class TestClassifyApi(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY,
apiServer=conf.BASE_PATH,
retinaName=conf.RETINA_NAME)
def testCreateCategoryFilter(self):
filterName = "filter1"
filter1 = self.client.createCategoryFilter(filterName,
positiveExamples,
negativeExamples)
self.assertGreater(len(filter1.positions), 50)
self.assertEqual(filter1.categoryName, filterName)
def testErrors(self):
expectedException = False
try:
self.client.createCategoryFilter(None, positiveExamples,
negativeExamples)
except CorticalioException:
expectedException = True
self.assertTrue(expectedException)
expectedException = False
try:
self.client.createCategoryFilter("filterName", [],
negativeExamples)
except CorticalioException:
expectedException = True
self.assertTrue(expectedException)
class TestClassifyApi(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY, apiServer=conf.BASE_PATH, retinaName=conf.RETINA_NAME)
def testCreateCategoryFilter(self):
filterName = "filter1"
filter1 = self.client.createCategoryFilter(filterName, positiveExamples, negativeExamples)
self.assertGreater(len(filter1.positions), 50)
self.assertEqual(filter1.categoryName, filterName)
def testErrors(self):
expectedException = False
try:
self.client.createCategoryFilter(None, positiveExamples, negativeExamples)
except CorticalioException:
expectedException = True
self.assertTrue(expectedException)
expectedException = False
try:
self.client.createCategoryFilter("filterName", [], negativeExamples)
except CorticalioException:
expectedException = True
self.assertTrue(expectedException)
class TestClientImageApi(unittest.TestCase):
"""INFO:
'base64/png' image data can be written to disk like this (python2.7):
with open("imageToSave.png", "wb") as f:
f.write(imageData.decode('base64'))
"""
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY, apiServer=conf.BASE_PATH, retinaName=conf.RETINA_NAME)
def testImage(self):
imageData = self.client.getImage(inputJSON)
self.assertNotEqual(imageData, None)
self.assertGreater(len(imageData), 1000)
def testCompare(self):
imageData = self.client.compareImage(inputJSONarray)
self.assertNotEqual(imageData, None)
self.assertGreater(len(imageData), 1000)
def testBulk(self):
images = self.client.getImages(inputJSONarray3, getFingerprint=True)
self.assertEqual(len(images), 3)
for image in images:
self.assertNotEqual(image, None)
self.assertNotEqual(image.image_data, None)
self.assertGreater(len(image.image_data), 1000)
self.assertGreater(len(image.fingerprint.positions), 50)
class TestClientImageApi(unittest.TestCase):
"""INFO:
'base64/png' image data can be written to disk like this (python2.7):
with open("imageToSave.png", "wb") as f:
f.write(imageData.decode('base64'))
"""
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY,
apiServer=conf.BASE_PATH,
retinaName=conf.RETINA_NAME)
def testImage(self):
imageData = self.client.getImage(inputJSON)
self.assertNotEqual(imageData, None)
self.assertGreater(len(imageData), 1000)
def testCompare(self):
imageData = self.client.compareImage(inputJSONarray)
self.assertNotEqual(imageData, None)
self.assertGreater(len(imageData), 1000)
def testBulk(self):
images = self.client.getImages(inputJSONarray3, getFingerprint=True)
self.assertEqual(len(images), 3)
for image in images:
self.assertNotEqual(image, None)
self.assertNotEqual(image.image_data, None)
self.assertGreater(len(image.image_data), 1000)
self.assertGreater(len(image.fingerprint.positions), 50)
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY,
apiServer=conf.BASE_PATH,
retinaName=conf.RETINA_NAME)
self.jsonBulkExpression = None
# path relative to current working dir
with open(bulkInput, "r") as f:
self.jsonBulkExpression = "".join(f.readlines())
class TestClientTextApi(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY,
apiServer=conf.BASE_PATH,
retinaName=conf.RETINA_NAME)
def testText(self):
fp = self.client.getFingerprintForText(inputText)
self.assertNotEqual(fp, None)
self.assertGreater(len(fp.positions), 500)
def testKeywords(self):
termList = self.client.getKeywordsForText(inputText)
self.assertGreater(len(termList), 2)
self.assertTrue(conf.isString(termList[0]))
def testTokenize(self):
sentences = self.client.getTokensForText(inputText)
self.assertGreater(len(sentences), 10)
self.assertTrue(conf.isString(sentences[0]))
firstSentence = sentences[0].split(',')
self.assertEqual(firstSentence[0], "george")
self.assertGreater(len(firstSentence), 10)
verbsSentences = self.client.getTokensForText(inputText, POStags="VB")
for verb in verbsSentences[0].split(","):
self.assertTrue("VERB" in self.client.getTerms(
term=verb)[0].pos_types)
def testSlices(self):
texts = self.client.getSlicesForText(inputText,
getFingerprint=True,
startIndex=0,
maxResults=2)
self.assertEqual(len(texts), 2)
self.assertEqual(texts[0].text.split(' ')[0], "George")
self.assertGreater(len(texts[0].fingerprint.positions), 100)
def testBulk(self):
fingerprints = self.client.getFingerprintsForTexts(bulkTexts,
sparsity=1.0)
self.assertEqual(len(fingerprints), 4)
for fp in fingerprints:
self.assertGreater(len(fp.positions), 100)
def testLanguageDetection(self):
self.assertEqual(
self.client.getLanguageForText("I have a dream!").language,
"English")
self.assertEqual(
self.client.getLanguageForText("Ich bin ein").wiki_url,
"http://en.wikipedia.org/wiki/German_language")
self.assertEqual(
self.client.getLanguageForText(
"Der var så dejligt ude på landet.").iso_tag, "da")
class TestClientTermsApi(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY,
apiServer=conf.BASE_PATH,
retinaName=conf.RETINA_NAME)
def testTerms(self):
terms = self.client.getTerms(term="apple",
getFingerprint=True,
startIndex=0,
maxResults=5)
self.assertTrue(len(terms) == 1)
self.assertTrue(terms[0].term == "apple")
self.assertTrue("NOUN" in terms[0].pos_types)
self.assertTrue(terms[0].df > 0.0001)
self.assertGreater(len(terms[0].fingerprint.positions), 100)
terms100 = self.client.getTerms("app*", startIndex=0, maxResults=100)
self.assertEqual(len(terms100), 100)
def testContexts(self):
contexts = self.client.getContextsForTerm(term="apple",
getFingerprint=True,
startIndex=0,
maxResults=3)
self.assertTrue(contexts != None)
self.assertEqual(3, len(contexts))
c0 = contexts[0]
self.assertGreater(len(c0.fingerprint.positions), 100)
self.assertTrue(conf.isString(c0.context_label))
self.assertTrue(c0.context_id == 0)
def testSimilarTerms(self):
terms = self.client.getSimilarTermsForTerm(term="apple",
contextId=0,
posType="NOUN",
getFingerprint=True,
startIndex=0,
maxResults=8)
self.assertTrue(terms != None)
self.assertEqual(8, len(terms))
t0 = terms[0]
self.assertTrue(len(t0.fingerprint.positions) > 0)
self.assertTrue(t0 != None)
self.assertTrue("NOUN" in t0.pos_types)
def testExceptionTerms(self):
exceptionOccurred = False
try:
terms = self.client.getSimilarTermsForTerm(term="apple",
posType="wrong")
except CorticalioException:
exceptionOccurred = True
self.assertTrue(exceptionOccurred)
class TestClientTextApi(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY, apiServer=conf.BASE_PATH, retinaName=conf.RETINA_NAME)
def testText(self):
