# HG changeset patch
# User Paul Boddie <paul@boddie.org.uk>
# Date 1285532536 -7200
# Node ID b9d34861708c6557f4a8755e7bce60f9110841fd
# Parent  f036b082e99b59be0e035e32dbaa0613ed31fc21
Added control over which parts of page names will be shown in search results.

diff -r f036b082e99b -r b9d34861708c ImprovedMoinSearch.py
--- a/ImprovedMoinSearch.py	Tue Sep 14 00:21:31 2010 +0200
+++ b/ImprovedMoinSearch.py	Sun Sep 26 22:22:16 2010 +0200
@@ -15,13 +15,30 @@
 paragraph_regexp = re.compile(r"(?P<paragraph>(?:^[^#=\s].*$\n)+)", re.UNICODE | re.MULTILINE)
 
 def range_groups(min_name, max_name):
-    return r"(?P<%s>\d+)?(?:\s*-\s*(?P<%s>\d+))?" % (min_name, max_name)
+    return r"(?P<%s>-?\d+)?(?:\s*-\s*(?P<%s>-?\d+))?" % (min_name, max_name)
 
 format_options_regexp = re.compile(r"("
     "(?P<heading>(heading|title|h)\s*" + range_groups("min_heading", "max_heading") + ")"
     "|(?P<paragraph>(paragraph|para|p)\s*(?P<paragraph_number>\d+)?)"
+    "|(?P<name>(name|page)\s*" + range_groups("first", "last") + ")"
     ")", re.UNICODE)
 
+def convert_index(i, length):
+
+    """
+    Convert from a 1-based indexing scheme to a 0-based scheme for the given
+    index 'i' in a sequence having the given 'length'.
+    """
+
+    if i is None:
+        return i
+    elif i > 0:
+        return i - 1
+    elif i < 0:
+        return length + i
+    else:
+        return i
+
 def getSearchResultPages(request, query, **kw):
 
     """
@@ -62,6 +79,14 @@
 
 def getParagraph(request, page, start=0, number=None):
 
+    """
+    Using the given 'request', return from the given 'page', starting from the
+    optional 'start' offset (or the beginning, if no such offset is specified),
+    the first paragraph or, if the optional 'number' is given, the paragraph
+    whose position corresponds to that number, with a number of 1 being the
+    first paragraph found, 2 being the second, and so on.
+    """
+
     full_page = Page(request, page.page_name)
     body = full_page.get_raw_body()
     if start != 0:
@@ -73,8 +98,37 @@
 
     return None
 
-def getPageName(request, page):
-    return page.page_name
+def getPageName(request, page, start=0, first=None, last=None):
+
+    """
+    Using the given 'request', return the name of the given 'page'. The optional
+    'start' offset refers to the body of the page and is returned as the start
+    and end of the result span if specified.
+
+    If the optional 'first' or 'last' parameters are specified, only the
+    specified span of parts extracted from the page name will be returned, where
+    the parts of the name are obtained by splitting the full name where the
+    slash ("/") character is found. The first part has an index of 1, and the
+    last part can be referred to using an index of -1.
+    """
+
+    parts = page.page_name.split("/")
+
+    first = convert_index(first, len(parts))
+    last = convert_index(last, len(parts))
+
+    if first is None:
+        if last is None:
+            pass
+        else:
+            parts = parts[:last+1]
+    else:
+        if last is None:
+            parts = parts[first:]
+        else:
+            parts = parts[first:last+1]
+
+    return "/".join(parts), (start, start)
 
 def formatResultPages(request, formatter, pages, paging, format, page_from=0):
 
@@ -96,6 +150,8 @@
                 actions.append((getFirstPageHeading, map(int_or_none, (match.group("min_heading"), match.group("max_heading")))))
             elif match.group("paragraph"):
                 actions.append((getParagraph, map(int_or_none, (match.group("paragraph_number"),))))
+            elif match.group("name"):
+                actions.append((getPageName, map(int_or_none, (match.group("first"), match.group("last")))))
     else:
         actions.append((getPageName, ()))