Person's Names (NER)

Filter

This filter identifies person's names in text using a PhEye model. It is configured as an entry in the pheyes array. The earlier single-object person key is deprecated but still accepted, with the same properties.

Required Parameters

This filter has no required parameters.

Optional Parameters

Parameter	Description	Default Value
phEyeConfiguration	Connection settings for the PhEye service, or a modelPath naming a local model to run in-process (see the Models API). A policy may name only an endpoint the deployment allows, and may not set a bearerToken; see Redaction services and the Policy Schema.	None
removePunctuation	When set to true, punctuation will be removed prior to analysis.	false
phEyeFilterStrategies	A list of filter strategies.	None
enabled	When set to false, the filter will be disabled and not applied	true
ignored	A list of terms to be ignored by the filter.	None

Filter Strategies

The filter may have zero or more filter strategies. When no filter strategy is given the default strategy of REDACT is used. When multiple filter strategies are given, they are evaluated in the order listed and only the first one whose condition is satisfied, or that has no condition, is applied. See Filter Strategies for details.

Strategy	Description
REDACT	Replace the sensitive text with a placeholder.
RANDOM_REPLACE	Replace the sensitive text with a similar, random value.
STATIC_REPLACE	Replace the sensitive text with a given value.
CRYPTO_REPLACE	Replace the sensitive text with its encrypted value.
HASH_SHA256_REPLACE	Replace the sensitive text with its SHA256 hash value.
ABBREVIATE	Replace the sensitive text with the initials of the text.

Conditions

Each filter strategy may have one condition. See Conditions for details.

Conditional	Description	Operators
TOKEN	Compares the value of the sensitive text.	==, !=, startswith
CONTEXT	Compares the filtering context.	== , !=
CONFIDENCE	Compares the confidence in the sensitive text against a threshold value.	< , <=, > , >=, ==, !=

Long Text

A PhEye model examines at most a fixed number of words in one call, its max_len (512 for the ph-eye-pii-en-* models), and ignores the rest. The filter divides longer text into overlapping windows that each fit, detects in each, and reports every name once, so a name anywhere in the text is examined. Words are counted as the model counts them, so each punctuation mark is a word. Text that fits is sent in one call.

The window size comes from the model: a model run in-process (modelPath) has its max_len read from its gliner_config.json, and the PhEye service's limit is read from its /model endpoint. A PhEye service without that endpoint, before ph-eye 1.4.0, is sent windows of 256 words. Windows overlap so that a name at a window's edge is found whole. No policy setting is needed, and text splitting is not required for this filter.

Example Policy

{
   "identifiers": {
      "pheyes": [
         {
            "phEyeFilterStrategies": [
               {
                  "strategy": "REDACT",
                  "redactionFormat": "{{{REDACTED-%t}}}"
               }
            ]
         }
      ]
   }
}