pydantic-anonymizer 0.3.0

anonymize sensitive data in pydantic models

pydantic-anonymizer

anonymize sensitive data in pydantic models

---

                 ,--.                  ,--.  ,--.       
 ,---.,--. ,--.,-|  | ,--,--.,--,--, ,-'  '-.`--' ,---. 
| .-. |\  '  /' .-. |' ,-.  ||      \'-.  .-',--.| .--' 
| '-' ' \   ' \ `-' |\ '-'  ||  ||  |  |  |  |  |\ `--. 
|  |-'.-'  /   `---'  `--`--'`--''--'  `--'  `--' `---' 
`--'  `---'                                             
                                                                            
                                                  ,--.                      
 ,--,--.,--,--,  ,---. ,--,--, ,--. ,--.,--,--,--.`--',-----. ,---. ,--.--. 
' ,-.  ||      \| .-. ||      \ \  '  / |        |,--.`-.  / | .-. :|  .--' 
\ '-'  ||  ||  |' '-' '|  ||  |  \   '  |  |  |  ||  | /  `-.\   --.|  |    
 `--`--'`--''--' `---' `--''--'.-'  /   `--`--`--'`--'`-----' `----'`--'    
                               `---'                                        

📦 installation

pip install pydantic-anonymizer

---

📑 quick start

from pydantic import BaseModel, Field
from pydantic_anonymizer import Anonymizer

class UserProfile(BaseModel, Anonymizer):
    username: str
    email: str = Field(json_schema_extra={"anonymize": True})
    card_number: str = Field(json_schema_extra={"anonymize": "card"})
    phone_number: str = Field(json_schema_extra={"anonymize": "phone"})

user = UserProfile(
    username="ivan_dev",
    email="[e-mail]",
    card_number="4242111122223333",
    phone_number="[phone]"
)

print(user.model_dump())
# {'username': 'ivan_dev', 'email': '[e-mail]', 'card_number': '4242111122223333', 'phone_number': '[phone]'}

print(user.model_dump_anonymized())
# {'username': 'ivan_dev', 'email': 'i***@***.com', 'card_number': '4242-****-****-3333', 'phone_number': '+380 (***) ***-**-85'}

---

🧩 features

• 🔐 automatic masking - configure via json_schema_extra in model fields
• 🎭 decorator - alternative to mixin via @Anonymize
• 📧 email masking - partial mask for emails (i***@***.com)
• 💳 card masking - format 4242-****-****-3333
• 📱 phone masking - correct country code parsing with phonenumbers
• 🌐 IP masking - 192.168.1.100 → 192.168.***.***
• 🎂 date masking - 15.03.1990 → **.**.****
• 👤 name masking - John Doe → J*** D**
• 🏦 IBAN masking - UA213996...6712 → UA**...****6712
• 🏗️ nested models - recursive processing of nested Pydantic models
• 📋 lists - support for list[Model] with masking of each element
• 🛠️ custom strategies - your own masking functions via MaskRegistry
• 📝 logging integration - AnonymizedFormatter for automatic masking in logs
• ⚡ async support - model_dump_anonymized_async() with async mask function support
• ✅ reliable - 105 tests covering all scenarios (99% coverage)
• 🪶 minimal dependencies - only pydantic>=2.0 and phonenumbers>=8.13

---

📖 usage

basic usage

from pydantic import BaseModel, Field
from pydantic_anonymizer import Anonymizer

class User(BaseModel, Anonymizer):
    name: str
    email: str = Field(json_schema_extra={"anonymize": True})

user = User(name="ivan", email="[e-mail]")

# original data
user.model_dump()  # {'name': 'ivan', 'email': '[e-mail]'}

# masked data
user.model_dump_anonymized()  # {'name': 'ivan', 'email': 'i***@***.com'}

# JSON string
user.model_dump_json_anonymized()  # '{"name": "ivan", "email": "i***@***.com"}'

nested models

from pydantic import BaseModel, Field
from pydantic_anonymizer import Anonymizer

class Address(BaseModel, Anonymizer):
    city: str
    street: str = Field(json_schema_extra={"anonymize": True})

class UserProfile(BaseModel, Anonymizer):
    name: str
    address: Address

user = UserProfile(
    name="ivan",
    address=Address(city="Moscow", street="Lenina 1")
)

result = user.model_dump_anonymized()
# {'name': 'ivan', 'address': {'city': 'Moscow', 'street': 'L***a 1'}}

lists of models

from pydantic import BaseModel, Field
from pydantic_anonymizer import Anonymizer

class Card(BaseModel, Anonymizer):
    number: str = Field(json_schema_extra={"anonymize": "card"})

class Wallet(BaseModel, Anonymizer):
    cards: list[Card]

wallet = Wallet(cards=[
    Card(number="4242111122223333"),
    Card(number="5555666677778888")
])

result = wallet.model_dump_anonymized()
# {'cards': [{'number': '4242-****-****-3333'}, {'number': '5555-****-****-8888'}]}

custom strategies

from pydantic import BaseModel, Field
from pydantic_anonymizer import Anonymizer, MaskRegistry

