text("nope", encoding="utf-8")

    config = load_config(project_config)
    result = run_pipeline(
        input_dir=input_dir,
        output_dir=output_dir,
        config=config,
    )

    run_path = result.run_dir.path
    assert (run_path / "images").is_dir()
    assert (run_path / "review").is_dir()
    assert (run_path / "debug").is_dir()
    assert (run_path / "processing_report.csv").is_file()
    assert (run_path / "run.log").is_file()
    assert result.stats.total == 2
    assert result.stats.errors == 0