Mercurial > pylearn

annotate test_dataset.py @ 80:40476a7746e8

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
bugfix
author Frederic Bastien <bastienf@iro.umontreal.ca>
date Mon, 05 May 2008 10:56:58 -0400
parents dde1fb1b63ba
children b4159cbdc06b 4b0859606d05
rev   line source
51
59757365a057 the script can be autorun
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 45
diff changeset
1 #!/bin/env python
45
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
2 from dataset import *
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
3 from math import *
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
4 import numpy
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
5
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
6 def test1():
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
7 global a,ds
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
8 a = numpy.random.rand(10,4)
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
9 print a
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
10 ds = ArrayDataSet(a,{'x':slice(3),'y':3,'z':[0,2]})
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
11 print "len(ds)=",len(ds)
54
70147d00615a added assert
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 52
diff changeset
12 assert(len(ds)==10)
45
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
13 print "example 0 = ",ds[0]
54
70147d00615a added assert
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 52
diff changeset
14 # assert
45
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
15 print "x=",ds["x"]
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
16 print "x|y"
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
17 for x,y in ds("x","y"):
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
18 print x,y
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
19 minibatch_iterator = ds.minibatches(fieldnames=['z','y'],n_batches=1,minibatch_size=3,offset=4)
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
20 minibatch = minibatch_iterator.__iter__().next()
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
21 print "minibatch=",minibatch
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
22 for var in minibatch:
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
23 print "var=",var
48
b6730f9a336d Fixing MinibatchDataSet getitem
bengioy@grenat.iro.umontreal.ca
parents: 45
diff changeset
24 print "take a slice and look at field y",ds[1:6:2]["y"]
45
a5c70dc42972 Test functions for dataset.py
bengioy@grenat.iro.umontreal.ca
parents:
diff changeset
25
58
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
26 def test_ArrayDataSet():
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
27 #don't test stream
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
28 #tested only with float value
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
29 a = numpy.random.rand(10,4)
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
30 print a
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
31 ds = ArrayDataSet(a,{'x':slice(3),'y':3,'z':[0,2]})
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
32 assert len(ds)==10
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
33 #assert ds==a? should this work?
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
34 for i in range(len(ds)):
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
35 assert ds[i]['x'].all()==a[i][:2].all()
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
36 assert ds[i]['y']==a[i][3]
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
37 assert ds[i]['z'].all()==a[i][0:3:2].all()
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
38 print "x=",ds["x"]
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
39 print "x|y"
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
40 i=0
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
41 for x in ds('x','y'):
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
42 assert numpy.append(x['x'],x['y']).all()==a[i].all()
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
43 i+=1
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
44 # i=0
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
45 # for x in ds['x','y']: # don't work
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
46 # assert numpy.append(x['x'],x['y']).all()==a[i].all()
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
47 # i+=1
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
48 # for (x,y) in (ds('x','y'),a): #don't work # haven't found a variant that work.
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
49 # assert numpy.append(x,y)==z
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
50 i=0
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
51 for x,y in ds('x','y'):
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
52 assert numpy.append(x,y).all()==a[i].all()
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
53 i+=1
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
54 for minibatch in ds.minibatches(['x','z'], minibatch_size=3):
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
55 assert minibatch[0][:,0:3:2].all()==minibatch[1].all()
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
56 for x,z in ds.minibatches(['x','z'], minibatch_size=3):
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
57 assert x[:,0:3:2].all()==z.all()
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
58
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
59 # for minibatch in ds.minibatches(['z','y'], minibatch_size=3):
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
60 # print minibatch
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
61 # minibatch_iterator = ds.minibatches(fieldnames=['z','y'],n_batches=1,minibatch_size=3,offset=4)
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
62 # minibatch = minibatch_iterator.__iter__().next()
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
63 # print "minibatch=",minibatch
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
64 # for var in minibatch:
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
65 # print "var=",var
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
66 # print "take a slice and look at field y",ds[1:6:2]["y"]
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
67 have_thrown = False
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
68 try:
63
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
69 ds['h'] # h is not defined...
58
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
70 except :
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
71 have_thrown = True
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
72 assert have_thrown == True
63
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
73 assert len(ds.fields())==3
58
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
74 for field in ds.fields():
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
75 for field_value in field: # iterate over the values associated to that field for all the ds examples
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
76 pass
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
77 for field in ds('x','z').fields():
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
78 pass
60
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 58
diff changeset
79 for field in ds.fields('x','y'):
58
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
80 pass
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
81 for field_examples in ds.fields():
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
82 for example_value in field_examples:
17729d7104fa added function test_ArrayDataSet
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 54
diff changeset
83 pass
63
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
84
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
85 assert ds == ds.fields().examples()
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
86
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
87
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
88 #test missing value
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
89
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
90 assert len(ds[:3])==3
66
dde1fb1b63ba fixed test and removed print
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 65
diff changeset
91 i=0
63
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
92 for x,z in ds[:3]('x','z'):
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
93 assert ds[i]['z'].all()==a[i][0:3:2].all()
66
dde1fb1b63ba fixed test and removed print
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 65
diff changeset
94 i+=1
63
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
95 #ds[i1:i2:s]# returns a ds with the examples i1,i1+s,...i2-s.
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
96
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
97 #ds[i]# returns an Example.
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
98
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
99 #ds[[i1,i2,...in]]# returns a ds with examples i1,i2,...in.
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
100
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
101 #ds[fieldname]# an iterable over the values of the field fieldname across
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
102 #the ds (the iterable is obtained by default by calling valuesVStack
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
103 #over the values for individual examples).
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
104
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
105 #ds.<property># returns the value of a property associated with
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
106 #the name <property>. The following properties should be supported:
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
107 # - 'description': a textual description or name for the ds
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
108 # - 'fieldtypes': a list of types (one per field)
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
109 #* ds1 | ds2 | ds3 == ds.hstack([ds1,ds2,ds3])
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
110 #* ds1 & ds2 & ds3 == ds.vstack([ds1,ds2,ds3])
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
111
14589f02a372 more test
Frederic Bastien <bastienf@iro.umontreal.ca>
parents: 60
diff changeset
112
65
d48eba49a2f4 fixed the infinite loop
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 64
diff changeset
113 test_ArrayDataSet()
64
863da25a60f1 trying to fix infinite loop
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents: 63
diff changeset
114