# register your own masking function
def mask_ssn(value: str) -> str:
    return "***-**-" + value[-4:]

MaskRegistry.register("ssn", mask_ssn)

class Person(BaseModel, Anonymizer):
    ssn: str = Field(json_schema_extra={"anonymize": "ssn"})

person = Person(ssn="123-45-6789")
person.model_dump_anonymized()  # {'ssn': '***-**-6789'}

@Anonymize decorator

alternative to mixin - decorator adds model_dump_anonymized() without inheritance:

from pydantic import BaseModel
from pydantic_anonymizer import Anonymize

@Anonymize(email=True, card="card", phone="phone")
class Payment(BaseModel):
    email: str
    card: str
    phone: str

payment = Payment(email="[e-mail]", card="4242111122223333", phone="[phone]")
payment.model_dump_anonymized()
# {'email': 'a***@***.com', 'card': '4242-****-****-3333', 'phone': '+380 (***) ***-**-85'}

async support

async methods support both sync and async mask functions:

import asyncio
from pydantic import BaseModel, Field
from pydantic_anonymizer import Anonymizer, MaskRegistry

# async mask function (e.g., external API call)
async def mask_external(value: str) -> str:
    # masking logic with async I/O
    return "MASKED:" + value[:2] + "***"

MaskRegistry.register("external", mask_external)

class User(BaseModel, Anonymizer):
    email: str = Field(json_schema_extra={"anonymize": "external"})

async def process():
    user = User(email="[e-mail]")
    data = await user.model_dump_anonymized_async()
    json_str = await user.model_dump_json_anonymized_async()

asyncio.run(process())

---

🎭 built-in strategies

Strategy	Field	Input	Output
True / "email"	email	[e-mail]	i***@***.com
"card"	card number	4242111122223333	4242-****-****-3333
"phone"	phone	[phone]	+380 (***) ***-**-85
"ip"	IP address	192.168.1.100	192.168.***.***
"birthdate"	date of birth	15.03.1990	**.**.****
"name"	full name	John Doe	J*** D**
"iban"	IBAN	UA213996...6712	UA**...****6712

country code support

phone masking works correctly with any country codes:

Country	Code	Example
Ukraine	+380	[phone] → +380 (***) ***-**-85
USA	+1	[phone] → +1 (***) ***-**-71
UK	+44	[phone] → +44 (***) ***-**-56
Russia	+7	[phone] → +7 (***) ***-**-67
Germany	+49	[phone] → +49 (***) ***-**-78
China	+86	[phone] → +86 (***) ***-**-78

---

📝 examples

FastAPI logging

from fastapi import FastAPI
from pydantic import BaseModel, Field
from pydantic_anonymizer import Anonymizer

app = FastAPI()

class UserCreate(BaseModel, Anonymizer):
    username: str
    email: str = Field(json_schema_extra={"anonymize": True})
    card_number: str = Field(json_schema_extra={"anonymize": "card"})

@app.post("/users")
def create_user(user: UserCreate):
    # log with masking
    print(user.model_dump_anonymized())
    # {'username': 'admin', 'email': 'a***@***.com', 'card_number': '4242-****-****-3333'}

    # original data for processing
    return user.model_dump()

safe logging

import logging
from pydantic import BaseModel, Field
from pydantic_anonymizer import Anonymizer

logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

class Payment(BaseModel, Anonymizer):
    card_number: str = Field(json_schema_extra={"anonymize": "card"})
    amount: float

def process_payment(payment: Payment):
    # safe to log - card number is masked
    logger.info("payment: %s", payment.model_dump_anonymized())

    # work with original data
    charge_card(payment.card_number, payment.amount)

automatic masking in logs (AnonymizedFormatter)

import logging
from pydantic import BaseModel, Field
from pydantic_anonymizer import Anonymizer, AnonymizedFormatter

handler = logging.StreamHandler()
handler.setFormatter(AnonymizedFormatter("%(message)s"))

logger = logging.getLogger(__name__)
logger.addHandler(handler)
logger.setLevel(logging.INFO)

class User(BaseModel, Anonymizer):
    email: str = Field(json_schema_extra={"anonymize": True})

# automatically masks model in logs
logger.info("User: %s", User(email="[e-mail]"))
# outputs: User: {'email': 'i***@***.com'}

---

🧪 tests and coverage

# install dev dependencies
pip install -e ".[dev]"

# run tests
pytest

# run tests with coverage
pytest --cov=pydantic_anonymizer --cov-report=term-missing

---

📜 license

MIT