Quick Start¶

Install¶

django-dandelion is available on https://pypi.python.org/pypi/django-dandelion/ install it simply with:

$ pip install django-dandelion

Configure¶

Settings¶

Add django_dandelion to your INSTALLED_APPS

INSTALLED_APPS = [
    # ...
    'django_dandelion',
]

Add the entry DANDELION_TOKEN. The recommended method is to setup your production keys using environment variables. This helps to keep them more secure. Your test keys can be displayed in your code directly.

The following entry look for your DANDELION_TOKEN in your environment and, if it can’t find them,uses your test keys values instead:

DANDELION_TOKEN = os.environ.get("DANDELION_TOKEN", "<your dandelion token>")

Register on dandelion to obtain your authentication keys and enrich your application with our semantic intelligence.

You can also change the url of the host and decide whether to use the cache:

DANDELION_HOST = 'api.dandelion.eu'  # Default 'api.dandelion.eu'
DANDELION_USE_CACHE = True  # Default True

Requests¶

DataTXT¶

EntityExtraction (Documentation)

>>> from django_dandelion.datatxt import EntityExtraction
>>> entityExtraction = EntityExtraction()
>>> entityExtraction.params = 'text', u'They say Apple is better than Windows'
>>> entityExtraction.params = 'lang', u'en'
>>> entityExtraction.analyze()
{u'annotations': [{u'confidence': 0.7353,
   u'end': 14,
   u'id': 856,
   u'label': u'Apple',
   u'spot': u'Apple',
   u'start': 9,
   u'title': u'Apple Inc.',
   u'uri': u'http://en.wikipedia.org/wiki/Apple_Inc.'},
  {u'confidence': 0.6904,
   u'end': 37,
   u'id': 18890,
   u'label': u'Microsoft Windows',
   u'spot': u'Windows',
   u'start': 30,
   u'title': u'Microsoft Windows',
   u'uri': u'http://en.wikipedia.org/wiki/Microsoft_Windows'}],
 u'lang': u'en',
 u'time': 0,
 u'timestamp': u'2017-03-09T16:10:46.703'}

>>> entityExtraction = EntityExtraction()
>>> spots = entityExtraction.UserDefinedSpots()  # spots = EntityExtraction.UserDefinedSpots()
>>> obj = spots.create(data=u'{"description": "My botanical custom spots","lang": "en","list": [{"spot": "làres","topic": "Larix decidua"},{"spot": "stropèr","topic": "Salix viminalis"},{"spot": "noselèr","topic": "Corylus avellana"},{"spot": "pomèr","topic": "Malus domestica"},{"spot": "brugnèra","topic": "Prunus domestica"}]}')
>>> obj
{u'created': u'2017-03-10T11:34:45',
 u'data': {u'defaults': {u'exactMatch': False,
   u'greedy': True,
   u'namedEntity': False},
  u'description': u'My botanical custom spots',
  u'lang': u'en',
  u'list': [{u'exactMatch': False,
    u'greedy': True,
    u'namedEntity': False,
    u'spot': u'l\xe0res',
    u'topic': u'Larix decidua'},
   {u'exactMatch': False,
    u'greedy': True,
    u'namedEntity': False,
    u'spot': u'strop\xe8r',
    u'topic': u'Salix viminalis'},
   {u'exactMatch': False,
    u'greedy': True,
    u'namedEntity': False,
    u'spot': u'nosel\xe8r',
    u'topic': u'Corylus avellana'},
   {u'exactMatch': False,
    u'greedy': True,
    u'namedEntity': False,
    u'spot': u'pom\xe8r',
    u'topic': u'Malus domestica'},
   {u'exactMatch': False,
    u'greedy': True,
    u'namedEntity': False,
    u'spot': u'brugn\xe8ra',
    u'topic': u'Prunus domestica'}]},
 u'dataType': u'custom-spots',
 u'id': u'b1290d8a-4af3-4c70-85cc-1eb6a97f7725',
 u'modified': u'2017-03-10T11:34:45',
 u'timestamp': u'2017-03-09T16:10:46.703'}
>>> entityExtraction.params = 'text', u'Larix decidua, commonly called European or common larch, is a deciduous conifer although it looks like a needled evergreen in summer. It is a large tree that will grow to 60-100’ tall with a pyramidal shape, horizontal branching and drooping branchlets.'
>>> entityExtraction.params = 'custom_spots', obj.id
>>> entityExtraction.params = 'min_confidence', 0.8
>>> entityExtraction.analyze()
{u'annotations': [{u'confidence': 0.8759,
   u'end': 13,
   u'id': 1584580,
   u'label': u'European larch',
   u'spot': u'Larix decidua',
   u'start': 0,
   u'title': u'Larix decidua',
   u'uri': u'http://en.wikipedia.org/wiki/Larix_decidua'},
  {u'confidence': 0.8184,
   u'end': 55,
   u'id': 99384,
   u'label': u'Larch',
   u'spot': u'larch',
   u'start': 50,
   u'title': u'Larch',
   u'uri': u'http://en.wikipedia.org/wiki/Larch'},
  {u'confidence': 0.829,
   u'end': 71,
   u'id': 66722,
   u'label': u'Deciduous',
   u'spot': u'deciduous',
   u'start': 62,
   u'title': u'Deciduous',
   u'uri': u'http://en.wikipedia.org/wiki/Deciduous'},
  {u'confidence': 0.9061,
   u'end': 79,
   u'id': 68085,
   u'label': u'Conifer',
   u'spot': u'conifer',
   u'start': 72,
   u'title': u'Pinophyta',
   u'uri': u'http://en.wikipedia.org/wiki/Pinophyta'}],
 u'lang': u'en',
 u'langConfidence': 1.0,
 u'time': 12,
 u'timestamp': u'2017-03-09T16:10:46.703'}