fp = self.client.getFingerprintForText(inputText)
self.assertNotEqual(fp, None)
self.assertGreater(len(fp.positions), 500)
def testKeywords(self):
termList = self.client.getKeywordsForText(inputText)
self.assertGreater(len(termList), 2)
self.assertTrue(conf.isString(termList[0]))
def testTokenize(self):
sentences = self.client.getTokensForText(inputText)
self.assertGreater(len(sentences), 10)
self.assertTrue(conf.isString(sentences[0]))
firstSentence = sentences[0].split(',')
self.assertEqual(firstSentence[0], "george")
self.assertGreater(len(firstSentence), 10)
verbsSentences = self.client.getTokensForText(inputText, POStags="VB")
for verb in verbsSentences[0].split(","):
self.assertTrue("VERB" in self.client.getTerms(term=verb)[0].pos_types)
def testSlices(self):
texts = self.client.getSlicesForText(inputText, getFingerprint=True, startIndex=0, maxResults=2)
self.assertEqual(len(texts), 2)
self.assertEqual(texts[0].text.split(' ')[0], "George")
self.assertGreater(len(texts[0].fingerprint.positions), 100)
def testBulk(self):
fingerprints = self.client.getFingerprintsForTexts(bulkTexts, sparsity=1.0)
self.assertEqual(len(fingerprints), 4)
for fp in fingerprints:
self.assertGreater(len(fp.positions), 100)
def testLanguageDetection(self):
self.assertEqual(self.client.getLanguageForText("I have a dream!").language, "English")
self.assertEqual(self.client.getLanguageForText("Ich bin ein").wiki_url, "http://en.wikipedia.org/wiki/German_language")
self.assertEqual(self.client.getLanguageForText("Der var så dejligt ude på landet.").iso_tag, "da")
class TestClientTermsApi(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY, apiServer=conf.BASE_PATH, retinaName=conf.RETINA_NAME)
def testTerms(self):
terms = self.client.getTerms(term="apple", getFingerprint=True, startIndex=0, maxResults=5)
self.assertTrue(len(terms) == 1)
self.assertTrue(terms[0].term == "apple")
self.assertTrue("NOUN" in terms[0].pos_types)
self.assertTrue(terms[0].df > 0.0001)
self.assertGreater(len(terms[0].fingerprint.positions), 100)
terms100 = self.client.getTerms("app*", startIndex=0, maxResults=100)
self.assertEqual(len(terms100), 100)
def testContexts(self):
contexts = self.client.getContextsForTerm(term="apple", getFingerprint=True, startIndex=0, maxResults=3)
self.assertTrue(contexts != None)
self.assertEqual(3, len(contexts))
c0 = contexts[0]
self.assertGreater(len(c0.fingerprint.positions), 100)
self.assertTrue(conf.isString(c0.context_label))
self.assertTrue(c0.context_id == 0)
def testSimilarTerms(self):
terms = self.client.getSimilarTermsForTerm(
term="apple", contextId=0, posType="NOUN", getFingerprint=True, startIndex=0, maxResults=8
)
self.assertTrue(terms != None)
self.assertEqual(8, len(terms))
t0 = terms[0]
self.assertTrue(len(t0.fingerprint.positions) > 0)
self.assertTrue(t0 != None)
self.assertTrue("NOUN" in t0.pos_types)
def testExceptionTerms(self):
exceptionOccurred = False
try:
terms = self.client.getSimilarTermsForTerm(term="apple", posType="wrong")
except CorticalioException:
exceptionOccurred = True
self.assertTrue(exceptionOccurred)
class TestClientRetinasApi(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY, apiServer=conf.BASE_PATH, retinaName=conf.RETINA_NAME)
def testRetinas(self):
retinas = self.client.getRetinas()
self.assertNotEqual(retinas, None)
self.assertNotEqual(retinas[0], None)
self.assertNotEqual(retinas[1], None)
self.assertTrue("en_synonymous" == retinas[0].retinaName or "en_associative" == retinas[0].retinaName)
self.assertTrue("en_synonymous" == retinas[1].retinaName or "en_associative" == retinas[1].retinaName)
