doc-page-extractor 0.0.8__py3-none-any.whl → 0.0.9__py3-none-any.whl

doc_page_extractor/overlap.py

@@ -88,7 +88,7 @@ def regroup_lines(origin_fragments: list[OCRFragment]) -> list[OCRFragment]:
         x2 = max(x2, x)
         y2 = max(y2, y)
 
-    if len(proto_texts_len) == 0:
+    if proto_texts_len == 0:
       continue
 
     fragments.append(OCRFragment(
@@ -115,17 +115,19 @@ def _split_fragments_into_groups(fragments: list[OCRFragment]) -> Generator[list
     height = y2 - y1
     median = (y1 + y2) / 2.0
 
+    if height == 0:
+      continue
+
     if len(group) > 0:
       next_mean_median = (sum_median + median) / (len(group) + 1)
       next_mean_height = (sum_height + height) / (len(group) + 1)
 
-      if next_mean_height > 0:
-        deviation_rate = abs(median - next_mean_median) / next_mean_height
-        if deviation_rate > max_deviation_rate:
-          yield group
-          group = []
-          sum_height = 0.0
-          sum_median = 0.0
+      deviation_rate = abs(median - next_mean_median) / next_mean_height
+      if deviation_rate > max_deviation_rate:
+        yield group
+        group = []
+        sum_height = 0.0
+        sum_median = 0.0
 
     group.append(fragment)
     sum_height += height

doc_page_extractor/plot.py

@@ -1,20 +1,73 @@
 from typing import Iterable
 from PIL import ImageDraw
+from PIL.ImageFont import load_default, FreeTypeFont
 from PIL.Image import Image
 from .types import Layout, LayoutClass
+from .rectangle import Point
 
 _FRAGMENT_COLOR = (0x49, 0xCF, 0xCB) # Light Green
+_Color = tuple[int, int, int]
 
-def plot(image: Image, layouts: Iterable[Layout]):
+def plot(image: Image, layouts: Iterable[Layout]) -> None:
+  layout_font = load_default(size=35)
+  fragment_font = load_default(size=25)
   draw = ImageDraw.Draw(image, mode="RGBA")
+
+  def _draw_number(position: Point, number: int, font: FreeTypeFont, bold: bool, color: _Color) -> None:
+    nonlocal draw
+    x, y = position
+    text = str(object=number)
+    width = len(text) * font.size
+    offset = round(font.size * 0.15)
+
+    for dx, dy in _generate_delta(bold):
+      draw.text(
+        xy=(x + dx - width - offset, y + dy),
+        text=text,
+        font=font,
+        fill=color,
+      )
+
   for layout in layouts:
-    draw.polygon([p for p in layout.rect], outline=_layout_color(layout), width=5)
+    draw.polygon(
+      xy=[p for p in layout.rect],
+      outline=_layout_color(layout),
+      width=5,
+    )
 
   for layout in layouts:
-    for fragments in layout.fragments:
-      draw.polygon([p for p in fragments.rect], outline=_FRAGMENT_COLOR, width=3)
+    for fragment in layout.fragments:
+      draw.polygon(
+        xy=[p for p in fragment.rect],
+        outline=_FRAGMENT_COLOR,
+        width=3,
+      )
+      _draw_number(
+        position=fragment.rect.lt,
+        number=fragment.order + 1,
+        font=fragment_font,
+        bold=False,
+        color=_FRAGMENT_COLOR,
+      )
+
+  for i, layout in enumerate(layouts):
+    _draw_number(
+      position=layout.rect.lt,
+      number=i + 1,
+      font=layout_font,
+      bold=True,
+      color=_layout_color(layout),
+    )
+
+def _generate_delta(bold: bool):
+  if bold:
+    for dx in range(-1, 2):
+      for dy in range(-1, 2):
+        yield dx, dy
+  else:
+    yield 0, 0
 
-def _layout_color(layout: Layout) -> tuple[int, int, int]:
+def _layout_color(layout: Layout) -> _Color:
   cls = layout.cls
   if cls == LayoutClass.TITLE:
     return (0x0A, 0x12, 0x2C) # Dark

{doc_page_extractor-0.0.8.dist-info → doc_page_extractor-0.0.9.dist-info}/METADATA

@@ -1,6 +1,6 @@
-Metadata-Version: 2.2
+Metadata-Version: 2.4
 Name: doc-page-extractor
-Version: 0.0.8
+Version: 0.0.9
 Summary: doc page extractor can identify text and format in images and return structured data.
 Home-page: https://github.com/Moskize91/doc-page-extractor
 Author: Tao Zeyu
@@ -19,6 +19,7 @@ Dynamic: author-email
 Dynamic: description
 Dynamic: description-content-type
 Dynamic: home-page
+Dynamic: license-file
 Dynamic: requires-dist
 Dynamic: summary
 

