[WIP] Thumbnail extraction

scripts/extract_paper_thumbnails.py

@@ -0,0 +1,31 @@
+import os
+import PIL
+import pdf2image
+
+
+HEIGHT = 150  # thumbnail height in pixel
+
+# output path
+path_thumbs = '../database/thumbnails/2017'
+os.makedirs(path_thumbs, exist_ok=True)
+
+path_pdf = '../database/pdfs/2017/ArigaFG17.pdf'
+cur_fn = os.path.splitext(os.path.basename(path_pdf))[0]
+
+# open pdf
+try:
+    images = pdf2image.convert_from_path(path_pdf)
+except pdf2image.exceptions.PDFPageCountError:
+    print('Could not open {}'.format(path_pdf))
+
+for cur_idx, cur_img in enumerate(images):
+    cur_ratio = cur_img.size[0] / cur_img.size[1]
+    new_width = int(HEIGHT * cur_ratio)
+
+    # resize and save
+    out = cur_img.resize((new_width, HEIGHT), resample=PIL.Image.LANCZOS)
+    path_out = os.path.join(path_thumbs, '{}_{}.png'.format(cur_fn, cur_idx))
+    try:
+        out.save(path_out)
+    except IOError:
+        print('Cannot save image: {}, {}'.format(cur_fn, cur_idx))