scan2pdf: import

scan2pdf/shell.nix

@@ -0,0 +1,48 @@
+{ pkgs ? import <nixpkgs> {} }:
+let
+  crop_ppm = (pkgs.writers.writePython3Bin "crop_ppm" {libraries = [pkgs.python3Packages.opencv4];} ''
+    import cv2
+    import sys
+
+    for path in sys.argv[1:]:
+        img = cv2.imread(path)[30:-30, 30:]
+        gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
+        _, thresh = cv2.threshold(gray, 1, 255, cv2.THRESH_BINARY)
+        contours, hierarchy = cv2.findContours(
+            thresh,
+            cv2.RETR_EXTERNAL,
+            cv2.CHAIN_APPROX_SIMPLE
+        )
+        cnt = max(contours, key=cv2.contourArea)
+        x, y, w, h = cv2.boundingRect(cnt)
+        print(path, x, y, w, h)
+        crop = img[y:y + h, x:x + w]
+        cv2.imwrite(path, crop)
+  '');
+in
+pkgs.mkShell {
+  packages = [
+    pkgs.poppler_utils
+    pkgs.mupdf
+    crop_ppm
+    (pkgs.writers.writeDashBin "generate" ''
+      set -x
+      pwd=$(pwd)
+      tmpdir=$(mktemp -d)
+      cd "$tmpdir"
+      [ $# -eq 1 ] || {
+        echo Please provide one PDF file as argument. >&2
+        exit 1
+      }
+      [ -f "$1" ] || {
+        echo "$1" does not exist. >&2
+        exit 1
+      }
+      basename=$(basename "$1" .pdf)
+      ${pkgs.poppler_utils}/bin/pdftoppm "$1" exploded
+      ${crop_ppm}/bin/crop_ppm exploded*.ppm
+      ${pkgs.imagemagick}/bin/convert exploded*.ppm "$pwd/cropped-$basename.pdf"
+      ${pkgs.mupdf}/bin/mutool poster -x 2 "$pwd/cropped-$basename.pdf" "$pwd/split-cropped-$basename.pdf"
+    '')
+  ];
+}