self.assertGreater(retinas[0].numberOfTermsInRetina, 50000)
def testException(self):
exceptionOccurred = False
try:
self.client.getRetinas("nonexisting_retina")
except CorticalioException:
exceptionOccurred = True
self.assertTrue(exceptionOccurred)
class TestClientRetinasApi(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY,
apiServer=conf.BASE_PATH,
retinaName=conf.RETINA_NAME)
def testRetinas(self):
retinas = self.client.getRetinas()
self.assertNotEqual(retinas, None)
self.assertNotEqual(retinas[0], None)
self.assertNotEqual(retinas[1], None)
self.assertTrue("en_synonymous" == retinas[0].retinaName
or "en_associative" == retinas[0].retinaName)
self.assertTrue("en_synonymous" == retinas[1].retinaName
or "en_associative" == retinas[1].retinaName)
self.assertGreater(retinas[0].numberOfTermsInRetina, 50000)
def testException(self):
exceptionOccurred = False
try:
self.client.getRetinas("nonexisting_retina")
except CorticalioException:
exceptionOccurred = True
self.assertTrue(exceptionOccurred)
class TestClientExpreissions(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY, apiServer=conf.BASE_PATH, retinaName=conf.RETINA_NAME)
self.jsonBulkExpression = None
# path relative to current working dir
with open(bulkInput, "r") as f:
self.jsonBulkExpression = "".join(f.readlines())
def testExpressions(self):
fp = self.client.getFingerprintForExpression(oneTermInputJSON, sparsity=0.5)
self.assertGreater(len(fp.positions), 100)
def testContexts(self):
contexts = self.client.getContextsForExpression(simpleExpression, getFingerprint=True, startIndex=0, maxResults=3, sparsity=1.0)
self.assertTrue(contexts != None)
self.assertEqual(3, len(contexts))
c0 = contexts[0]
self.assertGreater(len(c0.fingerprint.positions), 100)
self.assertTrue(conf.isString(c0.context_label))
self.assertTrue(c0.context_id == 0)
def testSimilarTerms(self):
terms = self.client.getSimilarTermsForExpression(simpleExpression, contextId=None, posType="NOUN", getFingerprint=True, startIndex=0, maxResults=8, sparsity=1.0)
self.assertTrue(terms != None)
self.assertEqual(8, len(terms))
for term in terms:
self.assertGreater(len(term.fingerprint.positions), 100)
self.assertTrue(term != None)
self.assertTrue("NOUN" in term.pos_types)
def testExpressionBulk(self):
fps = self.client.getFingerprintsForExpressions(self.jsonBulkExpression)
self.assertEqual(6, len(fps))
for fp in fps:
self.assertGreater(len(fp.positions), 50)
def testExpressionContextsBulk(self):
contextsList = self.client.getContextsForExpressions(self.jsonBulkExpression, getFingerprint=True, startIndex=0, maxResults=3)
self.assertEqual(len(contextsList), 6)
for contextList in contextsList:
for i, context in enumerate(contextList):
self.assertGreater(len(context.fingerprint.positions), 50)
self.assertTrue(conf.isString(context.context_label))
self.assertTrue(context.context_id == i)
def testExpressionSimilarTermsBulk(self):
termsLists = self.client.getSimilarTermsForExpressions(self.jsonBulkExpression, getFingerprint=True, maxResults=7)
self.assertTrue(len(termsLists) == 6)
for termList in termsLists:
self.assertTrue(len(termList) == 7)
self.assertGreater(len(termList[0].fingerprint.positions), 100)
class TestClientCompareApi(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY, apiServer=conf.BASE_PATH, retinaName=conf.RETINA_NAME)
def testCompare(self):
resultMetric = self.client.compare(inputJSONarray)
self.assertGreater(resultMetric.cosineSimilarity, 0.1)
self.assertGreater(resultMetric.euclideanDistance, 0.1)
self.assertGreater(resultMetric.jaccardDistance, 0.1)
