tweak comments

Author: JoshData

README.md

@@ -20,21 +20,20 @@ Installation
 	node make singlefile
 	cd ..
 
-	## for the renderer
+	# for the renderer
 
 	sudo pip3 install pillow
 
 Running
 -------
 
-	# Compute the changes.
-
-	# Unfortunately the pdf.js prints warnings on STDOUT, so we have
-	# to filter those out.
+Compute the changes (writes a JSON file):
 
 	node index.js before.pdf after.pdf | grep -v "^Warning:" > changes.json
 
-	# Render the changes.
+(Unfortunately the pdf.js library prints warnings on STDOUT, so we have to filter those out.)
+
+Render the changes (turns the PDFs + JSON file into a big PNG image):
 
 	python3 render.py < changes.json > test.png
 

render.py

@@ -7,7 +7,7 @@
 def pdftopng(pdffile, pagenumber, width=900):
     pngbytes = subprocess.check_output(["/usr/bin/pdftoppm", "-f", str(pagenumber), "-l", str(pagenumber), "-scale-to", str(width), "-png", pdffile])
     im = Image.open(io.BytesIO(pngbytes))
-    return im.convert("RGBA") # greyscale is required by .invert()
+    return im.convert("RGBA")
 
 # Load all of the pages named in changes.
 
@@ -20,6 +20,7 @@ def pdftopng(pdffile, pagenumber, width=900):
         pages[change["pdf"]["index"]][change["page"]["number"]] = pdftopng(change["pdf"]["file"], change["page"]["number"])
 
 # Draw red boxes around changes.
+
 for change in changes:
     if change == "*": continue # not handled yet
 
@@ -53,7 +54,7 @@ def pdftopng(pdffile, pagenumber, width=900):
     # do crop
     for pg in pages[idx]:
         im = pages[idx][pg]
-        bbox = ImageOps.invert(im.convert("L")).getbbox()
+        bbox = ImageOps.invert(im.convert("L")).getbbox() # .invert() requires a grayscale image
         vpad = int(.02*im.size[1])
         pages[idx][pg] = im.crop( (minx, max(0, bbox[1]-vpad), maxx, min(im.size[1], bbox[3]+vpad) ) )