Kinora 0.2.1

A maintenance release focused on the agent geometry, playback performance, and a more robust dependency installer.

Highlights

• Cylinder agents. Agents now render as cylinders instead of spheres, with a resilient shared mesh and full artefact cleanup between loads.
• Faster playback. Agent visibility is toggled only when it actually changes, fixing slow scrubbing and playback on larger scenes.
• More robust dependency installer. The addon now reuses Blender's own bundled numpy instead of swapping it at runtime, avoiding numpy-version collisions across Blender 4.x / 5.0 / 5.1.

Fixes

• HDF5 loading. pedpy's archive loaders are now called with the trajectory_file keyword, restoring compatibility with current pedpy.
• Clean reloads. All Kinora artefacts are wiped at the start of each load, so reloading no longer leaves stale objects behind.

Under the hood

• CI now matrix-tests installation across Ubuntu and macOS on Blender 5.0 and 5.1.
• README slimmed down; detailed docs moved to DEVELOPER.md.
• Release zip excludes leftover dependency folders.

Install

Download Kinora_0.2.1.zip below, then in Blender (4.0+): Edit > Preferences > Add-ons > Install..., pick the ZIP, enable Kinora, expand it and click Install Dependencies (give it a minute or two), then restart Blender.

Full Changelog: 0.2.0...0.2.1

Kinora 0.2.0

First release under the new name. The addon formerly known as BlenderJPS is now Kinora — see #11 for the naming discussion. Same project, same authors, broader positioning: a Blender visualiser for pedestrian-dynamics trajectory data (JuPedSim simulation output today; HDF5 from experimental sources and other simulators welcome).

Highlights

• Rebranded to Kinora — package, addon name, sidebar tab, operator IDs, collection names, logs, and docs all renamed. JuPedSim references are kept where they describe the simulator or its file format.
• HDF5 loading now works on Blender 4.x. Blender ships its own (older) numpy in site-packages, which used to shadow the numpy our deps installer pulled in for pedpy/scipy. The addon now evicts shadowed dependencies from sys.modules at load time so our bundled copy wins, regardless of which numpy Blender bundles (1.24 in 4.x, 1.26 in 5.0, 2.x in 5.1+).
• New Kinora logo (dark / light / bracketed B&W variants) by Fabian.

⚠️ Breaking changes — read before upgrading

If you have scripts or .blend files that reference the addon's old identifiers, they will need updating:

Old (BlenderJPS 0.1.x)	New (Kinora 0.2.0)
jupedsim.load_simulation / jupedsim.select_file / jupedsim.install_dependencies / jupedsim.uninstall_dependencies	kinora.load_simulation / kinora.select_file / kinora.install_dependencies / kinora.uninstall_dependencies
bpy.context.scene.jupedsim_props	bpy.context.scene.kinora_props
Collections JuPedSim_Agents / JuPedSim_Geometry	Kinora_Agents / Kinora_Geometry
Sidebar tab "JuPedSim"	"Kinora"
Module blender_jps	kinora

Installation

1. Download Kinora.zip below.
2. In Blender (4.0 – 5.0 supported; 5.1+ likely works, untested), open Edit → Preferences → Add-ons → Install... and pick the zip.
3. Enable Kinora - Pedestrian Data Visualiser.
4. Click Install Dependencies in the addon preferences. Restart Blender once it finishes.
5. The Kinora tab appears in the 3D Viewport sidebar (N).

If you previously had BlenderJPS installed, remove that addon first to avoid the old jupedsim.* operators sticking around.

Changes since 0.1.6

• Rebrand BlenderJPS → Kinora and bump version to 0.2.0 (#18, closes #11)
• Force addon's bundled numpy to win over Blender's site-packages copy (#18)
• Add Kinora logo SVGs (#18)
• README repositioned as pedestrian-data (not JuPedSim-only) visualiser (#18)
• Logo height tweak in README (#19)

Known issues / follow-ups

• Blender 5.1+ (Python 3.13, numpy 2.x) is not actively tested. The dependency-shadowing fix should handle it, but pedpy will need to publish its numpy-2 release before the install actually completes there.

