Issue 6685 - Offer incremental filter list downloads

abp/filters/renderer.py

Index: abp/filters/renderer.py
===================================================================
--- a/abp/filters/renderer.py
+++ b/abp/filters/renderer.py
@@ -24,7 +24,7 @@
 from .parser import parse_filterlist, Comment, Metadata
 from .sources import NotFound
 
-__all__ = ['IncludeError', 'MissingHeader', 'render_filterlist']
+__all__ = ['IncludeError', 'MissingHeader', 'render_filterlist', 'render_diff']
 
 _logger = logging.getLogger(__name__)
 
@@ -124,11 +124,12 @@
     """Remove duplicate metadata and headers."""
     # Always remove checksum -- a checksum coming from a fragment
     # will not match for the rendered list.
-    seen = {'Checksum'}
+    seen = {'checksum'}
     for i, line in enumerate(lines):
         if line.type == 'metadata':
-            if line.key not in seen:
-                seen.add(line.key)
+            key = line.key.lower()
+            if key not in seen:
+                seen.add(key)
                 yield line
         elif line.type == 'header':
             if i == 0:
@@ -180,3 +181,51 @@
                  _validate]:
         lines = proc(lines)
     return lines
+
+
+def _split_list_for_diff(list_in):
+    filterlist, metadata, keys = set(), set(), set()
+    for line in parse_filterlist(list_in):
+        if line.type == 'metadata':
+            metadata.add(line.to_string())
+            keys.add(line.key.lower())
+        elif line.type == 'filter':
+            filterlist.add(line.to_string())
+    return filterlist, metadata, keys
+
+
+def render_diff(base, latest):
+    """Return a diff between two filter lists.
+
+    Parameters
+    ----------
+    base : iterator of str
+        The base (old) list that we want to update to latest.
+    lastest : iterator  of str
+        The latest (most recent) list that we want to update to.
+
+    Returns
+    -------
+    iterable of str
+        A diff between two lists (https://issues.adblockplus.org/ticket/6685)
+
+    """
+    latest_fl, latest_md, latest_keys = _split_list_for_diff(latest)
+    base_fl, base_md, base_keys = _split_list_for_diff(base)
+
+    new_md = latest_md - base_md
+    removed_keys = base_keys - latest_keys
+    add_fl = latest_fl - base_fl
+    remove_fl = base_fl - latest_fl
+
+    yield '[Adblock Plus Diff]'
+    for item in new_md:
+        yield item
+    for key in removed_keys:
+        # If a special comment has been removed, enter it with a blank value
+        # so the client will set it back to the default value
+        yield '! {}:'.format(key)
+    for item in remove_fl:
+        yield '- {}'.format(item)
+    for item in add_fl:
+        yield '+ {}'.format(item)