PDF2PPt

PDF2PPt converts slide-oriented PDF files into editable PowerPoint decks using PDF rasterization, Gemini OCR, OpenCV inpainting, and python-pptx.

Features

• Converts selected PDF page ranges into PowerPoint slide decks.
• Provides both a Tkinter desktop interface and a command-line workflow.
• Uses Gemini to extract text and bounding boxes from rendered page images.
• Removes detected source text from slide images before rebuilding editable text layers.
• Saves partial output when at least one page is processed before a later page fails.
• Can keep temporary page images for debugging through the CLI.

Requirements

• Python 3.10 or newer.
• A Gemini API key available as GEMINI_API_KEY.
• Python packages listed in requirements.txt: google-generativeai, opencv-python, Pillow, PyMuPDF, python-dotenv, and python-pptx.
• On Linux, a Tkinter-capable Python installation is required for the desktop GUI.

Setup

1. Create and activate a virtual environment.
2. Install dependencies:

pip install -r requirements.txt

3. Copy .env.example to a local .env file.
4. Set the Gemini API key:

GEMINI_API_KEY=your-real-key

The GUI can also create or update the local environment file when the API key is entered and saved from the application.

Usage

Run the desktop application:

python gui.py

Run the command-line converter:

python main.py input.pdf --output output.pptx --start-page 1 --end-page 5

CLI options:

• --output: output PowerPoint path. Defaults to output.pptx.
• --start-page: first page to process, using 1-based numbering. Defaults to 1.
• --end-page: last page to process, using 1-based numbering.
• --api-key: Gemini API key override. When omitted, the converter uses GEMINI_API_KEY from the local environment file.
• --keep-temp: keeps generated page images and cleaned images beside the output file for debugging.

Project Structure

• main.py: command-line entry point and end-to-end conversion pipeline.
• gui.py: Tkinter desktop interface for selecting a PDF, output folder, API key, and page range.
• pdf_utils.py: PDF page counting and page-to-image rendering with PyMuPDF.
• ocr_service.py: Gemini configuration, OCR prompting, and bounding-box parsing.
• image_processor.py: OpenCV-based text mask creation and inpainting.
• ppt_generator.py: PowerPoint slide generation with background images and editable text boxes.
• requirements.txt: Python runtime dependencies.
• .env.example: sample environment file containing GEMINI_API_KEY.
• pdf2pptsoft.spec: PyInstaller specification for packaging the GUI application.

摘要

PDF2PPt 是一個將簡報型 PDF 轉成可編輯 PowerPoint 檔案的 Python 工具。 主要流程會先把 PDF 頁面轉成圖片，再使用 Gemini 擷取文字與位置資訊。 程式會用 OpenCV 嘗試移除原圖片上的文字，接著用 python-pptx 重建投影片。 使用前需安裝 requirements.txt 內的套件，並在本機環境檔設定 GEMINI_API_KEY。 桌面版可執行 python gui.py，命令列版可執行 python main.py input.pdf --output output.pptx。 若轉換中途發生 OCR 或 API 錯誤，只要已有成功處理的頁面，程式仍會先輸出部分結果。