Farsi Faker | فارسی فیکر

✨ Features

• Embedded name database — Persian first and family names shipped with the package
• Gender-specific generation — separate male and female first-name pools
• Dynamic extension — extend_name_pools() to merge your own names at runtime
• Profile fields — national ID (کد ملی with checksum), mobile, email, postal code, address
• Banking fields — IBAN (ISO 13616, Luhn mod-97), Shetab/Shaparak bank card (Luhn-10), bank name
• Vehicle fields — car plate number (پلاک خودرو) + registration province
• CLI — python -m farsi_faker for JSON/CSV fixtures
• Fast cold start — pickle-backed name cache, shared across instances
• Reproducible — seed support for stable fixtures
• Zero required dependencies — production install needs only the standard library
• Concurrent instantiation — shared name cache is lock-guarded and immutable
• Typed — type hints plus a py.typed marker (PEP 561)
• Tested — unit, packaging, concurrency, profile, and data-quality gates in CI
• Unicode — Persian/Farsi text output
• Optional pandas — DataFrame output for data-science workflows
• Cleaning helpers — farsi_faker.cleaning for OCR-split and ZWNJ repair

Data provenance: see DATA_PROVENANCE.md. Core common names (محمد, فاطمه, …) ship in the embedded pool as of v1.5.0.

📦 Installation

From PyPI (Recommended)

pip install farsi-faker

With pandas support (for DataFrame output)

pip install farsi-faker[pandas]

From Source

git clone https://github.com/alisadeghiaghili/farsi-faker.git
cd farsi-faker
pip install -e ".[dev]"

Requirements

• Python 3.9+
• No external dependencies for production use
• Optional: pandas for DataFrame output (as_dataframe=True)

🚀 Quick Start

Basic Usage

from farsi_faker import FarsiFaker

faker = FarsiFaker()

# Generate a random person
person = faker.full_name()
print(person)
# {'name': 'علی صادقی عقیلی', 'first_name': 'علی', 'last_name': 'صادقی عقیلی', 'gender': 'male'}

# Generate male name
male = faker.full_name('male')
print(male['name'])   # علی صادقی عقیلی

# Generate female name
female = faker.full_name('female')
print(female['name'])  # سپیده جلیلی

Synthetic Profile Fields (v1.3.0+)

from farsi_faker import FarsiFaker, is_valid_national_id

faker = FarsiFaker(seed=42)
person = faker.profile('male')
# {
#   'name': '...', 'first_name': '...', 'last_name': '...', 'gender': 'male',
#   'national_id': '10-digit کد ملی',
#   'mobile': '0912...',
#   'email': 'first.last@gmail.com',
#   'postal_code': '10-digit',
# }
assert is_valid_national_id(person['national_id'])
assert person['mobile'].startswith('09')

Banking Fields (v1.8.0+)

from farsi_faker import FarsiFaker, is_valid_iban, is_valid_bank_card

faker = FarsiFaker(seed=42)

iban = faker.iban()                  # 'IR..' — 28 chars, Luhn mod-97 valid
card = faker.bank_card_number()      # '6277..' — 16 digits, Luhn-10 valid
bank = faker.bank_name()             # e.g. 'بانک ملت'

assert is_valid_iban(iban)
assert is_valid_bank_card(card)

Vehicle Fields (v1.9.0+)

from farsi_faker import FarsiFaker, is_valid_car_plate

faker = FarsiFaker(seed=42)

plate = faker.car_plate_number()   # 'بث ۱۲۳۴' — two letters + four digits
prov  = faker.car_province()       # e.g. 'اصفهان'
rec   = faker.vehicle_record()     # {'plate': 'بث ۱۲۳۴', 'province': 'تهران'}

assert is_valid_car_plate(plate)

CLI

python -m farsi_faker --count 5 --seed 42
python -m farsi_faker --count 100 --profile --format csv > people.csv
python -m farsi_faker --count 3 --gender female --format json

Generate Multiple Names

# 10 random names as a list (default)
people = faker.generate_names(10)

# 50 male names as a list
men = faker.generate_names(50, 'male')

# 30 female names as a pandas DataFrame
women_df = faker.generate_names(30, 'female', as_dataframe=True)
print(women_df.shape)          # (30, 4)
print(list(women_df.columns))  # ['name', 'first_name', 'last_name', 'gender']
print(women_df.head(2))
#          name first_name last_name  gender
# 0  فاطمه احمدی     فاطمه    احمدی  female
# 1  زینب رضایی      زینب    رضایی  female