TextSimilarity (Documentation)

>>> from django_dandelion.datatxt import TextSimilarity
>>> textSimilarity = TextSimilarity()
>>> textSimilarity.params = 'text1', u'Reports that the NSA eavesdropped on world leaders have "severely shaken" relations between Europe and the U.S., German Chancellor Angela Merkel said.'
>>> textSimilarity.params = 'text2', u'Germany and France are to seek talks with the US to settle a row over spying, as espionage claims continue to overshadow an EU summit in Brussels.'
>>> textSimilarity.analyze()
{u'annotations': [{u'confidence': 0.7353,
   u'end': 14,
   u'id': 856,
   u'label': u'Apple',
   u'spot': u'Apple',
   u'start': 9,
   u'title': u'Apple Inc.',
   u'uri': u'http://en.wikipedia.org/wiki/Apple_Inc.'},
  {u'confidence': 0.6904,
   u'end': 37,
   u'id': 18890,
   u'label': u'Microsoft Windows',
   u'spot': u'Windows',
   u'start': 30,
   u'title': u'Microsoft Windows',
   u'uri': u'http://en.wikipedia.org/wiki/Microsoft_Windows'}],
 u'lang': u'en',
 u'time': 0,
 u'timestamp': u'2017-03-09T16:10:46.703'}

TextClassification (Documentation)

>>> from django_dandelion.datatxt import TextClassification
>>> textClassification = TextClassification()
>>> classifier = textClassification.UserDefinedClassifiers()  # classifier = TextClassification.UserDefinedClassifiers()
>>> obj = classifier.create(data=u'{"description": "My first model for classifying news","lang": "en","categories": [{"name": "Sport","topics": {"http://en.wikipedia.org/wiki/Sport": 2.0,"http://en.wikipedia.org/wiki/Baseball": 1.0,"http://en.wikipedia.org/wiki/Basketball": 1.0,"http://en.wikipedia.org/wiki/Football": 1.0}},{"name": "Politics","topics": {"Politics": 2.0,"Politician": 1.5,"David Cameron": 1.0,"Angela Merkel": 1.0}}]}')
>>> obj
{u'created': u'2017-03-09T16:10:46.703',
 u'data': {u'categories': [{u'name': u'Sport',
    u'topics': {u'http://en.wikipedia.org/wiki/Baseball': 1.0,
     u'http://en.wikipedia.org/wiki/Basketball': 1.0,
     u'http://en.wikipedia.org/wiki/Football': 1.0,
     u'http://en.wikipedia.org/wiki/Sport': 2.0}},
   {u'name': u'Politics',
    u'topics': {u'Angela Merkel': 1.0,
     u'David Cameron': 1.0,
     u'Politician': 1.5,
     u'Politics': 2.0}}],
  u'description': u'My first model for classifying news',
  u'lang': u'en'},
 u'dataType': u'cl-model',
 u'id': u'7a5a4c4f-8e4a-484a-9f65-b3240819bcfa',
 u'modified': u'2017-03-09T16:36:02',
 u'timestamp': u'2017-03-09T16:10:46.703'}
>>> textClassification.params = 'text', u'See how the main parties are doing in the latest opinion polls on voting intention'
>>> textClassification.params = 'model', obj.id
>>> textClassification.analyze()
{u'categories': [{u'name': u'Politics', u'score': 0.34198442},
  {u'name': u'Sport', u'score': 0.10255624}],
 u'lang': u'en',
 u'time': 1,
 u'timestamp': u'2017-03-09T16:10:46.703'}

LanguageDetection (Documentation)

>>> from django_dandelion.datatxt import LanguageDetection
>>> languageDetection = LanguageDetection()
>>> languageDetection.params = 'text', u'Le nostre tre M sono: mafia, mamma, mandolino'
>>> languageDetection.analyze()
{u'detectedLangs': [{u'confidence': 1.0, u'lang': u'it'}],
 u'time': 1,
 u'timestamp': u'2017-03-09T16:10:46.703'}

SentimentAnalysis (Documentation)

>>> from django_dandelion.datatxt import SentimentAnalysis
>>> sentimentAnalysis = SentimentAnalysis()
>>> sentimentAnalysis.params = 'text', u'Grande applicazione, grande social! Viva Twitter'
>>> sentimentAnalysis.analyze()
{u'lang': u'it',
 u'langConfidence': 1.0,
 u'sentiment': {u'score': 0.85, u'type': u'positive'},
 u'time': 0,
 u'timestamp': u'2017-03-09T16:10:46.703'}

Datagraph¶

Wikisearch (Documentation)

>>> from django_dandelion.datagraph import Wikisearch
>>> wikisearch = Wikisearch()
>>> wikisearch.params = 'text', u'Duomo di Trento'
>>> wikisearch.params = 'lang', u'it'
>>> wikisearch.params = 'limit', 3
>>> wikisearch.analyze()
{u'count': 1117,
 u'entities': [{u'id': 925191,
   u'label': u'Cattedrale di San Vigilio',
   u'title': u'Cattedrale di San Vigilio',
   u'uri': u'http://it.wikipedia.org/wiki/Cattedrale_di_San_Vigilio',
   u'weight': 35.093304},
  {u'id': 730261,
   u'label': u'Duomo',
   u'title': u'Duomo',
   u'uri': u'http://it.wikipedia.org/wiki/Duomo',
   u'weight': 30.126003},
  {u'id': 100137,
   u'label': u'Trento',
   u'title': u'Trento',
   u'uri': u'http://it.wikipedia.org/wiki/Trento',
   u'weight': 23.563704}],
 u'lang': u'it',
 u'query': u'full',
 u'time': 3,
 u'timestamp': u'2017-03-09T16:10:46.703'}