Commitment Extractor

--A rule based approach by Hui Guo (a PhD student at NCSU) converted this to Python.

A Python class that extracts commitments from text using a list of rules.

A commitment is normative relation between a subject and an object in which the subject is committed to the object to bring about a consequent if the anticident holds. The following is an example commitment:

If you agree to the list, I'll go ahead and submit them.

• subject: I
• object: you
• anticident: you agree to the list
• consequent: go ahead and submit them

However, in everyday conversations, the object and the anticident are frequently ommitted. This class identifies the commitment in a sentence, and its subject and consequent only. The subject has to include I or we.

Note that sometimes people include compositions in their commitments. The example above actually includes two consequents: go head and submit them (well, whether or not they are actually two separate things is a differnt matter). This class also identifies the position of a composition: in the subject, in the consequent, or other places.

Requirements

This Python code requires spacy and its models to parse texts. We recommend the en_web_core_lg model.

Other

This class checks whether or not a verb phrase is a consequent by checking whether the verb is commisive. We have included a list of commisive verbs under the words folder. Please pay attention to the file path.

The verb be can be either commisive or not commisive. For example:

I'll be eight tomorrow.

I'll be there tomorrow.

Whether or not it is commisive depends on the word after it. We have included a list of be phrases under the words folder. Please check our example code to see how to use it.

Example

The following is a sample output.

[{       
        'para_id': 0,
        'words': "If you agree to the list , I 'll go ahead and submit them .",
        'rule': 'Rule 1: MD+V',
        'subject': 'I',
        'subject_id': [7],
        'commisive': True,
        'commisive_word': 'go',
        'commisive_word_id': 9,
        'compositions': [{'composition_type': 'Consequent',
                          'composition_word': 'and',
                          'composition_word_id': 11,
                          'compositioned_commisive_word': 'submit',
                          'compositioned_commisive_word_id': 12}]
},
{
        'para_id': 1, 
        'words': "I 'll be eight tomorrow .", 
        'commisive': False
},
{       
        'para_id': 2,
        'words': "I 'll be there tomorrow .",
        'rule': 'Rule 1a: MD+be+other',
        'subject': 'I',
        'subject_id': [0],
        'commisive': True,
        'commisive_word': 'be there',
        'commisive_word_id': 3,
        'compositions': [],
}]

Please cite

@inproceedings{Kalia+SCC+2013,
  author    = {Anup K. Kalia and Hamid R. Motahari Nezhad and Claudio Bartolini and Munindar P. Singh},
  title     = {Monitoring Commitments in People-Driven Service Engagements},
  booktitle = {Proceedings of IEEE International Conference on Services Computing},
  pages     = {160--167},
  publisher = {{IEEE} Computer Society},
  year      = {2013},
  address = {Santa Clara, US}
  }