Natural Language Generation for Python using AcceleratedText

Installation

$ python -m pip install acctext

Usage

from acctext import AcceleratedText

at = AcceleratedText(host='http://127.0.0.1:3001')

Make sure Accelerated Text application is running. Refer to documentation for launch instructions.

at.health()

{'health': 'Ok'}

Interacting with Dictionary

items = [{'key': 'house',
          'category': 'N',
          'forms': ['house', 'houses']},
         {'key': 'hill',
          'category': 'N',
          'forms': ['hill', 'hills']},
         {'key': 'on',
          'forms': ['on'],
          'category': 'Prep',
          'attributes': {'Operation': 'Syntax.on_Prep/Prep'}},
         {'key': 'the',
          'forms': ['the'],
          'category': 'Det',
          'attributes': {'Operation': 'Syntax.the_Det/Det'}}]

for item in items:
    at.create_dictionary_item(**item)
    
items = at.list_dictionary_items()
items

[{'id': 'the_Det_Eng',
  'key': 'the',
  'forms': ['the'],
  'category': 'Det',
  'language': 'Eng',
  'attributes': {'Operation': 'Syntax.the_Det/Det'}},
 {'id': 'hill_N_Eng',
  'key': 'hill',
  'forms': ['hill', 'hills'],
  'category': 'N',
  'language': 'Eng',
  'attributes': {}},
 {'id': 'house_N_Eng',
  'key': 'house',
  'forms': ['house', 'houses'],
  'category': 'N',
  'language': 'Eng',
  'attributes': {}},
 {'id': 'on_Prep_Eng',
  'key': 'on',
  'forms': ['on'],
  'category': 'Prep',
  'language': 'Eng',
  'attributes': {'Operation': 'Syntax.on_Prep/Prep'}}]

Working with Data

Upload a local file

at.upload_data_file('example_data.csv')

{'message': 'Succesfully uploaded file', 'id': 'example_data.csv'}

Create a data file from scratch

at.create_data_file('example_data_2.csv', ['a', 'b'], [['1', '2'], ['3', '4']])

{'id': 'example_data_2.csv'}

List available data files

[x['id'] for x in at.list_data_files()]

['example_data.csv', 'example_data_2.csv']

Fetch data file

at.get_data_file('example_data_2.csv')

{'id': 'example_data_2.csv',
 'filename': 'example_data_2.csv',
 'header': ['a', 'b'],
 'rows': [['1', '2'], ['3', '4']]}

Delete data file

at.delete_data_file('example_data_2.csv')

{'message': 'Succesfully deleted file', 'id': 'example_data_2.csv'}

Languages and Readers

Fetch existing language properties

at.get_language('Eng')

{'id': 'Eng', 'name': 'English', 'flag': '🇬🇧', 'default': True}

Add new language

at.add_language('Ger', 'German')

{'id': 'Ger', 'name': 'German', 'flag': '🇩🇪', 'default': False}

List available languages

at.list_languages()

[{'id': 'Eng', 'name': 'English', 'flag': '🇬🇧', 'default': True},
 {'id': 'Ger', 'name': 'German', 'flag': '🇩🇪', 'default': False}]

Create new reader type

at.create_reader('Dc', 'Discount Customer', '(DC)')

{'id': 'Dc', 'name': 'Discount Customer', 'flag': '(DC)', 'default': False}

at.create_reader('Lc', 'Loyal Customer', '(LC)')

{'id': 'Lc', 'name': 'Loyal Customer', 'flag': '(LC)', 'default': False}

List available readers

at.list_readers()

[{'id': 'Dc', 'name': 'Discount Customer', 'flag': '(DC)', 'default': False},
 {'id': 'Lc', 'name': 'Loyal Customer', 'flag': '(LC)', 'default': False}]

Document plans

Open Accelerated Text document plan editor (http://127.0.0.1:8080 by default) and create a new document plan named "House description". More detailed instructions can be found in documentation.

Fetch single document plan

dp = at.get_document_plan(name='House description')
dp['documentPlan']

{'type': 'Document-plan',
 'segments': [{'children': [{'modifier': {'name': 'size',
      'type': 'Cell-modifier',
      'srcId': 'B-D0i/`TL4@ja%{U!?2G',
      'child': {'name': 'color',
       'type': 'Cell-modifier',
       'srcId': '!2b?}PBIB?i]%*/(~?XM',
       'child': {'name': 'house',
        'type': 'Dictionary-item',
        'srcId': '+5JLY;_/2/zEOcZ._$,4',
        'kind': 'N',
        'itemId': 'house_N_Eng'}}},
     'type': 'Modifier',
     'srcId': '`62!swypAqp_jK_lr1Ow',
     'child': {'name': 'on',
      'type': 'Dictionary-item-modifier',
      'srcId': ']MNfAFBjxy,c?G55a04@',
      'kind': 'Prep',
      'child': {'name': 'the',
       'type': 'Dictionary-item-modifier',
       'srcId': '62%#$13DP}Gj8=n4NCI.',
       'kind': 'Det',
       'child': {'name': 'hill',
        'type': 'Dictionary-item',
        'srcId': 'Ol68tPXKblg(pUghVhb@',
        'kind': 'N',
        'itemId': 'hill_N_Eng'},
       'itemId': 'the_Det_Eng'},
      'itemId': 'on_Prep_Eng'}}],
   'type': 'Segment',
   'srcId': ']H[rfMhNu,^(wX6[%.+w'}],
 'srcId': 'Li$gv+b_9o-n$z^FnSl~'}

Delete document plan

at.delete_document_plan(dp['id'])

True

Restore document plan

at.create_document_plan(**dp)['name']

'House description'

List document plans

[x['name'] for x in at.list_document_plans(kind='Document')]

['House description']

Text generation

result = at.generate('House description', data={"size": "small", "color": "red"})
result['variants']

['Small red house on the hill.']

Bulk generation

results = at.generate_bulk('House description', data=[{"size": "small", "color": "red"}, 
                                                      {"size": "big", "color": "green"}])
[x['variants'] for x in results]

[['Small red house on the hill.'], ['Big green house on the hill.']]

Fetch specific result

at.get_result(result['resultId'])

{'resultId': 'a364335f-5021-443d-9c77-fe40c296ecef',
 'offset': 0,
 'totalCount': 1,
 'ready': True,
 'updatedAt': 1628173135,
 'variants': ['Small red house on the hill.']}

Working with state

Export

at.export_state('state.zip')

Clear

at.clear_state()

Restore

at.restore_state('state.zip')

accelerated-text / accelerated-text-py