replace_strategies

`replace_strategies` ¶

Classes:

Name	Description
`ReplaceMethodBase`	Shared configuration for all replacement methods.
`Substitute`	Replace entities with LLM-generated synthetic values.
`Redact`	Replace each entity with a configurable redaction template.
`Annotate`	Tag each entity with a readable label token.
`Hash`	Replace each entity with a deterministic hash token.

`ReplaceMethodBase` `pydantic-model` ¶

Bases: BaseModel

Shared configuration for all replacement methods.

`Substitute` `pydantic-model` ¶

Bases: ReplaceMethodBase

Replace entities with LLM-generated synthetic values.

Fields:

instructions (str | None)

`instructions = None` `pydantic-field` ¶

Additional instructions for the LLM replacement generator.

`Redact` `pydantic-model` ¶

Bases: ReplaceMethodBase

Replace each entity with a configurable redaction template.

Fields:

format_template (str)
normalize_label (bool)

Validators:

validate_format_template → format_template

`format_template = '[REDACTED_{label}]'` `pydantic-field` ¶

Template with optional {label} placeholder.

`normalize_label = True` `pydantic-field` ¶

Uppercase and clean label before substitution.

`replace(text, label)` ¶

Apply the redaction template to a single entity occurrence.

Parameters:

Name	Type	Description	Default
`text`	`str`	Original entity text (e.g. `"Alice"`).	required
`label`	`str`	Entity label (e.g. `"first_name"`).	required

Source code in src/anonymizer/config/replace_strategies.py

def replace(self, text: str, label: str) -> str:
    """Apply the redaction template to a single entity occurrence.

    Args:
        text: Original entity text (e.g. ``"Alice"``).
        label: Entity label (e.g. ``"first_name"``).
    """
    normalized_label = _format_label_for_redaction(label) if self.normalize_label else label
    return self._render_template(
        template=self.format_template,
        text=text,
        label=normalized_label,
    )

`Annotate` `pydantic-model` ¶

Bases: ReplaceMethodBase

Tag each entity with a readable label token.

Fields:

format_template (str)

Validators:

validate_format_template → format_template

`format_template = '<{text}, {label}>'` `pydantic-field` ¶

Template with {text} and {label} placeholders.

`replace(text, label)` ¶

Apply the annotation template to a single entity occurrence.

Parameters:

Name	Type	Description	Default
`text`	`str`	Original entity text (e.g. `"Alice"`).	required
`label`	`str`	Entity label (e.g. `"first_name"`).	required

Source code in src/anonymizer/config/replace_strategies.py

def replace(self, text: str, label: str) -> str:
    """Apply the annotation template to a single entity occurrence.

    Args:
        text: Original entity text (e.g. ``"Alice"``).
        label: Entity label (e.g. ``"first_name"``).
    """
    return self._render_template(
        template=self.format_template,
        text=text,
        label=label,
    )

`Hash` `pydantic-model` ¶

Bases: ReplaceMethodBase

Replace each entity with a deterministic hash token.

Fields:

algorithm (Literal['sha256', 'sha1', 'md5'])
digest_length (int)
format_template (str)

Validators:

validate_format_template → format_template

`algorithm = 'sha256'` `pydantic-field` ¶

Hash algorithm.

`digest_length = 12` `pydantic-field` ¶

Number of hex characters to keep from the hash digest.

`format_template = '<HASH_{label}_{digest}>'` `pydantic-field` ¶

Template with {digest} required and optional {label}.

`replace(text, label)` ¶

Apply the hash template to a single entity occurrence.

Parameters:

Name	Type	Description	Default
`text`	`str`	Original entity text (e.g. `"Alice"`).	required
`label`	`str`	Entity label (e.g. `"first_name"`).	required

Source code in src/anonymizer/config/replace_strategies.py

def replace(self, text: str, label: str) -> str:
    """Apply the hash template to a single entity occurrence.

    Args:
        text: Original entity text (e.g. ``"Alice"``).
        label: Entity label (e.g. ``"first_name"``).
    """
    digest = hashlib.new(self.algorithm, text.encode("utf-8")).hexdigest()[: self.digest_length]
    return self._render_template(
        template=self.format_template,
        text=text,
        label=label.upper(),
        digest=digest,
    )

replace_strategies