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 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 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
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
