حوّل PDF إلى نص نظيف

Act as a careful document conversion specialist. Convert the attached PDF file or page image into clean, faithful Markdown that I can paste into any AI chat instead of uploading the original file.

Document language: Arabic
Pages to convert: all pages

Rules:
1. Identical content: convert the text one-to-one.
   - No summarizing, no rephrasing, no translation.
   - Do not remove or add anything, except the elements listed in rule 2.
   - Do not correct spelling or grammar. Keep the original wording, punctuation, quotation marks, diacritics and numbers exactly as printed, including any mistakes.
   - Keep words in other languages (names, technical terms, code) exactly where they appear inside each sentence.
2. Skip: repeated page headers and footers, page numbers, watermarks and logos. Output nothing for them.
3. Structure:
   - Match heading levels (#, ##, ###) to the visual hierarchy of the document.
   - Join lines that only wrap because of the page width into one paragraph. Start a new paragraph only where the document does.
   - Convert bulleted lists to "-" and numbered lists to "1.", keeping any nesting.
   - Use **bold** and *italic* only where the document shows them.
   - Rebuild tables as Markdown tables with the same rows and columns, in the document's reading order.
   - Put code, commands and file paths in code fences or inline code.
   - Keep hyperlinks as Markdown links when the address is visible or embedded.
   - Replace every image, chart or signature with a one-line description of it in square brackets, written in the document's language.
4. Unreadable text: if you cannot read a word with confidence, write [?] in its place. Never guess or invent missing text.

Output: only the Markdown, inside one code block, with nothing before or after it.