CHANGELOG.md

v0.10.0

• Add new scatter plot markers: "|", and "-"
• Add drag-and-drop functionality (drag_drop_source, drag_drop_target)
• Add TF.hover attribute for more robust hover functionality than Image.hover_tag.
• Re-introduce the maximizied window flag, because just setting window size to a large number doesn't properly work on Windows
• Fix a crash on slider and drag widgets

Breaking Changes

Remove the name argument from c.child

v0.9.3

• concur.core.replace, concur.core.replace_tag, concur.core.replace_value added. These make it convenient to transform events without lambdas. Lambdas don't work properly inside loops/comprehensions in Python.

v0.9.2

• concur.widgets.dummy implemented
• concur.widgets.input_text flags argument added. This enables creating a password input field, among other use-cases.
• Make the imageio dependency non-mandatory.

v0.9.1

• Update ImGui to the newest version in the docking branch
• Improve some docstrings

v0.9.0

Mainly interactive invocation and documentation improvements.

Breaking Changes

• c.main argument order switched up to this. Widget is now first, then the window name (newly optional), then width and height (also newly optional). This simplifies the Hello, world application and other simple examples considerably.

Fixes

• Destroy contexts on main exit correctly. This enables multiple main invocations with fresh ImGui state.
• Better cleanup on exceptions. Sigkill in IPython doesn't hang the window.

v0.8.0

Breaking Changes

• concur.extra_widgets.frame has new arguments width, height consistent with concur.extra_widgets.image.
• PanZoom-derived widgets now handle hover, down, and drag events in a different way. Instead of being yielded, they are passed to the content_gen as a widget. This makes it possible to react to the events in content_gen, which is better for modularization, and it is newly possible to use the information in tf as a part of the reaction. The cost is more complicated content_gen signature. Better solution might be possible. Affected widgets:
  • concur.extra_widgets.pan_zoom
  • concur.extra_widgets.image
  • concur.extra_widgets.frame
• PanZoom-derived widgets are now passed a keyword argument tf, instead of a positional argument. This will break code which used other names for the transformation info. Affected widgets:
  • concur.extra_widgets.pan_zoom
  • concur.extra_widgets.image
  • concur.extra_widgets.frame
• concur.draw.image argument list was reworked. It now takes additional mandatory arguments x and y, and non-mandotory arguments uv_a, uv_b.

Additions

• concur.widgets.font: render a widget with a specific font

Fixes

• concur.extra_widgets.image now correctly handles images with dimensions not divisible by 4. Previously, those may get severely distorted.
• concur.extra_widgets.image now has transparent background even for images without alpha channel.
• concur.widgets.key_press: fix a a crash on glfw.KEY_SPACE keypress.
• Make the system clipboard work by upgrading PyImGui to v1.3.1.
• Add a call to refresh_font_texture to GLFW initialization. This enables user to specify a font and have it integrated, and is hopefully harmless in all other cases.
• Fix a crash on certain greyscale NPOT images
• Fix UI ghosts appearing because the buffer was not cleared.

v0.7.1

• New windows are now fixed size on creation, if they aren't in imgui.ini. Previously, they were fit to contents, which works OK only when the contents do not fit to window. In the common case of a window containing only an image widget, the window would have been created outright tiny. This could lead to user not even noticing it.
• Add:
  • widget tree_node
  • widget input_text_multiline
  • widget text_wrapped
  • scatter square markers ('s')

v0.7.0

Breaking Changes

• Change colored widgets text_colored and color_button to use the same color spec as the draw functions.
• Rename key_pressed to key_press
• Remove TF.hovered. Left mouse click & drag is now handled by the pan_zoom object.
• Add left mouse interaction support to image and frame. New arguments drag_tag and down_tag, and hover_tag.
  • pan_zoom.is_hovered is no longer present.
  • image event handling was substantially reworked. Starting the drag outside image window no longer triggers panning.
• key_press now respects local keyboard layout

Other changes

• add TF.inv_transform to tranasform from screen-space to content-space
• Make scatter also work with empty one-dimensional arrays
• Add an invisible_button, and an image_button
• Add selectable text
• Enable empty polylines
• Add drag_int2, drag_int3, and drag_int4 widgets
• Add modifier key support to key_press
• Add main_menu_bar, menu, and menu_item

v0.6.4

• Fix a bug where long polylines were sometimes corrupted (use-after-free).

v0.6.3

• Move the documentation and homepage to a separate repository

v0.6.2

• Implement the square marker '.' for scatter plot.
• Implement draw.polygon and draw.polygons - filled convex polygons.

v0.6.1

• Add concur.widgets.mouse_click widget, which returns mouse position on click
• Add c.quick_* functions for quick asynchronous plotting without worrying about the threading and event details
• Add a fps argument to c.main
  • this is breaking if screencast arguments were used positionally.
• Flexible color specification including xkcd strings. Colors like (0,1,1), 'dark red', ('blue', 0.5), or 0xffaa0000 are now possible.

v0.6.0

Breaking Changes

• Change the argument order for c.draw.text from (x, y, color, text) to (text, x, y, color).

Other Changes

• Improve frame tick labels
• Add this changelog
• Add c.draw.polylines for optimized multiple-polyline rendering. It is possible to draw ~100k lines in 60 FPS using this function (instead of ~500 lines). Using this function, it was possible to implement:
  • c.draw.ellipses for optimized multiple ellipse drawing
  • c.draw.rects for optimized multiple rectangle drawing
  • c.draw.scatter for scatter plots with fancy markers
• Add experimental benchmarking setup