tuplesize<0 works with single points

i.e. I can do this:

  gp.plot( np.array((3,40)),
           tuplesize = -2 )

This patch adds the feature and a test

Author: dkogan

demo.py

@@ -1,6 +1,7 @@
 #!/usr/bin/python
 
 import numpy as np
+import numpysane as nps
 import time
 import sys
 
@@ -52,6 +53,18 @@ def check_expected_error(what, f):
 # first, some very basic stuff. Testing implicit domains, multiple curves in
 # arguments, packed broadcastable data, etc
 gp.plot(x**2, wait=1)
+gp.plot(( np.transpose(nps.cat(x,x**2)),
+          dict(_with='linespoints pt 4 ps 2'),
+         ),
+        ( 5,60,
+          dict(tuplesize=2,
+               _with='linespoints pt 5 ps 2'),
+        ),
+        ( np.array((3,40)),
+          dict(_with='linespoints pt 6 ps 2'),
+        ),
+        tuplesize = -2,
+        wait=1)
 gp.plot(-x, x**3, wait=1)
 gp.plot((x**2), wait=1)
 gp.plot((-x, x**3, {'with': 'lines'}), (x**2,), wait=1)

gnuplotlib.py

@@ -1885,7 +1885,7 @@ def reformat(curve):
                 if len(curve) != 1:
                     raise GnuplotlibError("tuplesize<0 means that only a single numpy array of data should be given: all data is in this array")
                 d['tuplesize'] = -d['tuplesize']
-                d['_data']     = list(nps.mv(curve[0], -1, 0))
+                d['_data']     = list(nps.mv(nps.atleast_dims(curve[0],-2), -1, 0))
             else:
                 d['_data'] = list(curve)
             return d