img_2_pdf

This repository now contains a practical document-processing toolkit focused on one main goal:

1. Accept camera captures or imported files/folders.
2. Run document preprocessing and cleanup.
3. Export a clean merged PDF.
4. Add optional OCR later as a controlled extension.

Current Main App

Run:

python camscan_hybrid_tool.py

camscan_hybrid_tool.py is the current unified variant for your workflow.

New Unified App (In Progress)

A new package-based unified application is being built under src/uniscan.

Run (from repository root):

set PYTHONPATH=src
python -m uniscan.cli

One-script launcher (recommended on Windows):

.\run_uniscan.cmd

Or after installation:

uniscan

Quick workflow (Office Lens style):

1. Open tab 1. Import (main mode) and load files/folder, or use 2. Scan for camera capture.
2. Import and Scan are acquisition-only: they load/capture raw pages into session.
3. App switches to 3. Review: reorder, rotate, deskew, auto crop, manual corners, and side-by-side Before/After preview.
4. All processing controls are in Review: quick dropdowns (Lens, Post, Preset), Advanced... popup sliders, and an Apply all changes to all files scope checkbox.
5. Review uses lightweight previews by default (Full HD); uncheck it to work directly with full-resolution previews.
6. Auto Crop... opens a page browser with auto-detect and manual corner editing for one page or all pages.
7. Open 4. Export, choose OCR engine if needed, then save merged PDF or image files.

Current implemented modules in this new app:

1. Capture: live preview, single capture, burst capture, camera configuration
2. Import: folder/files (multi-select)/PDF import into one session
3. Pages: page list management (preview, reorder, select/delete)
4. Export: merged PDF and separate image export

Implementation notes:

1. Session pages are disk-backed (uniscan cache) with lazy reads to reduce RAM usage on large batches.
2. Pages review now shows Before/After preview for preprocessing visibility.
3. Capture/import keep originals first; processing is only applied from Review.
4. Export tab supports OCR engine selection with dependency status checks.
5. Import supports multi-file selection and background loading.
6. Import order is preserved end-to-end: folder order, document page order, and mixed import order are kept as selected.
7. Searchable PDF is currently wired for pytesseract, OCRmyPDF, and PyMuPDF OCR.
8. PaddleOCR, Surya, and MinerU are available as selectable OCR backends with readiness checks (searchable-PDF wiring pending).

What The App Does

camscan_hybrid_tool.py supports three source modes:

1. Import folder
2. Import files
3. Camera capture

Processing features:

1. Document detection and perspective extraction using third-party logic from camscan_suhren: camscan.scanner.main
2. Postprocessing effects from camscan_suhren: None, Sharpen, Grayscale, Black and White
3. Optional two-page split (left/right) for book-like captures.
4. Merged PDF export from all source modes.
5. Quality profiles (Fast, Balanced, Best quality) for practical output control.

Note:

1. OCR is intentionally left as the next stage and is not active in camscan_hybrid_tool.py yet.

Processing Pipeline

For folder/files mode:

1. Load input images (and PDF pages if PDF files are provided and pymupdf is installed).
2. Optionally detect and extract document contour.
3. Apply selected postprocessing function.
4. Optionally split each page into left/right halves.
5. Convert processed pages into one merged PDF.

For camera mode:

1. Capture N shots from selected camera index.
2. Wait configured delay between shots.
3. Apply the same processing pipeline as above.
4. Export merged PDF.

Setup

Recommended Python:

1. Python 3.11+

Install dependencies:

pip install opencv-python numpy pillow img2pdf pymupdf

Optional OCR dependencies in the new app:

pip install pytesseract pypdf ocrmypdf paddleocr pymupdf

Also install CLI/system tools where needed:

1. Tesseract OCR engine in PATH for pytesseract and PyMuPDF OCR mode.
2. ocrmypdf command in PATH for OCRmyPDF mode.

Experimental engine packages:

1. Surya (or marker package path that bundles Surya OCR).
2. MinerU (mineru or magic_pdf package).

If you plan to use legacy scripts with OCR, install additionally:

pip install ocrmypdf pypdf

External OCR tools for legacy OCR scripts:

1. Tesseract OCR
2. Ghostscript
3. qpdf
4. Poppler (pdftoppm, pdfunite) for only_tesseract.py in PDF mode

How To Run

Main app:

python camscan_hybrid_tool.py

Alternative app (images/file + optional OCR already integrated):

python unified_pdf_tool.py

Legacy apps (kept for reference/fallback):

python fast.py
python img_2_pdf.py
python only_tesseract.py
python "prepare pdf to tesseract.py"

Script Map

File	Role
camscan_hybrid_tool.py	Main hybrid app (camera + files/folder) using third-party processing logic from camscan_suhren
unified_pdf_tool.py	Unified app for folder/file workflows with optional OCR path
fast.py	OCR-focused GUI with batch PDF support
img_2_pdf.py	Photo-to-PDF app with OpenCV preprocessing and optional OCR
only_tesseract.py	OCR pipeline using direct tesseract.exe calls
imgs_and_pdfs_ocr_fast_STABLE.py	Stable previous OCR GUI version
prepare pdf to tesseract.py	PDF conditioning helper before OCR
camscan_suhren/	Third-party camera scanner project used as source of preprocessing logic

Known Limitations

1. OCR in camscan_hybrid_tool.py is not enabled yet (planned next).
2. Camera mode is shot-based capture (not a full continuous preview UI).
3. PDF import in hybrid mode requires pymupdf.

Troubleshooting

1. Error about missing camscan modules: Ensure folder camscan_suhren exists directly inside repo root.
2. Cannot open camera: Check camera index and close other apps using webcam.
3. PDF import error: Install pymupdf (pip install pymupdf).

Short Roadmap

1. Add optional OCR to camscan_hybrid_tool.py with toggle and language setting.
2. Add stronger camera UX (preview/retake/selection before export).
3. Add job queue for large folder batches.
4. Add tests for hybrid pipeline stages.