تطوير برنامج Python لاستخراج بيانات من ملفات PDF العربية باستخدام OCR إلى Excel

I have roughly 500 Arabic-language PDFs (books and administrative letters) stored in a single folder on my Windows machine. Some are digitally generated while others are scanned images, so reliable, high-accuracy OCR is essential; ordinary solutions have failed because the text quality is only medium and many pages need careful pre-processing before recognition.

What I need the script to do:

• Read every PDF in the folder in one pass.
• For each file pull out four data points—رقم الكتاب (book number), تاريخ الكتاب (in the existing DD-MM-YYYY format), موضوع الكتاب (subject) and اسم الدائرة أو الجهة المرسلة (issuing department).
• Append these fields to an existing Excel workbook, filling only the blank rows so nothing already entered is duplicated. Each new row must also contain a working hyperlink that opens the corresponding PDF.
• Run fully offline on Windows; no cloud calls or external APIs.
• Handle the mixed batch gracefully: where the PDF already contains selectable text, extract it directly; where it is an image, trigger high-quality OCR after image enhancement (noise removal, skew correction, contrast boost, etc.).
• Deliver the full, well-documented Python source code, ready to execute, along with a concise “how to run” guide.

For testing, I will supply the Excel file, a subset of real PDFs and sample expected output. I will consider the job complete when the script processes the test set end-to-end without manual intervention and shows near-perfect field accuracy.