def test(docs, allpairs, kernelpairs, classifiers=[SLK_PRED, SST_PRED],
dditype=type, tag="", backup=False, printstd=False):
#expects: ddi_sentences.md
#data = ddi_train_slk.model, ddi_train_sst.model
tempfiles = []
excludesentences = []
if SLK_PRED in classifiers:
logging.info("**Testing SLK classifier %s ..." % (tag,))
#testpairdic = ddi_kernels.fromddiDic(testdocs)
kernelmodels.generatejSREdata(kernelpairs, docs, tag + "ddi_test_jsre.txt", dditype=dditype,
excludesentences=excludesentences)
kernelmodels.testjSRE(tag + "ddi_test_jsre.txt", tag + "ddi_test_result.txt",
model=tag + "ddi_train_slk.model")
allpairs = kernelmodels.getjSREPredicitons(tag + "ddi_test_jsre.txt", tag + "ddi_test_result.txt",
allpairs, kernelpairs, dditype=dditype)
tempfiles.append(kernelmodels.basedir + tag + "ddi_test_jsre.txt")
tempfiles.append(kernelmodels.basedir + tag + "ddi_test_result.txt")
if SST_PRED in classifiers:
logging.info("****Testing SST classifier %s ..." % (tag,))
allpairs = kernelmodels.testSVMTK(docs, kernelpairs, allpairs, dditype=dditype,
model=tag + "ddi_train_sst.model", tag=tag,
excludesentences=excludesentences)
return tempfiles, allpairs
def train(docs, kernelpairs, classifiers, dditype="all", tag="", backup=False):
tempfiles = []
excludesentences = []
#data = ddi_sentences.build_data_frame(docs)
#logging.info("excluding %s/%s sentences from the train data",
# len(excludesentences), len(docs))
if SLK_PRED in classifiers:
logging.info("**Training SLK classifier %s ..." % (tag,))
#trainpairdic = ddi_kernels.fromddiDic(traindocs)
kernelmodels.generatejSREdata(kernelpairs, docs, tag + "ddi_train_jsre.txt",
dditype=dditype, excludesentences=excludesentences,
train=True)
kernelmodels.trainjSRE(tag + "ddi_train_jsre.txt", tag + "ddi_train_slk.model")
logging.info("done.")
#logging.info("pred: %s \ntest_y: %s", labels, test_y)
#tempfiles.append(ddi_kernels.basedir + tag + "ddi_train_jsre.txt")
#tempfiles.append(ddi_kernels.basedir + tag + "ddi_train_slk.model")
if SST_PRED in classifiers:
logging.info("****Training SST classifier %s ..." % (tag,))
kernelmodels.trainSVMTK(docs, kernelpairs, model=tag + "ddi_train_sst.model",
dditype=dditype, excludesentences=excludesentences)
tempfiles.append("ddi_models/" + tag + "ddi_train_sst.model")
logging.info("done.")
print tag + " training complete"
if backup:
print "backing up these files:", tempfiles
backupFiles("train_results_", tempfiles)
return tempfiles
def test(docs,
allpairs,
kernelpairs,
classifiers=[SLK_PRED, SST_PRED],
dditype=type,
tag="",
backup=False,
printstd=False):
#expects: ddi_sentences.md
#data = ddi_train_slk.model, ddi_train_sst.model
tempfiles = []
excludesentences = []
if SLK_PRED in classifiers:
logging.info("**Testing SLK classifier %s ..." % (tag, ))
#testpairdic = ddi_kernels.fromddiDic(testdocs)
kernelmodels.generatejSREdata(kernelpairs,
docs,
tag + "ddi_test_jsre.txt",
dditype=dditype,
excludesentences=excludesentences)
kernelmodels.testjSRE(tag + "ddi_test_jsre.txt",
tag + "ddi_test_result.txt",
model=tag + "ddi_train_slk.model")
allpairs = kernelmodels.getjSREPredicitons(tag + "ddi_test_jsre.txt",
tag + "ddi_test_result.txt",
allpairs,
kernelpairs,
dditype=dditype)
tempfiles.append(kernelmodels.basedir + tag + "ddi_test_jsre.txt")
tempfiles.append(kernelmodels.basedir + tag + "ddi_test_result.txt")
if SST_PRED in classifiers:
logging.info("****Testing SST classifier %s ..." % (tag, ))
allpairs = kernelmodels.testSVMTK(docs,
kernelpairs,
allpairs,
dditype=dditype,
model=tag + "ddi_train_sst.model",
tag=tag,
excludesentences=excludesentences)
return tempfiles, allpairs
def train(docs, kernelpairs, classifiers, dditype="all", tag="", backup=False):
tempfiles = []
excludesentences = []
#data = ddi_sentences.build_data_frame(docs)
#logging.info("excluding %s/%s sentences from the train data",
# len(excludesentences), len(docs))
if SLK_PRED in classifiers:
logging.info("**Training SLK classifier %s ..." % (tag, ))
#trainpairdic = ddi_kernels.fromddiDic(traindocs)
kernelmodels.generatejSREdata(kernelpairs,
docs,
tag + "ddi_train_jsre.txt",
dditype=dditype,
excludesentences=excludesentences,
train=True)
kernelmodels.trainjSRE(tag + "ddi_train_jsre.txt",
tag + "ddi_train_slk.model")
logging.info("done.")
#logging.info("pred: %s \ntest_y: %s", labels, test_y)
#tempfiles.append(ddi_kernels.basedir + tag + "ddi_train_jsre.txt")
#tempfiles.append(ddi_kernels.basedir + tag + "ddi_train_slk.model")
if SST_PRED in classifiers:
logging.info("****Training SST classifier %s ..." % (tag, ))
kernelmodels.trainSVMTK(docs,
kernelpairs,
model=tag + "ddi_train_sst.model",
dditype=dditype,
excludesentences=excludesentences)
tempfiles.append("ddi_models/" + tag + "ddi_train_sst.model")
logging.info("done.")
print tag + " training complete"
if backup:
print "backing up these files:", tempfiles
backupFiles("train_results_", tempfiles)
return tempfiles