Choose which features you would like to run. Features include "entities", "pii", "keyword", "sentiment" and "summary".

At least one feature must be included and set to True otherwise the request will be rejected.

NOTE: sentiment is currently only supported on english.

body = {
    "features": {
        "entities": True,
        "pii": True,
        "sentiment": True,
        "keyword": False,
        "summary": False
    },
    "text": "the text"
}
response = requests.post(url, headers=headers, json=body)

Choose which entities you would like to return. All entity types are returned by default.

If you supply an entity type that does not exist, it will be ignored.

The list of entities that can be extracted differs by language. For a full list of entity types by languages, see Entities.

NOTE: If you configured custom entity, then those entity type names can also be included in this setting.

body = {
    "features": {
        "entities": True,
    },
    "settings": {
        "entity.types": ["PERSON", "ORG"]
    },
    "text": "the text"
}
response = requests.post(url, headers=headers, json=body)

Choose which PII types you would like to return.

If you supply a PII type that does not exist, it will be ignored.

For a full list of PII type names, see PII

NOTE: If you configured custom PII analysis, then those PII type names can also be included in this setting.

body = {
    "features": {
        "pii": True,
    },
    "settings": {
        "pii.proximity": ["GENERAL_EMAIL", "GENERAL_PERSON"],
    },
    "text": "the text"
}
response = requests.post(url, headers=headers, json=body)

If performing PII analysis, PII patterns that contain proximity rules (i.e. pattern near a keyword) default to a proximity of 100 characters.

You can change this default using the pii.proximity setting.

body = {
    "features": {
        "pii": True
    },
    "settings": {
        "pii.proximity": 200
    }
    "text": "the text"
}
response = requests.post(url, headers=headers, json=body)

If performing PII analysis, PII patterns that contain proximity rules (i.e. pattern near a keyword) default to a proximity of 100 characters.

You can change this default for a specific PII type using the pii.TYPE_NAME.proximity setting.

For a full list of PII type names, see PII

NOTE: If you configured custom PII analysis, then those PII type names can also be included in this setting.

body = {
        "features": {
            "pii": True
        },
        "settings": {
            "pii.GENERAL_PAYMENT_CARD.proximity": 200,
            "pii.GENERAL_SWIFT_CODE.proximity": 200
        },
        "text": "the text"
    }
    response = requests.post(url, headers=headers, json=body)

If using they keyword analyzer, the default number of keywords returned is 5.

It's possible to change this value using the keyword.max.results setting.

body = {
    "features": {
        "keyword": True
    },
    "settings": {
        "keyword.max.results": 10
    },
    "text": "the text"
}
response = requests.post(url, headers=headers, json=body)

If using they keyword analyzer, keywords are only returned if the number of words in the keyword is 4 or less.

It's possible to change this value using the keyword.max.words setting.

body = {
    "features": {
        "keyword": True
    },
    "settings": {
        "keyword.max.words": 2
    },
    "text": "the text"
}
response = requests.post(url, headers=headers, json=body)

If using they summary analyzer, the default number of sentences returned is 3.

It's possible to change this value using the summary.max.sentences setting.

body = {
    "features": {
        "summary": True
    },
    "settings": {
        "summary.max.sentences": 5
    },
    "text": "the text"
}
response = requests.post(url, headers=headers, json=body)

Diagnostics information can be returned with each request. Diagnostics information includes timings for various analyzers and may also include some warnings.

body = {
    "features": {
        "entities": True,
        "pii": True,
        "sentiment": True,
        "keyword": False,
        "summary": False
    },
    "text": "the text",
    "settings": {
        "diagnostics": True
    }
}
response = requests.post(url, headers=headers, json=body)