# HG changeset patch
# User Paul Boddie <paul@boddie.org.uk>
# Date 1317480029 -7200
# Node ID 6e1d369de4b38dc9934e2a3548d4fdc6b4fdf6cf
# Parent  48a194ecc68c42b32cb80f7b43a598c67582cfc0
Incorporated usage of accessors into readers so that they can provide records
and keys directly.

diff -r 48a194ecc68c -r 6e1d369de4b3 simplex/__init__.py
--- a/simplex/__init__.py	Sat Oct 01 16:01:22 2011 +0200
+++ b/simplex/__init__.py	Sat Oct 01 16:40:29 2011 +0200
@@ -31,13 +31,12 @@
 from simplex.readers import *
 import bisect
 
-def make_index(reader, accessor, interval):
+def make_index(reader, interval):
 
     """
-    Index a resource whose 'reader' provides records and whose 'accessor' can
-    yield the key for such records, creating an index entry for a record after a
-    given number of records, defined by 'interval', have been read since the
-    last entry was produced.
+    Index a resource whose 'reader' provides records and keys, creating an index
+    entry for a record after a given number of records, defined by 'interval',
+    have been read since the last entry was produced.
     """
 
     l = []
@@ -46,8 +45,7 @@
     current_key = None
     start_pos = 0
 
-    for i, record in enumerate(reader.get_records()):
-        key = accessor.get_key(record)
+    for i, (key, record) in enumerate(reader.get_records()):
 
         # Where duplicate keys are permitted, the first record employing the key
         # must be available as an index entry. Otherwise, records preceding the
@@ -65,13 +63,12 @@
 
     return l
 
-def find_with_index(reader, accessor, l, term):
+def find_with_index(reader, l, term):
 
     """
-    Find in the resource whose 'reader' provides records and whose 'accessor'
-    can yield the key for such records, using the given index list 'l', the
-    given 'term', returning a record employing the term or None if no such
-    record was found.
+    In the resource whose 'reader' provides records and keys, using the given
+    index list 'l', find the given 'term', returning a record employing the term
+    or None if no such record was found.
     """
 
     i = bisect.bisect_left(l, (term, None))
@@ -90,18 +87,17 @@
         found, pos = l[i]
 
     reader.seek(pos)
-    return find_in_file(reader, accessor, term)
+    return find_in_file(reader, term)
 
-def find_in_file(reader, accessor, term):
+def find_in_file(reader, term):
 
     """
-    Find in the resource whose 'reader' provides records and whose 'accessor'
-    can yield the key for such records, the given 'term', returning a record
-    employing the term or None if no such record was found.
+    In the resource whose 'reader' provides records and keys, find the given
+    'term', returning a record employing the term or None if no such record was
+    found.
     """
 
-    for record in reader.get_records():
-        key = accessor.get_key(record)
+    for key, record in reader.get_records():
         if term == key:
             return record
 
diff -r 48a194ecc68c -r 6e1d369de4b3 simplex/readers.py
--- a/simplex/readers.py	Sat Oct 01 16:01:22 2011 +0200
+++ b/simplex/readers.py	Sat Oct 01 16:40:29 2011 +0200
@@ -22,14 +22,34 @@
 
     "A wrapper around text files."
 
-    def __init__(self, f):
+    def __init__(self, f, accessor):
         self.f = f
+        self.accessor = accessor
 
     def seek(self, pos):
         self.f.seek(pos)
 
     def get_records(self):
-        return self.f.xreadlines()
+        return FileIterator(self.f, self.accessor)
+
+class FileIterator:
+
+    "An iterator over records employing record accessors."
+
+    def __init__(self, resource, accessor):
+        self.resource = resource
+        self.accessor = accessor
+        self.iterator = None
+
+    def __iter__(self):
+        self.iterator = iter(self.resource.xreadlines())
+        return self
+
+    def next(self):
+        if self.iterator is None:
+            iter(self)
+        record = self.iterator.next()
+        return self.accessor.get_key(record), record
 
 class DelimitedRecord:
 
@@ -47,6 +67,9 @@
 
         self.keys = keys or [0]
         self.delimiter = delimiter
+
+        # Define a conversion method.
+
         self.convert = numeric and self.convert_numeric or (lambda x: x)
 
     def convert_numeric(self, term):
diff -r 48a194ecc68c -r 6e1d369de4b3 test_indexed.py
--- a/test_indexed.py	Sat Oct 01 16:01:22 2011 +0200
+++ b/test_indexed.py	Sat Oct 01 16:40:29 2011 +0200
@@ -13,18 +13,18 @@
     sys.exit(1)
 
 f = open(filename)
-reader = TextFile(f)
 accessor = DelimitedRecord(keys, numeric=(numeric == "true"))
+reader = TextFile(f, accessor)
 try:
     t = time.time()
-    l = make_index(reader, accessor, int(interval))
+    l = make_index(reader, int(interval))
     print "Indexed in %s seconds." % (time.time() - t)
 
     # Now use the index.
 
     for term in terms:
         t = time.time()
-        line = find_with_index(reader, accessor, l, accessor.convert(term))
+        line = find_with_index(reader, l, accessor.convert(term))
         if line:
             print "Found (at %s seconds)...\n%s" % (time.time() - t, line)
 
diff -r 48a194ecc68c -r 6e1d369de4b3 test_scan.py
--- a/test_scan.py	Sat Oct 01 16:01:22 2011 +0200
+++ b/test_scan.py	Sat Oct 01 16:40:29 2011 +0200
@@ -13,14 +13,14 @@
     sys.exit(1)
 
 f = open(filename)
-reader = TextFile(f)
 accessor = DelimitedRecord(keys, numeric=(numeric == "true"))
+reader = TextFile(f, accessor)
 try:
     for term in terms:
         reader.seek(0)
 
         t = time.time()
-        line = find_in_file(reader, accessor, accessor.convert(term))
+        line = find_in_file(reader, accessor.convert(term))
         if line:
             print "Found (at %s seconds)...\n%s" % (time.time() - t, line)
 finally: