Skip to content

Changelog 🕓

0.0.13 - 2026-06-27

New Features

  • Timezone-aware datetimes: Pass any datetime.tzinfo subclass — including datetime.timezone constants and zoneinfo.ZoneInfo instances — through json_schema_extra={"timezone": ...} on a datetime field. Supporting attributes such as,
class Event(BaseModel):
    # stdlib UTC constant
    utc_ts: Annotated[datetime, Field(..., json_schema_extra={"timezone": timezone.utc})]
    # IANA timezone via zoneinfo
    ny_ts: Annotated[datetime, Field(..., json_schema_extra={"timezone": ZoneInfo("America/New_York")})]

0.0.12 - 2026-06-23

Major changes

  • Feature, allowing users to create more realistic and complex time series data. The implementation is designed to be extensible for future enhancements. (pyfake/core/timeseries.py)

  • Implemented support for noise, seasonality, anomalies, missing values, and value constraints in the Timeseries class. The class now includes methods to inject these features into generated data, and the constructor accepts new parameters for each

Docs

  • Added detailed documentation for the new timeseries features, including explanations of noise, seasonality, anomalies, missing values, and constraints, along with code examples and usage guidelines. (docs/timeseries/advanced-features.md)
  • Replaced mermaid chart with a static image for better presentation

0.0.11 - 2026-06-21

Major changes

  • Added a new Timeseries class to pyfake.core.timeseries, supporting generation of time series data with configurable start, periods, frequency, baseline, trend (upward, downward, flat), and reproducibility via seed. The implementation is extensible for future features like noise, seasonality, anomalies, and missing data, but currently implements only baseline and linear trend. (pyfake/core/timeseries.py)

  • The project now depends on Numpy for internal calculations, and the documentation has been updated to reflect this change. The installation verification instructions have been removed from the docs since Numpy is now a required dependency. (docs/index.md)

Docs

  • Added a comprehensive "Getting Started" guide for the new timeseries feature, including API usage, code examples, limitations, and tips for reproducibility. (docs/timeseries/getting-started.md)
  • Updated the main documentation to clarify that Numpy is now a requirement and removed outdated installation verification instructions. (docs/index.md)
  • Added a reference to the "how it works" section for further reading. (docs/index.md)

0.0.10 - 2026-04-19

Minor changes

  • Packaging and type hinting: Added a py.typed marker to the package data for pyfake in pyproject.toml, indicating that the package includes type hints for static type checkers.

0.0.9 - 2026-04-09

New Features - Support for multiple_of constraint for numeric fields, allowing users to specify that generated numbers must be multiples of a given value. This is implemented in the generator registry and integrated into the schema resolution process. - Improved support for the decimal_places option for both float and Decimal fields, with documentation clarifying floating-point precision nuances.

Docs - Added comprehensive new documentation pages for datetime and uuid usage, including examples, option explanations, and implementation notes. - Updated the README with a clearer quick example, a new feature matrix comparing pyfake to other generators, and a more concise project introduction.

0.0.8 - 2026-04-06

Major changes

  • Introduced the Fake class as a new, user-friendly API for generating fake data (from pyfake import fake), accessible via the fake singleton in pyfake/__init__.py. This class provides methods for generating data as dicts, models, or JSON, and supports seeding for reproducibility. (pyfake/core/api.py, pyfake/__init__.py)

New Features

  • Added support for the Decimal type, including generator registration and type mapping in the registry.
  • Improved metadata parsing for field constraints (e.g., pattern, decimal_places) and fixed a bug in how these are extracted from Pydantic field metadata.

Docs - Added comprehensive usage guides for booleans, numbers (including constraints and Decimal), and strings, with examples and notes on supported/unsupported features. - Updated the main README.md with new documentation links and fixed minor formatting issues.

0.0.7 - 2026-04-02

Major changes

  • Recursive generator support for nested Pydantic models
  • Recursive resolver for resolving nested Pydantic models in the input Schema

New Features

  • Support for datetime, date & time formats
  • Ability to set bounds using pydantic Field() function
  • Flag based return for either accepting a dictionary or the input Pydantic model (as_dict = True)
  • Updated ReadTheDocs documentation
  • Support for list, set and tuple types with length constraints using pydantic Field() function
  • Moving documentations to zensical for better experience and future support.

0.0.6 - 2026-01-30

Major changes

  • Migrated the entire project to uv for faster, deterministic dependency management and installs.
  • Introduced a new modular and expandable architecture, separating concerns into:
  • engine – execution and orchestration layer
  • registry – type to generator resolution
  • context – shared runtime state and randomness control

New Features

  • Added full support for UUID generation across all formats:
  • UUID1, UUID3, UUID4, UUID5, UUID6, UUID7, UUID8
  • Added support for legacy primitive types:
  • int
  • float
  • str
  • Introduced a reproducible generation context, allowing users to set a seed for deterministic fake data generation.

Testing

  • Added comprehensive test coverage across all generators and core components.
  • Tests now validate:
  • Correct type resolution
  • Deterministic output with seeded contexts
  • UUID format correctness
  • Edge cases across primitive and complex generators

0.0.5 - 2025-06-16

  • Initial release