StillHammer
acbe1b4769
Remove ClassGenSystem and obsolete tools from tracking, update course content files with latest revisions. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
96 lines
2.5 KiB
Python
96 lines
2.5 KiB
Python
#!/usr/bin/env python3
|
|
"""
|
|
Test different Y offsets to find the correct one.
|
|
Creates multiple pages with different offsets.
|
|
"""
|
|
|
|
import json
|
|
from pdf2image import convert_from_path
|
|
from reportlab.pdfgen import canvas
|
|
from reportlab.lib.utils import ImageReader
|
|
from reportlab.lib.colors import red, blue
|
|
from reportlab.pdfbase import pdfmetrics
|
|
from reportlab.pdfbase.ttfonts import TTFont
|
|
import tempfile
|
|
import os
|
|
|
|
# Font
|
|
try:
|
|
pdfmetrics.registerFont(TTFont('CF', 'C:/Windows/Fonts/msyh.ttc'))
|
|
font = 'CF'
|
|
except:
|
|
font = 'Helvetica'
|
|
|
|
# Load
|
|
with open('Raw/DevelopChinese/OCR/听力Intro.json') as f:
|
|
data = json.load(f)
|
|
|
|
# Convert
|
|
imgs = convert_from_path('Raw/DevelopChinese/PDF/听力Intro.pdf', dpi=300, first_page=1, last_page=1)
|
|
img = imgs[0]
|
|
w, h = img.size
|
|
print(f'Image: {w}x{h}')
|
|
|
|
# Save image once
|
|
with tempfile.NamedTemporaryFile(suffix='.png', delete=False) as tmp:
|
|
img.save(tmp.name, 'PNG')
|
|
tmp_path = tmp.name
|
|
|
|
# Create PDF with multiple test pages
|
|
c = canvas.Canvas('test_Y_OFFSETS.pdf')
|
|
|
|
# Test different Y offsets
|
|
test_offsets = [-200, -150, -100, -50, 0, 50, 100, 150, 200]
|
|
|
|
for offset in test_offsets:
|
|
c.setPageSize((w, h))
|
|
|
|
# Draw image
|
|
c.drawImage(ImageReader(tmp_path), 0, 0, width=w, height=h)
|
|
|
|
# Draw title
|
|
c.setFont('Helvetica', 20)
|
|
c.setFillColorRGB(1, 0, 0)
|
|
c.drawString(50, h - 50, f'Y OFFSET: {offset:+d} pixels')
|
|
|
|
# Draw boxes with offset
|
|
count = 0
|
|
for d in data['pages'][0]['details']:
|
|
text = d.get('text', '').strip()
|
|
if not text or count >= 30: # Limit to first 30 for clarity
|
|
continue
|
|
bbox = d.get('bbox', [])
|
|
if not bbox or len(bbox) != 4:
|
|
continue
|
|
|
|
xs = [p[0] for p in bbox]
|
|
ys = [p[1] for p in bbox]
|
|
x = min(xs)
|
|
width = max(xs) - x
|
|
|
|
# Apply offset to Y calculation
|
|
y_base = h - max(ys)
|
|
y = y_base + offset
|
|
height = max(ys) - min(ys)
|
|
|
|
c.setStrokeColor(red)
|
|
c.setLineWidth(0.5)
|
|
c.rect(x, y, width, height, stroke=1, fill=0)
|
|
|
|
font_size = max(1, height * 0.7)
|
|
c.setFont(font, font_size)
|
|
c.setFillColor(blue)
|
|
c.drawString(x, y, text)
|
|
|
|
count += 1
|
|
|
|
c.showPage()
|
|
print(f'Page with offset {offset:+d} done')
|
|
|
|
c.save()
|
|
os.unlink(tmp_path)
|
|
|
|
print(f'\nCreated test_Y_OFFSETS.pdf with {len(test_offsets)} pages')
|
|
print('Browse through pages to find which offset aligns boxes correctly!')
|
|
print('Offsets tested:', test_offsets)
|