LocalFoodAI_lanfr144/scratch/test_clean_save.py

import os
from markdown_pdf import MarkdownPdf, Section
import fitz

docs_dir = r"c:/Users/lanfr144/Documents/DOPRO1/Antigravity/Food/docs"
md_file = os.path.join(docs_dir, "Installation_Guide.md")

with open(md_file, "r", encoding="utf-8") as f:
    md_content = f.read()

pdf = MarkdownPdf()
section = Section(md_content, paper_size="A4", root=docs_dir)
pdf.add_section(section)

# Try saving WITH clean=True
doc = pdf._make_doc()
doc.save("scratch/test_clean_on.pdf", clean=True, garbage=3, deflate=True)
doc.close()

# Try saving WITHOUT clean=True (only garbage=3, deflate=True)
doc2 = fitz.open("scratch/test_clean_on.pdf")
print("WITH clean=True, Page 2 text length:", len(doc2[1].get_text()))
doc2.close()

pdf2 = MarkdownPdf()
section2 = Section(md_content, paper_size="A4", root=docs_dir)
pdf2.add_section(section2)
doc3 = pdf2._make_doc()
doc3.save("scratch/test_clean_off.pdf", clean=False, garbage=3, deflate=True)
doc3.close()

doc4 = fitz.open("scratch/test_clean_off.pdf")
print("WITHOUT clean=True, Page 2 text length:", len(doc4[1].get_text()))
doc4.close()