{doc_page_extractor-0.0.8.dist-info → doc_page_extractor-0.0.9.dist-info}/RECORD

@@ -5,8 +5,8 @@ doc_page_extractor/extractor.py,sha256=RPoqc_UtbHe4BoQvtaJxWuKyHwoDkK_fuYvuyA31p
 doc_page_extractor/layoutreader.py,sha256=BdC4oPbtpXoLmYhjuSFrKn6SNoT2zWw_gi95sGAUwrk,4031
 doc_page_extractor/ocr.py,sha256=tEalIlEz_80bJN8BXnHWi3UCMoPdJ5ylgkmaBV1dzVs,5201
 doc_page_extractor/ocr_corrector.py,sha256=RfRA1jESEuqC8_a2kUEvHblT_B4xBjE0OApLMl1JiRg,3917
-doc_page_extractor/overlap.py,sha256=eACIljYa-5cR3jD0fo7IAo_aDC4bl-2tFwr7ikEyK84,5339
-doc_page_extractor/plot.py,sha256=R8hbmdGjtw2pAH1lJkGc7Qbis4aRaaAkrkEo6WjbqyM,1378
+doc_page_extractor/overlap.py,sha256=Nx7Z6Ih_7SBE-JAjqlMx51PYiciXomPxIOAS37x6fco,5327
+doc_page_extractor/plot.py,sha256=4uibjS_x1SyEyjaJJd0YsBbzkgldDOCct4Ry2cOhdXU,2556
 doc_page_extractor/raw_optimizer.py,sha256=1KghECq_rJwuZZITTLQnGTKYivFKg_qDvMLN9g17sks,2844
 doc_page_extractor/rectangle.py,sha256=Tp__NPiY6JlYwYxejST7BUXhv_bl8tkmDXi4JgHCK6E,1539
 doc_page_extractor/rotation.py,sha256=QCZ-HqfDxIhnQw8KRHki2myj6-UusvNY7Mpjsu-wI-4,4334
@@ -24,10 +24,10 @@ doc_page_extractor/onnxocr/predict_rec.py,sha256=qQrCs5jzCf5PYp-iEKJ53pcx_xRoJdJ
 doc_page_extractor/onnxocr/predict_system.py,sha256=yoqXunAsoboPsWe7qQjvQf2_SMW1T1QMriEoiGdX3BM,2721
 doc_page_extractor/onnxocr/rec_postprocess.py,sha256=qZt5Ripal7z9hniKq5e7azOkD9e6NR1ylWpRpznhweg,29556
 doc_page_extractor/onnxocr/utils.py,sha256=AQoHgQyv-jpPo4BsVzq3r7_ze698EZ-a7LJobm2fwUI,1864
+doc_page_extractor-0.0.9.dist-info/licenses/LICENSE,sha256=TfPDBt3ar0uv_f9cqCDMZ5rIzW3CY8anRRd4PkL6ejs,34522
 tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 tests/test_history_bus.py,sha256=WaCUW3U75SESMcLq0f5FKnpVUVRDvmfxLFE7Zo83e48,2517
-doc_page_extractor-0.0.8.dist-info/LICENSE,sha256=TfPDBt3ar0uv_f9cqCDMZ5rIzW3CY8anRRd4PkL6ejs,34522
-doc_page_extractor-0.0.8.dist-info/METADATA,sha256=FRVMnS_BAr5U5Mnltx9zvDrnPiBIQ0MXi7RVzQ2_8qo,2203
-doc_page_extractor-0.0.8.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
-doc_page_extractor-0.0.8.dist-info/top_level.txt,sha256=ErNybD_lBzAmw8mVBAK4htsAH_hp14jioZVex-tUqvM,25
-doc_page_extractor-0.0.8.dist-info/RECORD,,
+doc_page_extractor-0.0.9.dist-info/METADATA,sha256=vwcUU9DE9rUhpbz1UigpDN3YhNnNSRbHGAyJroHPV7Q,2225
+doc_page_extractor-0.0.9.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
+doc_page_extractor-0.0.9.dist-info/top_level.txt,sha256=ErNybD_lBzAmw8mVBAK4htsAH_hp14jioZVex-tUqvM,25
+doc_page_extractor-0.0.9.dist-info/RECORD,,

{doc_page_extractor-0.0.8.dist-info → doc_page_extractor-0.0.9.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (75.8.0)
+Generator: setuptools (78.1.0)
 Root-Is-Purelib: true
 Tag: py3-none-any