DataSanitation Processor

Overview

The DataSanitation processor helps scrub sensitive information from incoming requests. It provides two approaches for configuration: either by specifying entities to inspect (whitelist-style) or entities to exclude (blacklist-style). This processor is ideal for ensuring that sensitive data such as credit card numbers, phone numbers, and email addresses are not passed downstream.

---

Input and Output

This processor operates on the request stream:

• Input Stream: Request – It intercepts and sanitizes the request body before forwarding it to downstream processors.
• Output Stream: Request – After sanitization, the request continues through the processing pipeline.

This ensures sensitive fields are removed or masked before any further processing or logging occurs.

---

Parameters

blocklisted_entities

Type: list_of_strings
Required: False
Default: [CreditCard, Email, Phone]
A list of specific entity types to be scrubbed from the request. When this parameter is used, only the listed entities will be scrubbed (whitelist behavior).

Example:

- key: blocklisted_entities
  value:
    - CreditCard
    - Email
    - Phone

---

ignored_entities

Type: list_of_strings
Required: False
A list of entity types that should be excluded from scrubbing. When this parameter is used, all entities except the ones listed here will be scrubbed (blacklist behavior).

Example:

- key: ignored_entities
  value:
    - Phone

---

Best Practices

• Use blocklisted_entities to explicitly control which fields get sanitized.
• Use ignored_entities when you need to scrub everything except specific fields.
• Always verify the output to ensure no sensitive data is leaked inadvertently.

---

DataSanitation Processor Template

DataSanitation Processor

SanitizeRequest:
  processor: DataSanitation
  parameters:
    - key: blocklisted_entities
      value:
        - CreditCard
        - Email
        - Phone

---

Use Case

You can use the DataSanitation processor to sanitize OpenAI requests. In this configuration, all entities except IPAddress will be sanitized before the request reaches the OpenAI provider.

  OpenAISanitization:
    processor: DataSanitation
    parameters:
      - key: ignored_entities
        value:
          - IPAddress