Changeset 332 for pyndexter/trunk/pyndexter/hyperestraier.py

Timestamp:

30/04/06 14:09:58 (4 years ago)

Author:

athomas

Message:

pyndexter:

• Added a horribly inefficient built-in indexer, default.DefaultIndexer. There seems to be a memory leak somewhere, so on large datasets the indexer will consume large amounts of memory.
• Added a util module with CacheDict, currently used by the default indexer.
• Added some more CAP_ bits.
• Source state data is now accumulated by the Source classes __iter__() method, no longer requiring a full walk of the source to collect state. This means an Indexer.update() will automagically do the right thing.
• Factored out some common environment initialisation code into Indexer._init_env().
• Factored FileSource include/exclude/predicate code into base Source class so it can be reused.

Files:

• pyndexter/trunk/pyndexter/hyperestraier.py (modified) (2 diffs)

pyndexter/trunk/pyndexter/hyperestraier.py

@@ -5 +5 @@
 class HyperestraierIndexer(Indexer):
     capabilities = CAP_READONLY | CAP_CONTENT | CAP_ATTRIBUTES | CAP_ORDERING |\
-                   CAP_HITCOUNT | CAP_LIST | CAP_RELEVANCE
+                   CAP_HITCOUNT | CAP_LIST | CAP_RELEVANCE | CAP_WHOLEWORD | \
+                   CAP_ASTERISK | CAP_INTERSECTION
 
     def __init__(self, path, source=None, mode=READWRITE, hype_mode=None):
         Indexer.__init__(self, source, mode, os.path.join(path, 'state.db'))
         self.path = path
-        if not os.path.exists(self.path):
-            if mode != READWRITE:
-                raise IndexerError("Index directory has not been initialised")
-            os.makedirs(self.path)
+        self._init_env(self.path)
         self.hype_path = os.path.join(self.path, 'hyperestraier.db')
         if hype_mode is None:
@@ -60 +58 @@
             else:
                 order_type = 'STR'
-            print order_ascending
             order = u'@%s %s%s' % (order_by, order_type,
                                    order_ascending and 'A' or 'D')