Skip to content

Releases: davewalker5/ChessAnalyser

Version 1.23.0

Choose a tag to compare

@davewalker5 davewalker5 released this 03 Sep 06:55
e3deeeb
  • This release contains dependency updates that address known security vulnerabilities reported by Dependabot
  • There are no changes to application functionality or behaviour

Version 1.22.0

Choose a tag to compare

@davewalker5 davewalker5 released this 26 Aug 18:58
fb13993

PGN Editor UI Improvements

This release improves the PGN Editor’s layout, game information, and new-game workflow.

What’s new

Captured material display

The editor now shows captured pieces and material totals above and below the board.

  • In the normal orientation, Black’s captures appear above the board and White’s captures below it
  • Flipping the board also reverses the placement of the summaries
  • Captured pieces use conventional values
    • Pawn: 1
    • Knight: 3
    • Bishop: 3
    • Rook: 5
    • Queen: 9
  • The summaries follow the currently displayed position during navigation and playback
  • Ordinary captures, en passant, promotion captures, and captured promoted pieces are handled correctly

Automatic game results

The PGN Result field is now updated automatically when a move ends the game.

  • Checkmate sets the result to 1-0 or 0-1
  • Draws recognised by python-chess, including stalemate, set the result to 1/2-1/2
  • Undoing the terminal move resets the result to *, even when Undo is selected while viewing an earlier position
  • Results can still be entered manually for resignation, agreed draws, and other outcomes not determined by the board position

New Game command

A new File → New Game command is available, with the Ctrl+N keyboard shortcut.

Starting a new game:

  • Clears all moves
  • Restores the standard starting position
  • Resets the PGN game details
  • Resets captured-material totals
  • Removes the current filename
  • Creates a clean, untitled document

If the current game contains unsaved changes, the editor offers the existing Save, Discard, and Cancel choices. A cancelled or unsuccessful save leaves the current game unchanged.

Clear Game improvements

Clear Game now resets both the move history and game metadata while retaining the current filename.

The editor asks for confirmation when clearing would discard moves or non-default game details. After clearing, saving continues to target the currently associated PGN file.

Version 1.21.0

Choose a tag to compare

@davewalker5 davewalker5 released this 24 Aug 15:35
e4f956c

Chess Analyser v1.21.0 introduces a new graphical PGN Editor providing a full interactive chessboard, legal move validation, PGN metadata editing and game playback, all wrapped in a dark desktop interface with cyan accents.

Introducing the PGN Editor

Games can be entered directly on a resizable graphical chessboard using drag-and-drop move entry.

Moves are validated against the current position using python-chess, so the editor understands the full rules of chess — including:

  • Captures
  • Castling
  • En passant
  • Pawn promotion
  • Checkmate
  • Stalemate

Legal destinations and captures are highlighted as pieces are moved, while the game is built automatically in Standard Algebraic Notation.

Edit, Explore and Replay

The editor provides facilities for navigating backwards and forwards through a game without altering it, or for replaying the entire game automatically at an adjustable speed from 0.25 to 5 seconds per move.

Controls are also provided for:

  • Undoing moves
  • Clearing the board
  • Flipping the board orientation
  • Moving through the game
  • Starting and stopping automatic playback

Full PGN Metadata Editing

PGN headers can be edited directly from the interface, including:

  • Event
  • Site
  • Date — with calendar picker
  • Round
  • White
  • Black
  • Result

Safer File Handling

PGN games can be opened and saved directly from the editor.

File operations use transactional error handling, and the editor keeps track of unsaved changes so an edited game can't be accidentally discarded without warning.

Starting the PGN Editor

After installing or updating Chess Analyser, launch the editor with:

pgn-editor

or:

python -m pgn_editor

The existing command-line Chess Analyser remains available through:

chess-analyser

and is completely unaffected by the addition of the graphical editor.

Third-Party Notices

Third-party notices and applicable licence texts have also been added or updated.

Version 1.20.0

Choose a tag to compare

@davewalker5 davewalker5 released this 24 Aug 08:45
af1c71f
  • This release contains dependency updates that address known security vulnerabilities reported by Dependabot
  • There are no changes to application functionality or behaviour

Version 1.19.1

Choose a tag to compare

@davewalker5 davewalker5 released this 24 Jun 10:04
692c8d0
  • Fixup movie export issues after pillow/moviepy update

Version 1.19.0

Choose a tag to compare

@davewalker5 davewalker5 released this 24 Jun 08:39
8437058
  • Applied the latest dependabot dependency recommendations

Version 1.18.2

Choose a tag to compare

@davewalker5 davewalker5 released this 26 Jan 18:13
1401e8d

Write Model Parameters to JSON

  • When the performance model notebooks run, they now output the static and calculated model parameters to a JSON file
  • The "performance over time" notebook reads the multi-engine model parameters from that exported file
  • The "run all" script ensures the notebooks are run in the correct order

Version 1.18.1

Choose a tag to compare

@davewalker5 davewalker5 released this 26 Jan 12:02
f48fe8c

Performance Model Run Scripts

  • Created scripts to run/schedule the performance model notebooks via Papermill

Version 1.18.0

Choose a tag to compare

@davewalker5 davewalker5 released this 24 Jan 12:25
455882e

Personal Performance Rating Report

  • Added an "average personal performance rating" by month report
  • This uses the performance model added in versions 1.16.0 and 1.17.0

Version 1.17.0

Choose a tag to compare

@davewalker5 davewalker5 released this 23 Jan 08:22
7b9bf6b

Handling Low ACPL

  • Implemented an ACPL floor in the performance model:
    • The performance model is sensitive to very low ACPL values
    • Yet there are real-world circumstances that ca yield such values (some examples are given in the performance modelling README)
    • In recognition of this, the model now applies a floor to ACPL values, based on game length
  • Added the chessprf performance calculator for the TI-84 Plus CE Python Edition calculator, using the model parameters to calculate personal performance ratings based on ACPL and game length