{"id":78958,"date":"2026-04-15T14:06:45","date_gmt":"2026-04-15T08:36:45","guid":{"rendered":"https:\/\/2thenew.xyz\/blog\/?p=78958"},"modified":"2026-04-22T11:27:42","modified_gmt":"2026-04-22T05:57:42","slug":"building-a-python-file-compressor-with-ghostscript-and-pillow","status":"publish","type":"post","link":"https:\/\/2thenew.xyz\/blog\/building-a-python-file-compressor-with-ghostscript-and-pillow\/","title":{"rendered":"Building a Python File Compressor with Ghostscript and Pillow"},"content":{"rendered":"

Building a Production\u2011Grade File Compression Utility in Python<\/span><\/h2>\n
We built an open\u2011source Python tool that shrinks 30\u201350 MB PDFs and images down to under 2 MB \u2014 with readable output, a desktop GUI, Docker support, CLI, and a one\u2011click .exe<\/code> build. This post explains how it works under the hood.<\/p>\n

The Problem Nobody Talks About<\/h2>\n

Every organization has the same quiet pain point: oversized files<\/strong>.<\/p>\n

    \n
  • A 42 MB scanned contract that won\u2019t attach to an email<\/li>\n
  • A folder of 35 MB site photos that choke a web upload form<\/li>\n
  • A SharePoint library ballooning because nobody compresses before uploading<\/li>\n<\/ul>\n

    Online compressors cap uploads, destroy quality, or charge monthly fees.
    \nManual Ghostscript commands? Great \u2014 if you memorize flags for a living.<\/p>\n

    We needed something that just works:
    \ntarget a size, hit compress, done.<\/strong><\/p>\n

    What This Utility Does<\/h2>\n\n\n\n\n\n\n\n\n\n\n\n\n\n
    Capability<\/th>\nDetails<\/th>\n<\/tr>\n<\/thead>\n
    Image compression<\/td>\nJPEG, PNG, TIFF, BMP, WebP \u2192 optimized JPEG<\/td>\n<\/tr>\n
    PDF compression<\/td>\nGhostscript (primary), raster pipeline, pikepdf fallback<\/td>\n<\/tr>\n
    Precision targeting<\/td>\nSet an exact target size (e.g. 1.8 MB)<\/td>\n<\/tr>\n
    Scanned PDF intelligence<\/td>\nAuto\u2011detects text vs scanned pages<\/td>\n<\/tr>\n
    Batch processing<\/td>\nParallel directory compression<\/td>\n<\/tr>\n
    Desktop GUI<\/td>\nTkinter-based Upload \u2192 Compress \u2192 Download workflow<\/td>\n<\/tr>\n
    CLI<\/td>\nClick-powered command-line interface<\/td>\n<\/tr>\n
    Docker<\/td>\nPreconfigured image with system dependencies<\/td>\n<\/tr>\n
    Standalone .exe<\/td>\nPyInstaller build for zero-install distribution<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

    Real\u2011World Compression Results<\/h2>\n\n\n\n\n\n\n
    File Type<\/th>\nInput Size<\/th>\nOutput Size<\/th>\nReduction<\/th>\n<\/tr>\n<\/thead>\n
    Images<\/td>\n30\u201350 MB<\/td>\n~500 KB<\/td>\n~98%<\/td>\n<\/tr>\n
    PDFs<\/td>\n30\u201350 MB<\/td>\n~1.8 MB<\/td>\n~96%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

    A 42 MB scanned PDF becomes a crisp 1.7 MB<\/strong> file that remains perfectly readable.<\/p>\n

