# HG changeset patch
# User Dan Stowell <danstowell@users.sourceforge.net>
# Date 1354187216 0
# Node ID 8a27868916feb5e65201f5d0f1d9b5717370519b
# Parent  5cca21430889dc2b7d4c2589b05997c55d8abf3d
allow command-line usage to have different train and test folders.
Adds a dependency on argparse (>=python 2.7)

diff -r 5cca21430889 -r 8a27868916fe README.md
--- a/README.md	Wed Nov 14 14:16:53 2012 +0000
+++ b/README.md	Thu Nov 29 11:06:56 2012 +0000
@@ -16,12 +16,14 @@
 
 Requirements
 ------------
-* Python 2.x (has been tested in python 2.7)
+* Python 2.7 or later (it uses the 'argparse' module, not available earlier)
 * Python modules:
     * numpy
     * scikits.audiolab
     * sklearn.mixture
 
+It has been tested on python 2.7 (on ubuntu 11.10 and 12.04).
+
 
 Usage example 1: commandline
 -------------
@@ -32,6 +34,11 @@
 This can be a useful test that the script is working;
 however, remember that for proper evaluation you must test using files independent from the training files.
 
+To train and test on different folders, you can run it like this:
+
+	python smacpy.py -t trainwavs -T testwavs
+
+
 Usage example 2: from your own code
 -------------
 In this hypothetical example we train on four audio files, labelled as either 'usa' or 'uk', and then test on a separate audio file of someone called hubert:
diff -r 5cca21430889 -r 8a27868916fe smacpy.py
--- a/smacpy.py	Wed Nov 14 14:16:53 2012 +0000
+++ b/smacpy.py	Thu Nov 29 11:06:56 2012 +0000
@@ -11,8 +11,8 @@
 # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 import os.path
-import sys
 import numpy as np
+import argparse
 from glob import glob
 from scikits.audiolab import Sndfile
 from scikits.audiolab import Format
@@ -142,34 +142,44 @@
 # If this file is invoked as a script, it carries out a simple runthrough
 # of training on some wavs, then testing (on the same ones, just for confirmation, not for eval)
 if __name__ == '__main__':
-	foldername = 'wavs'
-	if len(sys.argv) > 1:
-		foldername = sys.argv[1]
 
-	trainingdata = {}
-	pattern = os.path.join(foldername, '*.wav')
-	for wavpath in glob(pattern):
-		label = os.path.basename(wavpath).split('_')[0]
-		shortwavpath = os.path.relpath(wavpath, foldername)
-		trainingdata[shortwavpath] = label
-	if len(trainingdata)==0:
-		raise RuntimeError("Found no files using this pattern: %s" % pattern)
-	if verbose:
-		print "Class-labels and filenames to be used in training:"
-		for wavpath,label in sorted(trainingdata.iteritems()):
-			print " %s: \t %s" % (label, wavpath)
+	# Handle the command-line arguments for where the train/test data comes from:
+	parser = argparse.ArgumentParser()
+	parser.add_argument('-t', '--trainpath',  default='wavs', help="Path to the WAV files used for training")
+	parser.add_argument('-T', '--testpath',   default='wavs', help="Path to the WAV files used for testing")
+	parser.add_argument('-c', '--charsplit',  default='_',   help="Character used to split filenames: anything BEFORE this character is the class")
+	parser.add_argument('-V', dest='lessverbose', action='store_true', help="Be less verbose")
+	args = vars(parser.parse_args())
+	verbose = not args['lessverbose']
+
+	# Build up lists of the training and testing WAV files:
+	wavsfound = {'trainpath':{}, 'testpath':{}}
+	for onepath in ['trainpath', 'testpath']:
+		pattern = os.path.join(args[onepath], '*.wav')
+		for wavpath in glob(pattern):
+			label = os.path.basename(wavpath).split(args['charsplit'])[0]
+			shortwavpath = os.path.relpath(wavpath, args[onepath])
+			wavsfound[onepath][shortwavpath] = label
+		if len(wavsfound[onepath])==0:
+			raise RuntimeError("Found no files using this pattern: %s" % pattern)
+		if verbose:
+			print "Class-labels and filenames to be used from %s:" % onepath
+			for wavpath,label in sorted(wavsfound[onepath].iteritems()):
+				print " %s: \t %s" % (label, wavpath)
 
 	print "##################################################"
 	print "TRAINING"
-	model = Smacpy(foldername, trainingdata)
+	model = Smacpy(args['trainpath'], wavsfound['trainpath'])
 
 	print "##################################################"
-	print "TESTING (nb on the same files as used for training - for true evaluation please train and test on independent data):"
+	print "TESTING"
+	if args['trainpath'] == args['testpath']:
+		print " (nb testing on the same files as used for training - for true evaluation please train and test on independent data):"
 	ncorrect = 0
-	for wavpath,label in trainingdata.iteritems():
-		result = model.classify(os.path.join(foldername, wavpath))
+	for wavpath,label in wavsfound['testpath'].iteritems():
+		result = model.classify(os.path.join(args['testpath'], wavpath))
 		print " inferred: %s" % result
 		if result == label:
 			ncorrect += 1
-	print "Got %i correct out of %i" % (ncorrect, len(trainingdata))
+	print "Got %i correct out of %i (trained on %i classes)" % (ncorrect, len(wavsfound['testpath']), len(model.gmms))