Sentiment analysis

The sentiments app extracts expressed sentiments from your documents, positive or negative, as well as a percentage of positives and negatives per document.

Names as well as pronouns like he or she are resolved, i.e. He is a nice person. will appear as John Smith is a nice person. in the results.

Results

SentimentsResults

interface SentimentsResults {
    locations: Location[];
    positive: number;
    negative: number;
}

with:

Property	Description
locations	Sentiments found in the document
positive	Percentage of positive sentiments expressed in the document
negative	Percentage of negative sentiments expressed in the document

Location

interface Location {
  polarity: SentimentPolarity;
  text: string;
  begin_offset: number;
  end_offset: number;
  object?: string;
  adjective?: string;
  noun?: string;
  verb?: string;
  expression?: string;
}

with:

Property	Description
polarity	Polarity of the sentiment
text	Literal text of the sentiment
begin_offset	Begin offset within the document
end_offset	End offset within the document
object	Object to which the sentiment refers
adjective	When the polarity comes from an adjective, like beautiful
noun	When the polarity comes from a noun, like crisis
verb	When the polarity comes from a verb, like hate
expression	When the polarity comes from an expression, like making waves

SentimentPolarity

type SentimentPolarity = 'positive' | 'negative';

Examples

Sentiment analysis

The sentiments app extracts expressed sentiments from your documents, positive or negative, as well as a percentage of positives and negatives per document.

Names as well as pronouns like he or she are resolved, i.e. He is a nice person. will appear as John Smith is a nice person. in the results.

Results

SentimentsResults

interface SentimentsResults {
    locations: Location[];
    positive: number;
    negative: number;
}

with:

Property	Description
locations	Sentiments found in the document
positive	Percentage of positive sentiments expressed in the document
negative	Percentage of negative sentiments expressed in the document

Location

interface Location {
  polarity: SentimentPolarity;
  text: string;
  begin_offset: number;
  end_offset: number;
  object?: string;
  adjective?: string;
  noun?: string;
  verb?: string;
  expression?: string;
}

with:

Property	Description
polarity	Polarity of the sentiment
text	Literal text of the sentiment
begin_offset	Begin offset within the document
end_offset	End offset within the document
object	Object to which the sentiment refers
adjective	When the polarity comes from an adjective, like beautiful
noun	When the polarity comes from a noun, like crisis
verb	When the polarity comes from a verb, like hate
expression	When the polarity comes from an expression, like making waves

SentimentPolarity

type SentimentPolarity = 'positive' | 'negative';

API

Examples

You will need to install the english language module to run the sample. pip install wowool-english

Extracting sentiments

Note that in the results the sentiments have been resolved. For this to happen we need to add the entities in the pipeline.

from wowool.sdk import Pipeline
from wowool.sentiments.app_id import APP_ID

text = "John Dow is a very nice person, but he sucks at football."
pipeline = Pipeline(
    [
        "english",
        "entity",
        "sentiment",
        "sentiments.app",
    ]
)
document = pipeline(text)
print(document.results(APP_ID))

results:

{
    "positive": 50.0,
    "negative": 50.0,
    "locations": [
        {
            "polarity": "positive",
            "text": "John Dow is a very nice person",
            "begin_offset": 0,
            "end_offset": 30,
            "object": "John Dow",
            "adjective": "nice"
        },
        {
            "polarity": "negative",
            "text": "John Dow sucks",
            "begin_offset": 36,
            "end_offset": 44,
            "object": "John Dow",
            "expression": "suck",
            "comp": "intrans",
            "verb": "suck"
        }
    ]
}

License

In both cases you will need to acquirer a license file at https://www.wowool.com

Non-Commercial

This library is licensed under the GNU AGPLv3 for non-commercial use.  
For commercial use, a separate license must be purchased.  

Commercial license Terms

1. Grants the right to use this library in proprietary software.  
2. Requires a valid license key  
3. Redistribution in SaaS requires a commercial license.