    Architecture Overview<\/h2>\n
    \r\nsrc\/\r\n\u251c\u2500\u2500 compressor.py\r\n\u251c\u2500\u2500 image_compressor.py\r\n\u251c\u2500\u2500 pdf_compressor.py\r\n\u251c\u2500\u2500 config.py\r\n\u2514\u2500\u2500 utils.py\r\n\r\nmain.py\r\ncompression_service.py\r\nui_app.py\r\nui_actions.py\r\nui_styles.py\r\nstate.py\r\n  <\/code><\/pre>\n
    The key design decision:
    \none compression core<\/strong>.
    \nBoth the CLI and GUI call compress_core()<\/code>.
    \nNo duplicated logic. No divergence.<\/p>\n
    How the Image Pipeline Works<\/h2>\n
      \n
    1. Load & Analyze<\/strong> \u2014 dimensions, mode, alpha channel<\/li>\n
    2. Pre\u2011process<\/strong> \u2014 RGB conversion, alpha compositing<\/li>\n
    3. Resize<\/strong> \u2014 proportional downscale using LANCZOS<\/li>\n
    4. Optimize<\/strong> \u2014 dual\u2011axis quality + dimension tuning<\/li>\n
    5. Validate<\/strong> \u2014 verify size and compression ratio<\/li>\n<\/ol>\n
      This two\u2011axis approach (quality + dimensions) enables aggressive targets
      \nlike 500 KB from a 40 MB image<\/strong> without turning it into mush.<\/p>\n
      How the PDF Pipeline Works<\/h2>\n
      Strategy 1: Ghostscript<\/h3>\n
      \r\ngs -sDEVICE=pdfwrite\r\n   -dCompatibilityLevel=1.4\r\n   -dPDFSETTINGS=\/ebook\r\n   -dColorImageResolution=120\r\n   -dDownsampleColorImages=true\r\n   -sOutputFile=output.pdf\r\n   input.pdf\r\n  <\/code><\/pre>\n
      Presets and DPI values are tried in descending quality order until the target is met.<\/p>\n
      Strategy 2: Raster Pipeline<\/h3>\n
        \n
      • Render pages via pdf2image<\/li>\n
      • Enhance contrast and sharpness (OpenCV)<\/li>\n
      • Binary\u2011search JPEG quality<\/li>\n
      • Rebuild PDF via img2pdf<\/li>\n<\/ul>\n
        Strategy 3: Auto\u2011Detection<\/h3>\n
        Text pages are preserved. Scanned pages are raster\u2011compressed.
        \nOptional region\u2011aware compression preserves text while shrinking photos.<\/p>\n
        Strategy 4: pikepdf Fallback<\/h3>\n
        When all else fails, basic stream compression ensures some reduction.<\/p>\n
        \n
        The Desktop GUI<\/h2>\n
        A Tkinter-based interface with a simple workflow:
        \nUpload \u2192 Compress \u2192 Download<\/strong><\/p>\n
          \n
        • Threaded background compression<\/li>\n
        • KB \/ MB target selector<\/li>\n
        • Temp file safety (no overwrites)<\/li>\n
        • Minimal state management<\/li>\n<\/ul>\n
          \n
          The CLI<\/h2>\n
          \r\npython main.py compress document.pdf --target-size 1800\r\npython main.py compress-dir input\/ --target-size 500 --workers 8\r\npython main.py analyze document.pdf\r\npython main.py check\r\n  <\/code><\/pre>\n
          \n
          Using It as a Python Library<\/h2>\n
          \r\nfrom src.compressor import FileCompressor\r\n\r\ncompressor = FileCompressor(target_size_kb=1800)\r\nresult = compressor.compress(\"input.pdf\", \"output.pdf\")\r\n  <\/code><\/pre>\n
          Lessons Learned<\/h2>\n
            \n
          • Binary search beats linear quality stepping<\/li>\n
          • Two\u2011axis optimization is essential<\/li>\n
          • Scanned PDFs need raster treatment<\/li>\n
          • Ghostscript discovery on Windows is non\u2011trivial<\/li>\n
          • Temp outputs prevent accidental data loss<\/li>\n<\/ul>\n
            \n
            Getting Started<\/h2>\n
            \r\npip install -r requirements.txt\r\npython main.py compress large.pdf --target-size 1800\r\n  <\/code><\/pre>\n
            Built with Python, Ghostscript, Pillow, and a healthy frustration with file size limits.<\/p>\n<\/article>\n","protected":false},"excerpt":{"rendered":"
            Building a Production\u2011Grade File Compression Utility in Python We built an open\u2011source Python tool that shrinks 30\u201350 MB PDFs and images down to under 2 MB \u2014 with readable output, a desktop GUI, Docker support, CLI, and a one\u2011click .exe build. This post explains how it works under the hood. The Problem Nobody Talks About […]<\/p>\n","protected":false},"author":2246,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"iawp_total_views":61,"footnotes":""},"categories":[5879],"tags":[8512,8513,1358],"class_list":["post-78958","post","type-post","status-publish","format-standard","hentry","category-python","tag-ghostscript","tag-pillow","tag-python"],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t