first commit

image_test/e2e_test.py

@@ -0,0 +1,150 @@
+"""端到端验证：process_image 输出 vs 实际写入 docx 的图片字节。
+
+目的:
+    用户反馈"从文档复制出来的图底部有 1/4 空白"，但独立测试 process_image
+    输出已是紧致裁剪。本脚本走完整 sign_word 链路，把生成的 docx 解开来看
+    word/media/ 下真实的图片字节，对比 process_image 的输出，定位空白来源:
+
+    - 若字节完全一致 → 空白来自 Word 段落渲染（行高/对齐），不是图片本身
+    - 若字节不一致   → 我们的链路在某个环节给图片加了 padding，需进一步定位
+
+使用:
+    python image_test/e2e_test.py
+"""
+import io
+import sys
+import zipfile
+from pathlib import Path
+
+BASE_DIR = Path(__file__).resolve().parent
+PROJECT_ROOT = BASE_DIR.parent
+INPUT_IMAGE = BASE_DIR / "input" / "test1.jpg"
+OUT_DIR = BASE_DIR / "output" / "e2e"
+DOCX_OUT = OUT_DIR / "signed.docx"
+
+
+def _ensure_project_on_path():
+    if str(PROJECT_ROOT) not in sys.path:
+        sys.path.insert(0, str(PROJECT_ROOT))
+
+
+def _make_test_docx(path):
+    """构造一份带 SET USER_test 域代码的 docx，正文段落使用偏大行高。
+
+    段落行高 2cm（lineRule=exact），模拟用户文档里的固定行高场景。
+    """
+    from docx import Document
+    from docx.shared import Cm, Pt
+    from docx.enum.text import WD_LINE_SPACING
+    from lib import field_codes
+
+    doc = Document()
+    p = doc.add_paragraph()
+    p.paragraph_format.line_spacing_rule = WD_LINE_SPACING.EXACTLY
+    p.paragraph_format.line_spacing = Cm(2.0)
+    p.add_run("正文前 ")
+
+    runs = field_codes.build_set_field_runs("test")
+    for r in runs:
+        p._element.append(r)
+
+    p.add_run(" 正文中 ")
+    p.add_run("正文后")
+    doc.save(str(path))
+
+
+def _image_stats(data):
+    from PIL import Image, ImageChops
+    img = Image.open(io.BytesIO(data))
+    if img.mode != 'RGBA':
+        img = img.convert('RGBA')
+    r, g, b, a = img.split()
+    luma = Image.merge('RGB', (r, g, b)).convert('L')
+    alpha_mask = a.point(lambda p: 255 if p > 30 else 0)
+    luma_mask = luma.point(lambda p: 255 if p < 210 else 0)
+    combined = ImageChops.darker(alpha_mask, luma_mask)
+    total = img.size[0] * img.size[1] or 1
+    return {
+        'size': img.size,
+        'mode': img.mode,
+        'content_pct': 100.0 * sum(combined.histogram()[1:]) / total,
+        'alpha_pct': 100.0 * sum(alpha_mask.histogram()[1:]) / total,
+        'bbox': combined.getbbox(),
+        'sha256': __import__('hashlib').sha256(data).hexdigest()[:16],
+    }
+
+
+def main():
+    _ensure_project_on_path()
+
+    if not INPUT_IMAGE.exists():
+        print(f"测试图片不存在: {INPUT_IMAGE}")
+        sys.exit(1)
+
+    OUT_DIR.mkdir(parents=True, exist_ok=True)
+    src_docx = OUT_DIR / "input.docx"
+    _make_test_docx(src_docx)
+
+    with open(src_docx, 'rb') as f:
+        docx_data = f.read()
+    with open(INPUT_IMAGE, 'rb') as f:
+        image_bytes = f.read()
+
+    from lib.image_processor import process_image
+    from lib.word_signer import sign_word
+
+    processed = process_image(
+        image_bytes, is_signature=True,
+        noise_level=0, rotate_min=0, rotate_max=0,
+        log_source='e2e测试',
+    )
+
+    params = {
+        'docx_name': 'input.docx',
+        'match_mode': 'upload',
+        'stamps': [{
+            'marker': 'USER_test',
+            'image_base64': __import__('base64').b64encode(image_bytes).decode(),
+            'image_filename': 'test1.jpg',
+            'is_signature': True,
+        }],
+        'use_global_height': True,
+        'height': 2.5,
+    }
+
+    result, err = sign_word(docx_data, params, log_source='e2e测试', trace_id='e2e')
+    if err:
+        print(f"sign_word 失败: {err}")
+        sys.exit(1)
+
+    with open(DOCX_OUT, 'wb') as f:
+        f.write(result['data'])
+
+    print(f"已生成: {DOCX_OUT}")
+    print(f"warnings: {result.get('warnings', [])}")
+    print(f"success_count: {result.get('success_count', 0)}")
+    print()
+
+    print("=== process_image 直出 ===")
+    s1 = _image_stats(processed)
+    print(f"  尺寸={s1['size']} 内容={s1['content_pct']:.1f}% "
+          f"alpha={s1['alpha_pct']:.1f}% bbox={s1['bbox']} sha={s1['sha256']}")
+    print()
+
+    print("=== docx 内 word/media/ 实际图片 ===")
+    with zipfile.ZipFile(DOCX_OUT, 'r') as z:
+        medias = sorted(n for n in z.namelist() if n.startswith('word/media/'))
+        if not medias:
+            print("  未找到任何图片！")
+            return
+        for name in medias:
+            data = z.read(name)
+            s2 = _image_stats(data)
+            same = "一致" if data == processed else "不一致"
+            print(f"  [{name}] 尺寸={s2['size']} 内容={s2['content_pct']:.1f}% "
+                  f"alpha={s2['alpha_pct']:.1f}% bbox={s2['bbox']} sha={s2['sha256']} 字节={same}")
+            print(f"    原始 {len(processed)} bytes vs docx {len(data)} bytes")
+
+
+if __name__ == '__main__':
+    main()