# HG changeset patch
# User Paul Boddie <paul@boddie.org.uk>
# Date 1317422438 -7200
# Node ID b18d0bb03a08ff2a789b87bdfbffdde5a2f8dc58
# Parent  6397ceae2ddaa3aa1ea4c2d28a03338d580ff979
Added support for compound keys.
Fixed out of range index access.
Added a groups convenience function for slicing up sequences into fixed length
groups.

diff -r 6397ceae2dda -r b18d0bb03a08 simplex.py
--- a/simplex.py	Sat Oct 01 00:01:04 2011 +0200
+++ b/simplex.py	Sat Oct 01 00:40:38 2011 +0200
@@ -41,9 +41,9 @@
 
     "A wrapper around text files."
 
-    def __init__(self, f, key=0, delimiter=None):
+    def __init__(self, f, keys=None, delimiter=None):
         self.f = f
-        self.key = key
+        self.keys = keys or [0]
         self.delimiter = delimiter
 
     def seek(self, pos):
@@ -53,7 +53,8 @@
         return self.f.xreadlines()
 
     def get_key(self, record):
-        return record.split(self.delimiter)[self.key]
+        values = record.split(self.delimiter)
+        return [values[key] for key in self.keys]
 
 def index_file(f, interval):
 
@@ -95,7 +96,11 @@
     """
 
     i = bisect.bisect_left(l, (term, None))
-    found, pos = l[i]
+
+    try:
+        found, pos = l[i]
+    except IndexError:
+        return None
 
     # Since the index is more coarse than the underlying file, the bisect left
     # operation will most likely point to an index entry for later records than
@@ -132,4 +137,20 @@
     def find(self, term):
         return find_with_index(self.f, self.entries, term)
 
+def groups(l, length):
+
+    "Split 'l' into groups of the given 'length'."
+
+    if length <= 0:
+        raise ValueError, "Groups must be greater than zero."
+
+    i = 0
+    g = []
+
+    while i < len(l):
+        g.append(l[i:i+length])
+        i += length
+
+    return g
+
 # vim: tabstop=4 expandtab shiftwidth=4
diff -r 6397ceae2dda -r b18d0bb03a08 test_indexed.py
--- a/test_indexed.py	Sat Oct 01 00:01:04 2011 +0200
+++ b/test_indexed.py	Sat Oct 01 00:40:38 2011 +0200
@@ -3,14 +3,20 @@
 from simplex import *
 import sys, time
 
-filename, step = sys.argv[1:3]
-terms = sys.argv[3:]
+try:
+    separator = sys.argv.index("--")
+    filename, interval = sys.argv[1:3]
+    keys = map(int, sys.argv[3:separator])
+    terms = groups(sys.argv[separator+1:], len(keys))
+except (IndexError, ValueError):
+    print >>sys.stderr, "Usage: %s <filename> <interval> <key>... -- <term value>..." % sys.argv[0]
+    sys.exit(1)
 
 f = open(filename)
-tf = TextFile(f)
+tf = TextFile(f, keys)
 try:
     t = time.time()
-    l = index_file(tf, int(step))
+    l = index_file(tf, int(interval))
     print "Indexed in %s seconds." % (time.time() - t)
 
     # Now use the index.
diff -r 6397ceae2dda -r b18d0bb03a08 test_scan.py
--- a/test_scan.py	Sat Oct 01 00:01:04 2011 +0200
+++ b/test_scan.py	Sat Oct 01 00:40:38 2011 +0200
@@ -3,11 +3,17 @@
 from simplex import *
 import sys, time
 
-filename = sys.argv[1]
-terms = sys.argv[2:]
+try:
+    separator = sys.argv.index("--")
+    filename = sys.argv[1]
+    keys = map(int, sys.argv[2:separator])
+    terms = groups(sys.argv[separator+1:], len(keys))
+except (IndexError, ValueError):
+    print >>sys.stderr, "Usage: %s <filename> <key>... -- <term value>..." % sys.argv[0]
+    sys.exit(1)
 
 f = open(filename)
-tf = TextFile(f)
+tf = TextFile(f, keys)
 try:
     for term in terms:
         tf.seek(0)