Add revert flag

zundler/__main__.py

@@ -2,13 +2,19 @@ def main():
     from .args import parse_args
     args = parse_args()
 
-    from .embed import embed_assets
-    embed_assets(
-        args.input_path,
-        output_path=args.output_path,
-        append_pre=args.append_pre,
-        append_post=args.append_post,
-    )
+    from .embed import embed_assets, extract_assets
+    if args.revert:
+        extract_assets(
+            args.input_path,
+            output_path=args.output_path,
+        )
+    else:
+        embed_assets(
+            args.input_path,
+            output_path=args.output_path,
+            append_pre=args.append_pre,
+            append_post=args.append_post,
+        )
 
 
 if __name__ == "__main__":

zundler/args.py

@@ -25,6 +25,14 @@ parser.add_argument(
     version=__version__,
 )
 
+parser.add_argument(
+    '-R', '--revert',
+    default=False,
+    action='store_true',
+    help='set this flag for the reverse operation: deflate a Zundler file'
+         ' into its components (ouput_path must be a directory; "." by default)',
+)
+
 parser.add_argument(
     '-P', '--append-pre',
     default="",

zundler/embed.py

@@ -22,6 +22,7 @@ import json
 import logging
 import os
 from pathlib import Path
+import re
 import zlib
 
 SCRIPT_PATH = os.path.abspath(os.path.dirname(__file__))
@@ -211,7 +212,6 @@ def embed_css_resources(css, filename):
     Also, handle @import."""
     # This uses some heuristics which will fail in general.
     # Eventually a library like tinycss2 might be preferable.
-    import re
 
     # First, make sure all @import's are using url(), because these are both valid:
     # @import url("foo.css");
@@ -307,3 +307,41 @@ def load_filetree(base_dir, before=None, after=None, exclude_pattern=None):
             logger.debug('Packed file %s [%d]' % (key, len(result[key])))
 
     return result
+
+
+def extract_assets(input_path, output_path=None):
+    """Split a file generated by Zundler into its constituents
+
+    Import for debugging"""
+
+    if not output_path:
+        output_path = '.'
+
+    html = open(input_path, 'r').read()
+
+    try:
+        # Find large base64 blob
+        m = re.search('.*<script>.*window.*"(?P<blob>[A-Za-z0-9/+]{128,})".*</script>.*', html)
+        blob = m['blob']
+        blob = base64.b64decode(blob)
+        blob = zlib.decompress(blob).decode()
+        blob = json.loads(blob)
+        file_tree = blob['file_tree']
+    except Exception as e:
+        logger.error(str(e))
+        logger.error("Does not look like a Zundler output file: %s" % input_path)
+        exit(1)
+
+    for filename, file in file_tree.items():
+        filename = os.path.join(output_path, filename)
+        os.makedirs(os.path.dirname(filename) or '.', exist_ok=True)
+        data = file['data']
+        if file['base64encoded']:
+            data = base64.b64decode(data)
+        else:
+            data = data.encode()
+        open(filename, 'wb').write(data)
+        file['data'] = file['data'][:100] + '...'
+
+    with open(os.path.join(output_path, 'file_tree.json'), 'w') as fp:
+        json.dump(file_tree, fp, indent=2)