# HG changeset patch
# User Paul Boddie <paul@boddie.org.uk>
# Date 1254340971 -7200
# Node ID ff3800a700d52bde0b8e3f07d15b4521f2f352d9
# Parent  de111fdce60f8bfba452256fc4e10493c6493c16
Simplified the IndexWriter document cache, adopting a list of items instead of a
dictionary.
Added a get_document method to the IndexReader class in order to support more
convenient access to indexed documents, also adding dictionary-style methods to
the Document class for field access.
Removed inaccurate comments.

diff -r de111fdce60f -r ff3800a700d5 iixr/index.py
--- a/iixr/index.py	Sun Sep 27 23:03:19 2009 +0200
+++ b/iixr/index.py	Wed Sep 30 22:02:51 2009 +0200
@@ -41,10 +41,11 @@
 
     "A container of document information."
 
-    def __init__(self, docnum):
+    def __init__(self, docnum, fields=None):
         self.docnum = docnum
-        self.fields = []
+        self.fields = fields or []
         self.terms = {}
+        self.field_dict = None
 
     def add_position(self, term, position):
 
@@ -70,6 +71,18 @@
 
         self.fields = fields
 
+    def _ensure_dict(self):
+        if self.field_dict is None:
+            self.field_dict = dict(self.fields)
+
+    def keys(self):
+        self._ensure_dict()
+        return self.field_dict.keys()
+
+    def __getitem__(self, key):
+        self._ensure_dict()
+        return self.field_dict[key]
+
 class IndexWriter:
 
     """
@@ -86,7 +99,7 @@
         self.field_dict_partition = 0
 
         self.terms = {}
-        self.docs = {}
+        self.docs = []
 
         self.doc_counter = 0
 
@@ -100,7 +113,7 @@
         for term, positions in doc.terms.items():
             self.terms.setdefault(term, {})[doc.docnum] = positions
 
-        self.docs[doc.docnum] = doc.fields
+        self.docs.append((doc.docnum, doc.fields))
 
         self.doc_counter += 1
         if self.flush_interval and self.doc_counter >= self.flush_interval:
@@ -147,17 +160,16 @@
 
         # Get the documents in order.
 
-        docs = self.docs.items()
-        docs.sort()
+        self.docs.sort()
 
         field_dict_writer = self.get_field_writer()
 
-        for docnum, fields in docs:
+        for docnum, fields in self.docs:
             field_dict_writer.write_fields(docnum, fields)
 
         field_dict_writer.close()
 
-        self.docs = {}
+        self.docs = []
         self.field_dict_partition += 1
 
     def close(self):
@@ -192,6 +204,9 @@
     def get_fields(self, docnum):
         return self.field_dict_reader.get_fields(docnum)
 
+    def get_document(self, docnum):
+        return Document(docnum, self.get_fields(docnum))
+
     def close(self):
         self.dict_reader.close()
         self.field_dict_reader.close()
diff -r de111fdce60f -r ff3800a700d5 iixr/positions.py
--- a/iixr/positions.py	Sun Sep 27 23:03:19 2009 +0200
+++ b/iixr/positions.py	Wed Sep 30 22:02:51 2009 +0200
@@ -73,8 +73,6 @@
         to 'count'.
         """
 
-        # Duplicate the file handle.
-
         f = self.open("rb")
         return PositionIterator(f, offset, count)
 
@@ -121,8 +119,6 @@
         to 'doc_frequency'.
         """
 
-        # Duplicate the file handle.
-
         f = self.open("rb")
         return PositionIndexIterator(f, offset, doc_frequency)