FenScribe - A Smart PDF Layout Optimizer

Automatically detect and remove blank areas from images or PDFs, for reducing printing costs.

Home Page

Key Features

• Detect and remove long blank regions (by brightness threshold).[New - We may now disregard marginal scan edges! (in advanced option)]
• Split pages into content segments(pre line) and save them as individual images.
• Optional Word export: insert segments into a .docx template.
• [New Update!]Automatic image width fitting for Word documents (python-docx), reducing the need for post-processing macros.

Requirements

This project was developed on Windows and assumes Microsoft Word is available for macro injection features.

Python packages:

pip install PyMuPDF Pillow python-docx tkinterdnd2 pywin32

• pywin32 is required only if you want the program to inject/run VBA macros in Word (Windows only).
• If you only need image output or python-docx insertion, Word and pywin32 are optional.

---

Quick Start

1. Install dependencies (see Requirements).
2. Run the GUI:

python gui.pyw

Or on Windows you can double-click gui.pyw.

Typical Workflow

1. Select a PDF file, single image, or a folder containing images.
2. Tune parameters in Step 2 (threshold, DPI, min_height, blank_height).
3. Choose a .docx template in Step 3 if you want Word output.
4. Start processing. Output files:

• Images / segments are saved under _temp/<input>_segments.
• If Word insertion is enabled, the final document is written to __Output/output_<basename>.docx.

Configuration Parameters

• threshold (20–255): Brightness threshold for blank-line detection. Larger values are more conservative.
• dpi (72–600): Resolution used when converting PDF pages to images.
• min_height: Minimum segment height (px) to keep a segment.
• blank_height: Number of consecutive blank rows that separate paragraphs.
• lr_threshold_enabled: Enable left/right margin trimming for blank detection only.
• lr_threshold_percent: Percent width on each side to ignore when detecting blanks.
• no_insert_word: When set, only images are saved and Word insertion is skipped.

Macros and Word Automation

• The app can optionally inject and run a VBA macro (from the _Macros folder) after producing the .docx file. That requires Microsoft Word and pywin32.
• For macro injection to work, enable Trust access to the VBA project object model in Word: File → Options → Trust Center → Trust Center Settings → Macro Settings → Trust access to the VBA project object model.
• Note: the program now attempts to auto-fit images to the document/column width using python-docx. The macro remains available as an optional post-processing step but is no longer strictly required for common cases.

Troubleshooting

• If Word macro injection fails with an access error, verify the VBA trust setting described above.
• If Word is not installed or pywin32 is missing, use the Advanced option Do not insert Word to only save images.
• For scanned documents with slanted pages, run a deskew/preprocessing step first.

License

MIT License

Development Notes

• This version introduces batch image insertion, automatic width-fit using python-docx, and progress throttling to improve performance and UX.
• Contributions and bug reports are welcome.