Generate Balanced Dataset

# 100 people with 60% male ratio — as a list
dataset = faker.generate_dataset(100, male_ratio=0.6)
print(len(dataset))   # 100

# Same, but as a pandas DataFrame
df = faker.generate_dataset(500, male_ratio=0.5, as_dataframe=True)
print(df.shape)                      # (500, 4)
print(df['gender'].value_counts())
# male      250
# female    250
# Name: gender, dtype: int64

Reproducible Results

faker1 = FarsiFaker(seed=42)
faker2 = FarsiFaker(seed=42)
assert faker1.full_name() == faker2.full_name()  # True

Quick One-Off Generation

from farsi_faker import generate_fake_name

person = generate_fake_name('male')
print(person['name'])  # علی صادقی عقیلی

📖 Documentation

Class: FarsiFaker

Main class for generating Persian names.

Constructor

FarsiFaker(seed: Optional[int] = None)

Parameters:

• seed (int, optional): Random seed for reproducible results

Example:

faker = FarsiFaker()        # random
faker = FarsiFaker(seed=42) # reproducible

male_first_name() -> str

Return a random male first name.

faker.male_first_name()  # 'محمد'

female_first_name() -> str

Return a random female first name.

faker.female_first_name()  # 'فاطمه'

first_name(gender=None) -> Tuple[str, str]

Return a first name with its normalised gender.

Parameters:

• gender (str, optional): Any supported gender token (see Gender Input Options)

Returns: (name, gender) — gender is always 'male' or 'female'

name, g = faker.first_name('male')
# ('علی', 'male')

name, g = faker.first_name()   # random gender
# ('مریم', 'female')

last_name() -> str

Return a random Persian family name.

faker.last_name()  # 'احمدی'

full_name(gender=None) -> Dict[str, str]

Return a complete person record.

Returns: dict with keys name, first_name, last_name, gender

person = faker.full_name('female')
# {
#     'name': 'سپیده جلیلی',
#     'first_name': 'سپیده',
#     'last_name': 'جلیلی',
#     'gender': 'female'
# }
assert person['name'] == person['first_name'] + ' ' + person['last_name']

generate_names(count=10, gender=None, as_dataframe=False)

Generate multiple full-name records.

Parameters:

• count (int, default 10): Number of records to generate
• gender (str, optional): Gender applied to all records; random mix when None
• as_dataframe (bool, default False): Return a pandas.DataFrame instead of a list

Returns: List[Dict] or pandas.DataFrame with columns ['name', 'first_name', 'last_name', 'gender']

Raises: ValueError if count ≤ 0; ImportError if as_dataframe=True and pandas is not installed

# List (default)
people = faker.generate_names(5, 'male')
assert len(people) == 5
assert all(p['gender'] == 'male' for p in people)

# DataFrame
df = faker.generate_names(100, as_dataframe=True)
assert df.shape == (100, 4)
assert list(df.columns) == ['name', 'first_name', 'last_name', 'gender']
assert not df.isnull().any().any()
assert (df['name'] == df['first_name'] + ' ' + df['last_name']).all()

generate_dataset(count=100, male_ratio=0.5, as_dataframe=False)

Generate a balanced dataset with a configurable gender ratio.

Parameters:

• count (int, default 100): Total number of records
• male_ratio (float, default 0.5): Fraction of male records in [0.0, 1.0]
• as_dataframe (bool, default False): Return a pandas.DataFrame instead of a list

Returns: Shuffled List[Dict] or pandas.DataFrame

Raises: ValueError if count ≤ 0 or male_ratio outside [0.0, 1.0]; ImportError if pandas missing and as_dataframe=True

# List (default)
dataset = faker.generate_dataset(10, male_ratio=0.6)
assert len(dataset) == 10
assert sum(1 for p in dataset if p['gender'] == 'male') == 6

# DataFrame
df = faker.generate_dataset(100, male_ratio=0.5, as_dataframe=True)
assert df.shape == (100, 4)
assert df['gender'].value_counts().to_dict() == {'male': 50, 'female': 50}

# Edge cases
assert all(p['gender'] == 'female' for p in faker.generate_dataset(5, male_ratio=0.0))
assert all(p['gender'] == 'male'   for p in faker.generate_dataset(5, male_ratio=1.0))

get_stats() -> Dict[str, int]

Return statistics about the embedded names database.

