Skip to content

Latest commit

 

History

History
213 lines (183 loc) · 7.88 KB

CHANGELOG.rst

File metadata and controls

213 lines (183 loc) · 7.88 KB
Raw

Changelog

Rev 1.1

Rev 1.1.0

  • Added new flavors to the Burning ship impelmentations fractalshades.models.Burning_ship and fractalshades.models.Perturbation_burning_ship: “Celtic” and "Buffalo"
  • Added projection-related classes: abstract base class fractalshades.projection.Projection and its derived classes fractalshades.projection.Cartesian fractalshades.projection.Expmap fractalshades.projection.Generic_mapping
  • Added database export capabilities fractalshades.Fractal_plotter.save_db, fractalshades.db.Db, fractalshades.db.Frame, fractalshades.db.Exp_db, fractalshades.db.Exp_frame.
  • Added movie making API: fractalshades.movie.Movie, fractalshades.movie.Camera_pan, fractalshades.movie.Camera_zoom, fractalshades.movie.Custom_sequence.
  • Renamed function fractalshades.settings.remove_decompression_size_check to: fractalshades.settings.disable_decompression_size_check
  • Improved memory management of BLA iterations. Now images with up to 500 millions iterations can be computed on a system with 16 Mo RAM.

Rev 1.1.1

Bugfix revision, patch for numba version >0.57 (deprecated generated_jit)

Rev 1.0

Rev 1.0.3

Yet another bugfix.

Rev 1.0.2

Bugfix revision, fixing OS different behavior with memory mappings #2.

Rev 1.0.1

Bugfix revision, fixing OS Windows different behavior with memory mappings.

Rev 1.0.0

This revision is a major refactoring of the code, with optimised high-resolution rendering options. Several non backward-compatible changes have been introduced. If you use the GUI pre-made scripts available in the documentation section, you will need to update them (i.e., download them again) as the scripts running with 0.5.x will not be compatible with version 1.x.y onwards.

  • fractalshades.Fractal_plotter now implements supersampling and jitter parameters
  • antialiasing parameter from fractalshades.Fractal.zoom method deprecated: use fractalshades.Fractal_plotter supersampling parameter instead
  • BLA_params dict-like parameter replaced by a float parameter: BLA_eps
  • in fractalshades.colors.layers.Blinn_lighting.add_light_source method: 2-uplet float parameters angles replaced by 2 float parameters polar_angle, azimuth_angle
  • in fractalshades.colors.layers.Blinn_lighting.add_light_source: coords parameter deprecated
  • fractalshades.Fractal saved data files: files *.params renamed as *.fingerprint
  • fractalshades.postproc.Fieldlines_pp class: implementation entirely rewritten, parameter damping_ratio renamed endpoint_k
  • run method from fractalshades.Fractal removed. Calculation is automatically trigered through fractalshades.Fractal_plotter.plot method
  • in fractalshades.colors.layers.Color_layer, parameter probes_kind suppressed, all probes are now "absolute".
  • fractalshades.colors.layers.Normal_map_layer : A correction of the max slope calculation may impact shaded renderings (especially glossy renders). Usually a smaller slope angle is now needed to be used to achieve visually similar result (e.g. 45. -> 30.).
  • Color parameters typing : now use a dedicated type fscolors.Color type is used instead of QtGui.QColor: this allows to run batch files on systems without pyQt6 installed.
  • Avoiding banding for very high iteration numbers (banding may occur, usually above 10 million iterations): it is now possible to suppress banding by switching to float64 datatype for the postprocessing phase : see parameter fractalshades.settings.postproc_dtype

Regarding the Fractal models implemented:

  • Power n mandelbrot is now implemented in both standard and arbitrary precision and accessible from the GUI (in standard precision it features calculation of the limit cycle period and attaractivity with Newton method.)
  • The different Burning Ship variants have been factored into a single class with a flavor parameter to select the appropriate variant.

Regarding the Fractal post-processing implemented:

  • The fractalshades.postproc.Fieldlines_pp post-processing has been re-implemented as a truncated orbit average. It shall be used together with the calc_orbit and backshift parameters from the basic iteration computations (to define a starting point for the truncated orbit before divergence)

Several changes have also been implemented in the GUI:

  • Collapsible parameters groups (main panel) for easier selection
  • Dedicated GUI-editor for fractalshades.colors.layers.Blinn_lighting instances
  • Dedicated GUI-editor for fractalshades.Fractal subclasses instances. This allows the interactive selection of different fractal variant: exponent for Mandelbrot power n > 2 implementations (fractalshades.models.Mandelbrot_N, or fractalshades.models.Perturbation_mandelbrot_N) variant -flavor- of the Burning ship implementations (fractalshades.models.Burning_ship, fractalshades.models.Perturbation_mandelbrot_N) fractalshades.models.Perturbation_burning_ship)
  • Parametrized "functions" can now be passed (as string)
  • A GUI template function fractalshades.gui.guitemplates.std_zooming is now available and suitable for most of the implemented fractal models ; as a result lanching a GUI with all options for a fractal becomes almost a one-liner (see the updated GUI examples).
  • Saving and reloading fractalshades.colors.Fractal_colormap in binary format (through Python standard serialization module pickle ) is now available from the GUI (in addition to the already available python source-code format output)

Application-level logging has also been implemented (it may target both the standard output pipe and user-defined log files). Refer to fractalshades.settings.log_directory and fractalshades.log.set_log_handlers .

Please report any bug or unexpected behavior at: https://github.com/GBillotey/Fractalshades/issues

Rev 0.5

Rev 0.5.6

  • Added Shark Fin escape-time fractal (incl. arbitrary precision implementation)
  • "Show source" export button from the GUI now gives a ready-to-run script
  • Fixed bug in GUI where a modified colormap was not updated correctly
  • Early interior points detection for Mandelbrot deep zoom

Rev 0.5.5

  • Quick estimation of skew when no mini (based on a diverging point)
  • Improved documentation

Rev 0.5.4

  • Dark mode for html doc
  • Remove debugging test file ("knwon fails") from source distribution

Rev 0.5.3

  • Added more examples and reorganised the documentation
  • Fixed bug in Perpendicular Burning Ship

Rev 0.5.2

  • Added Perpendicular Burning Ship in arbitrary precision Rev 0.5.1
  • Added unskew option in interactive mode
  • Added Tetration (power tower) fractal

Rev 0.5.0

Rev 0.4

Rev 0.4.3

  • fixed typo in run_interactive.py

Rev 0.4.2

  • fixed concurrent.futures import

Rev 0.4.1

  • Improved Fieldlines default postproc when using mirrored cmap
  • Typo name Disp_Layer -> Disp_layer
  • added Collaz fractal
  • use gmpy2 bindings through Cython C-extension for faster full-precision calculations
  • improved glitch correction : use single-reference based method
  • use multi-threading + NOGIL compilation to improve portability under Windows
  • build under windows
  • added progress status bar
  • cleaner separation of parameters by themes in GUI
  • rotation in GUI
  • Newton search in GUI
  • When one quit and relaunch the GUI, all previous parameters are reloaded