self.assertGreater(resultMetric.weightedScoring, 0.1)
self.assertGreater(resultMetric.sizeRight, 10)
self.assertGreater(resultMetric.sizeLeft, 10)
self.assertGreater(resultMetric.overlappingLeftRight, 0.1)
self.assertGreater(resultMetric.overlappingAll, 10)
self.assertGreater(resultMetric.overlappingRightLeft, 0.1)
def testCompareBulk(self):
resultMetricList = self.client.compareBulk(bulkJSONarray)
self.assertEqual(len(resultMetricList), 2)
for resultMetric in resultMetricList:
self.assertGreater(resultMetric.cosineSimilarity, 0.1)
self.assertGreater(resultMetric.euclideanDistance, 0.1)
self.assertGreater(resultMetric.jaccardDistance, 0.1)
self.assertGreater(resultMetric.weightedScoring, 0.1)
self.assertGreater(resultMetric.sizeRight, 10)
self.assertGreater(resultMetric.sizeLeft, 10)
self.assertGreater(resultMetric.overlappingLeftRight, 0.1)
self.assertGreater(resultMetric.overlappingAll, 10)
self.assertGreater(resultMetric.overlappingRightLeft, 0.1)
def testException(self):
# testing using only one input element in the array
expectedException = False
try:
self.client.compare(oneTermInputJSONarray)
except CorticalioException:
expectedException = True
self.assertTrue(expectedException)
# testing JSON parse exception in the input
expectedException = False
try:
self.client.compare(syntaxErrorJSONarray)
except CorticalioException:
expectedException = True
self.assertTrue(expectedException)
class TestClientCompareApi(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY,
apiServer=conf.BASE_PATH,
retinaName=conf.RETINA_NAME)
def testCompare(self):
resultMetric = self.client.compare(inputJSONarray)
self.assertGreater(resultMetric.cosineSimilarity, 0.1)
self.assertGreater(resultMetric.euclideanDistance, 0.1)
self.assertGreater(resultMetric.jaccardDistance, 0.1)
self.assertGreater(resultMetric.weightedScoring, 0.1)
self.assertGreater(resultMetric.sizeRight, 10)
self.assertGreater(resultMetric.sizeLeft, 10)
self.assertGreater(resultMetric.overlappingLeftRight, 0.1)
self.assertGreater(resultMetric.overlappingAll, 10)
self.assertGreater(resultMetric.overlappingRightLeft, 0.1)
def testCompareBulk(self):
resultMetricList = self.client.compareBulk(bulkJSONarray)
self.assertEqual(len(resultMetricList), 2)
for resultMetric in resultMetricList:
self.assertGreater(resultMetric.cosineSimilarity, 0.1)
self.assertGreater(resultMetric.euclideanDistance, 0.1)
self.assertGreater(resultMetric.jaccardDistance, 0.1)
self.assertGreater(resultMetric.weightedScoring, 0.1)
self.assertGreater(resultMetric.sizeRight, 10)
self.assertGreater(resultMetric.sizeLeft, 10)
self.assertGreater(resultMetric.overlappingLeftRight, 0.1)
self.assertGreater(resultMetric.overlappingAll, 10)
self.assertGreater(resultMetric.overlappingRightLeft, 0.1)
def testException(self):
# testing using only one input element in the array
expectedException = False
try:
self.client.compare(oneTermInputJSONarray)
except CorticalioException:
expectedException = True
self.assertTrue(expectedException)
# testing JSON parse exception in the input
expectedException = False
try:
self.client.compare(syntaxErrorJSONarray)
except CorticalioException:
expectedException = True
self.assertTrue(expectedException)
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY, apiServer=conf.BASE_PATH, retinaName=conf.RETINA_NAME)
def __init__(self, apiKey):
self._fullClient = FullClient(apiKey, apiServer="http://api.cortical.io/rest", retinaName="en_associative")
class LiteClient(object):
"""Minimalistic client for accessing core features of Cortical.io's Retina API in a simple way."""