Returns: dict with keys male_names_count, female_names_count, last_names_count, total_names, possible_combinations

stats = faker.get_stats()
assert stats['possible_combinations'] == \
    (stats['male_names_count'] + stats['female_names_count']) * stats['last_names_count']
print(f"Possible combinations: {stats['possible_combinations']:,}")
# Possible combinations: 21,000,000

Function: generate_fake_name(gender=None, seed=None) -> Dict[str, str]

Convenience wrapper for one-off generation. For bulk generation prefer a FarsiFaker instance directly.

from farsi_faker import generate_fake_name

p1 = generate_fake_name('female', seed=99)
p2 = generate_fake_name('female', seed=99)
assert p1 == p2  # reproducible

🎨 Examples

Example 1: Django test fixtures

from farsi_faker import FarsiFaker
from myapp.models import User

faker = FarsiFaker(seed=42)
for person in faker.generate_dataset(100, male_ratio=0.5):
    User.objects.create(**person)

Example 2: Export to CSV

import csv
from farsi_faker import FarsiFaker

faker = FarsiFaker()
with open('people.csv', 'w', encoding='utf-8', newline='') as f:
    writer = csv.DictWriter(f, fieldnames=['name', 'first_name', 'last_name', 'gender'])
    writer.writeheader()
    writer.writerows(faker.generate_dataset(1000, male_ratio=0.6))

Example 3: pandas DataFrame for data science

from farsi_faker import FarsiFaker

faker = FarsiFaker(seed=123)
df = faker.generate_dataset(500, male_ratio=0.55, as_dataframe=True)

print(df.shape)                          # (500, 4)
print(df['gender'].value_counts())       # male 275 / female 225
print(df.groupby('gender')['last_name'].nunique())

Example 4: pytest fixture

import pytest
from farsi_faker import FarsiFaker

@pytest.fixture
def fake_users():
    return FarsiFaker(seed=42).generate_dataset(10, male_ratio=0.5)

def test_user_creation(fake_users):
    assert len(fake_users) == 10
    assert all('name' in u for u in fake_users)

Example 5: Flask mock API

from flask import Flask, jsonify
from farsi_faker import FarsiFaker

app = Flask(__name__)
faker = FarsiFaker()

@app.route('/api/users/random')
def random_user():
    return jsonify(faker.full_name())

@app.route('/api/users/<int:count>')
def multiple_users(count):
    return jsonify(faker.generate_names(min(count, 100)))

🎯 Gender Input Options

📊 Database Statistics

from farsi_faker import FarsiFaker

stats = FarsiFaker().get_stats()
print(f"Male names:            {stats['male_names_count']:,}")
print(f"Female names:          {stats['female_names_count']:,}")
print(f"Last names:            {stats['last_names_count']:,}")
print(f"Total names:           {stats['total_names']:,}")
print(f"Possible combinations: {stats['possible_combinations']:,}")

🧪 Testing

pip install -e ".[dev]"
pytest tests/ -v
pytest tests/ --cov=farsi_faker --cov-report=html

🛠️ Development

git clone https://github.com/alisadeghiaghili/farsi-faker.git
cd farsi-faker
python -m venv venv && source venv/bin/activate
pip install -e ".[all]"

# quality checks
black farsi_faker/ && isort farsi_faker/ && mypy farsi_faker/
pytest tests/ -v

📁 Project Structure

farsi-faker/
├── farsi_faker/
│   ├── __init__.py
│   ├── faker.py          ← core class
│   ├── _version.py
│   └── data/names.pkl
├── tests/test_faker.py
├── scripts/create_pickle.py
├── setup.py
├── pyproject.toml
├── CHANGELOG.md
└── README.md

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Add tests for new functionality
4. Run tests (pytest tests/)
5. Commit (git commit -m 'Add amazing feature')
6. Push and open a Pull Request

Code style: Black + isort. Type hints required. Docstrings required.

📄 License

MIT — see LICENSE.

📞 Contact

• Author: Ali Sadeghi Aghili
• GitHub: alisadeghiaghili/farsi-faker
• PyPI: pypi.org/project/farsi-faker
• Issues: github.com/alisadeghiaghili/farsi-faker/issues