StillHammer
a61a32b57f
- Move all Python scripts to tools/ directory - Move documentation files to docs/ directory - Create exams/ and homework/ directories for future use - Remove temporary test file (page1_preview.png) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
111 lines
2.9 KiB
Python
111 lines
2.9 KiB
Python
#!/usr/bin/env python3
|
|
"""
|
|
Draw actual quadrilaterals (not rectangles) to match rotated text bboxes.
|
|
"""
|
|
|
|
import json
|
|
from pathlib import Path
|
|
from PIL import Image
|
|
from pdf2image import convert_from_path
|
|
from reportlab.pdfgen import canvas
|
|
from reportlab.lib.utils import ImageReader
|
|
from reportlab.pdfbase import pdfmetrics
|
|
from reportlab.pdfbase.ttfonts import TTFont
|
|
from reportlab.lib.colors import red, blue
|
|
import tempfile
|
|
import os
|
|
|
|
# Register Chinese font
|
|
try:
|
|
pdfmetrics.registerFont(TTFont('ChineseFont', 'C:/Windows/Fonts/msyh.ttc'))
|
|
font_name = 'ChineseFont'
|
|
except:
|
|
font_name = 'Helvetica'
|
|
|
|
# Paths
|
|
json_path = "Raw/DevelopChinese/OCR/听力Intro.json"
|
|
pdf_path = "Raw/DevelopChinese/PDF/听力Intro.pdf"
|
|
output_path = "test_QUADRILATERAL.pdf"
|
|
|
|
# Load JSON
|
|
with open(json_path) as f:
|
|
data = json.load(f)
|
|
|
|
# Convert PDF to image at 300 DPI (same as OCR)
|
|
print("Converting PDF to image at 300 DPI...")
|
|
images = convert_from_path(pdf_path, dpi=300)
|
|
img = images[0]
|
|
img_width, img_height = img.size
|
|
print(f"Image size: {img_width} x {img_height}")
|
|
|
|
# Create PDF
|
|
c = canvas.Canvas(output_path)
|
|
c.setPageSize((img_width, img_height))
|
|
|
|
# Draw image
|
|
with tempfile.NamedTemporaryFile(suffix='.png', delete=False) as tmp:
|
|
img.save(tmp.name, 'PNG')
|
|
tmp_path = tmp.name
|
|
|
|
c.drawImage(ImageReader(tmp_path), 0, 0, width=img_width, height=img_height)
|
|
|
|
# Get page data
|
|
page_data = data['pages'][0]
|
|
details = page_data.get('details', [])
|
|
|
|
text_count = 0
|
|
for detail in details:
|
|
text = detail.get('text', '').strip()
|
|
if not text:
|
|
continue
|
|
|
|
bbox = detail.get('bbox', [])
|
|
if not bbox or len(bbox) != 4:
|
|
continue
|
|
|
|
# bbox is [[x0,y0], [x1,y1], [x2,y2], [x3,y3]] in image coordinates
|
|
# Convert Y (PDF has origin at bottom, image at top)
|
|
points_pdf = []
|
|
for point in bbox:
|
|
x = point[0]
|
|
y = img_height - point[1] # Flip Y
|
|
points_pdf.append((x, y))
|
|
|
|
# Draw the actual quadrilateral
|
|
c.setStrokeColor(red)
|
|
c.setLineWidth(0.5)
|
|
|
|
# Draw lines connecting all 4 points
|
|
path = c.beginPath()
|
|
path.moveTo(points_pdf[0][0], points_pdf[0][1])
|
|
path.lineTo(points_pdf[1][0], points_pdf[1][1])
|
|
path.lineTo(points_pdf[2][0], points_pdf[2][1])
|
|
path.lineTo(points_pdf[3][0], points_pdf[3][1])
|
|
path.close()
|
|
c.drawPath(path, stroke=1, fill=0)
|
|
|
|
# For text, use bottom-left corner (point 3)
|
|
x_text = points_pdf[3][0]
|
|
y_text = points_pdf[3][1]
|
|
|
|
# Calculate height from bbox
|
|
min_y = min(p[1] for p in points_pdf)
|
|
max_y = max(p[1] for p in points_pdf)
|
|
height = max_y - min_y
|
|
|
|
font_size = max(1, height * 0.7)
|
|
c.setFont(font_name, font_size)
|
|
c.setFillColor(blue)
|
|
c.drawString(x_text, y_text, text)
|
|
|
|
text_count += 1
|
|
|
|
print(f"Drew {text_count} quadrilaterals")
|
|
|
|
c.showPage()
|
|
c.save()
|
|
os.unlink(tmp_path)
|
|
|
|
print(f"\nDone! Check {output_path}")
|
|
print("Red quadrilaterals should now EXACTLY match the text areas!")
|