def __init__(self, apiKey):
self._fullClient = FullClient(apiKey, apiServer="http://api.cortical.io/rest", retinaName="en_associative")
def _createDictionary(self, textOrFingerprint):
if type(textOrFingerprint) == str:
return {"text": textOrFingerprint}
elif type(textOrFingerprint) == list:
return {"positions": textOrFingerprint}
else:
raise CorticalioException("Invalid argument, cannot create input from: '%s'" % (str(textOrFingerprint)))
def getSimilarTerms(self, textOrFingerprint):
"""Get the similar terms for a given text or fingerprint
Args:
textOrFingerprint, str OR list of integers
Returns:
list of str: the 20 most similar terms
Raises:
CorticalioException: if the request was not successful
"""
expression = self._createDictionary(textOrFingerprint)
terms = self._fullClient.getSimilarTermsForExpression(json.dumps(expression), maxResults=20)
return [t.term for t in terms]
def getKeywords(self, text):
"""Get a list of keywords from the text
Args:
text, str: The input document
Returns:
list of str
Raises:
CorticalioException: if the request was not successful
"""
terms = self._fullClient.getKeywordsForText(text)
return terms
def getFingerprint(self, text):
"""Get the semantic fingerprint of the input text.
Args:
text, str: The text to be evaluated
Returns:
list of str: the positions of the semantic fingerprint
Raises:
CorticalioException: if the request was not successful
"""
fp = self._fullClient.getFingerprintForText(text)
return fp.positions
def compare(self, textOrFingerprint1, textOrFingerprint2):
"""Returns the semantic similarity of texts or fingerprints. Each argument can be eiter a text or a fingerprint.
Args:
textOrFingerprint1, str OR list of integers
textOrFingerprint2, str OR list of integers
Returns:
float: the semantic similarity in the range [0;1]
Raises:
CorticalioException: if the request was not successful
"""
compareList = [self._createDictionary(textOrFingerprint1), self._createDictionary(textOrFingerprint2)]
metric = self._fullClient.compare(json.dumps(compareList))
return metric.cosineSimilarity
def createCategoryFilter(self, positiveExamples):
"""Creates a filter fingerprint.
Args:
positiveExamples, list(str): The list of positive example texts.
Returns:
list of int: the positions representing the filter representing the texts
Raises:
CorticalioException: if the request was not successful
"""
categoryFilter = self._fullClient.createCategoryFilter("CategoryFilter", positiveExamples)
return categoryFilter.positions
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY,
apiServer=conf.BASE_PATH,
retinaName=conf.RETINA_NAME)
class TestClientExpreissions(unittest.TestCase):
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY,
apiServer=conf.BASE_PATH,
retinaName=conf.RETINA_NAME)
self.jsonBulkExpression = None
# path relative to current working dir
with open(bulkInput, "r") as f:
self.jsonBulkExpression = "".join(f.readlines())
def testExpressions(self):
fp = self.client.getFingerprintForExpression(oneTermInputJSON,
sparsity=0.5)
self.assertGreater(len(fp.positions), 100)
def testContexts(self):
contexts = self.client.getContextsForExpression(simpleExpression,
getFingerprint=True,
startIndex=0,
maxResults=3,
sparsity=1.0)
self.assertTrue(contexts != None)
self.assertEqual(3, len(contexts))
c0 = contexts[0]
self.assertGreater(len(c0.fingerprint.positions), 100)
self.assertTrue(conf.isString(c0.context_label))
self.assertTrue(c0.context_id == 0)
def testSimilarTerms(self):
terms = self.client.getSimilarTermsForExpression(simpleExpression,
contextId=None,
posType="NOUN",
getFingerprint=True,
startIndex=0,
maxResults=8,
sparsity=1.0)
self.assertTrue(terms != None)
self.assertEqual(8, len(terms))
for term in terms:
self.assertGreater(len(term.fingerprint.positions), 100)
self.assertTrue(term != None)
self.assertTrue("NOUN" in term.pos_types)
def testExpressionBulk(self):
fps = self.client.getFingerprintsForExpressions(
self.jsonBulkExpression)
self.assertEqual(6, len(fps))
for fp in fps:
self.assertGreater(len(fp.positions), 50)
def testExpressionContextsBulk(self):
contextsList = self.client.getContextsForExpressions(
self.jsonBulkExpression,
getFingerprint=True,
startIndex=0,
maxResults=3)
self.assertEqual(len(contextsList), 6)
for contextList in contextsList:
for i, context in enumerate(contextList):
self.assertGreater(len(context.fingerprint.positions), 50)
self.assertTrue(conf.isString(context.context_label))
self.assertTrue(context.context_id == i)
def testExpressionSimilarTermsBulk(self):
termsLists = self.client.getSimilarTermsForExpressions(
self.jsonBulkExpression, getFingerprint=True, maxResults=7)
self.assertTrue(len(termsLists) == 6)
for termList in termsLists:
self.assertTrue(len(termList) == 7)
self.assertGreater(len(termList[0].fingerprint.positions), 100)
def __init__(self, apiKey):
self._fullClient = FullClient(apiKey,
apiServer="http://api.cortical.io/rest",
retinaName="en_associative")
class LiteClient(object):
"""Minimalistic client for accessing core features of Cortical.io's Retina API in a simple way."""
def __init__(self, apiKey):
self._fullClient = FullClient(apiKey,
apiServer="http://api.cortical.io/rest",
retinaName="en_associative")
def _createDictionary(self, textOrFingerprint):
if type(textOrFingerprint) == str:
return {"text": textOrFingerprint}
elif type(textOrFingerprint) == list:
return {"positions": textOrFingerprint}
else:
raise CorticalioException(
"Invalid argument, cannot create input from: '%s'" %
(str(textOrFingerprint)))
def getSimilarTerms(self, textOrFingerprint):
"""Get the similar terms for a given text or fingerprint
Args:
textOrFingerprint, str OR list of integers
Returns:
list of str: the 20 most similar terms
Raises:
CorticalioException: if the request was not successful
"""
expression = self._createDictionary(textOrFingerprint)
terms = self._fullClient.getSimilarTermsForExpression(
json.dumps(expression), maxResults=20)
return [t.term for t in terms]
def getKeywords(self, text):
"""Get a list of keywords from the text
Args:
text, str: The input document
Returns:
list of str
Raises:
CorticalioException: if the request was not successful
"""
terms = self._fullClient.getKeywordsForText(text)
return terms
def getFingerprint(self, text):
"""Get the semantic fingerprint of the input text.
Args:
text, str: The text to be evaluated
Returns:
list of str: the positions of the semantic fingerprint
Raises:
CorticalioException: if the request was not successful
"""
fp = self._fullClient.getFingerprintForText(text)
return fp.positions
def compare(self, textOrFingerprint1, textOrFingerprint2):
"""Returns the semantic similarity of texts or fingerprints. Each argument can be eiter a text or a fingerprint.
Args:
textOrFingerprint1, str OR list of integers
textOrFingerprint2, str OR list of integers
Returns:
float: the semantic similarity in the range [0;1]
Raises:
CorticalioException: if the request was not successful
"""
compareList = [
self._createDictionary(textOrFingerprint1),
self._createDictionary(textOrFingerprint2)
]
metric = self._fullClient.compare(json.dumps(compareList))
return metric.cosineSimilarity
def createCategoryFilter(self, positiveExamples):
"""Creates a filter fingerprint.
Args:
positiveExamples, list(str): The list of positive example texts.
Returns:
list of int: the positions representing the filter representing the texts
Raises:
CorticalioException: if the request was not successful
"""
categoryFilter = self._fullClient.createCategoryFilter(
"CategoryFilter", positiveExamples)
return categoryFilter.positions
def setUp(self):
self.client = FullClient(apiKey=conf.API_KEY, apiServer=conf.BASE_PATH, retinaName=conf.RETINA_NAME)
self.jsonBulkExpression = None
# path relative to current working dir
with open(bulkInput, "r") as f:
self.jsonBulkExpression = "".join(f.readlines())