Full diff: 0.1.6...0.2.0

blenderjps_v.0.1.5

BlenderJPS 0.1.5

Release notes for changes since v0.1.4.

Highlights

• Modular codebase: The monolithic operators.py has been split into focused modules — io/sqlite_reader.py, core/streaming.py, and core/geometry.py — making it straightforward to add new format loaders (e.g. HDF5).

Changes

Changed

• Extract SQLite I/O into io/sqlite_reader.py
• Extract streaming state and frame-change handler into core/streaming.py
• Extract Blender 3D object creation into core/geometry.py
• operators.py slimmed from ~873 to ~310 lines, acting purely as the modal orchestrator
• Reuse a cached SQLite cursor during streaming instead of recreating per frame

Fixed

• Reset frame_step to 1 on stream cleanup (was missing from reset logic)
• Close SQLite connection on add-on unregister to prevent resource leaks

Contributors

@FabianPlum
@chraibi — review feedback on PR #8

blenderjps_v.0.1.4

BlenderJPS 0.1.4

Release notes for changes since v0.1.3.

Highlights

• Uninstall Dependencies: You can now remove installed dependencies from addon preferences (Uninstall Dependencies button).
• Installation improvements: Dependency installation was refactored and is more reliable, including when using a development/symlink setup.
• CI and quality: The project now uses GitHub Actions for linting (Ruff) and Blender addon tests.

Changes

Added

• Uninstall Dependencies option in addon preferences
• CI workflow (Ruff lint + Blender addon tests) on push and PRs
• Improved addon metadata

Changed

• Refactored dependency installation (shared logic for UI and CI)
• More reliable dependency discovery so the operator finds deps correctly in all setups
• CI: Blender version updated; Blender 4 no longer tested in CI

Fixed

• Various indentation and formatting fixes
• Test path to example file

Contributors

• @chraibi (Mohcine Chraibi)
• @FabianPlum (Fabian Plum)

What's Changed

• CI: Add Blender runtime tests + ruff linting by @chraibi in #6

Full Changelog: 0.1.3...0.1.4

blenderjps_v.0.1.3

Release Notes

BlenderJPS v0.1.3

New features

New installation procedure

Installation no longer requires Administrator or root rights:

• Dependencies install into the addon folder — pedpy and numpy<2.0 are installed into the addon’s directory, so no system-wide or user-site install is needed.
• Run Blender as your normal user — Install the addon and click Install Dependencies from Edit → Preferences → Add-ons while running Blender as your everyday user (not as Administrator/root).

See the README for the full installation instructions.

Frame skips (Load Every Nth Frame)

• Correct frame-skip behavior — The Load Every Nth Frame option now implements proper frame skipping: Blender frame F displays SQLite frame F×N (e.g. with step 10, Blender frame 1 shows SQLite frame 10, frame 2 shows frame 20). The timeline is mapped so playback matches the intended simulation frames.
• Applies to paths — When Load Full Paths is enabled, path curves also respect the frame step so they stay in sync with the animation.

Contributors

• Fabian Plum
• Mohcine Chraibi — installation process improvements (no admin rights, install to addon location, pedpy path and warnings), merged via PR #3.

blenderjps_v.0.1.2

Release Notes

New

• Added Big Data Mode for very large datasets via particle-based streaming
• Added display controls for agent scale, geometry thickness, and frame rate
• Added optional Load Full Paths toggle for per-agent path curves
• Visualisation improvements (add ground plane, basic materials for agents and geometry)

blenderjps_v0.1.1

BlenderJPS v0.1.1 - Agent Path Visualization

What's New

• Agent Path Curves: Each agent's complete trajectory is now automatically created as a curve object when loading simulations
• Path Visibility Toggle: New checkbox in the Display Options section to show/hide all agent path curves
• Path curves are hidden by default but can be toggled on/off at any time

blenderjps_v0.1.0

BlenderJPS v0.1.0 - First Release

🎉 Initial release of BlenderJPS, a Blender addon for importing and visualizing JuPedSim pedestrian simulation data.

What's New

• Import JuPedSim SQLite files - Load trajectory data and walkable area geometry directly into Blender
• Animated agent visualization - Each agent is represented as an animated sphere following their trajectory path
• Geometry visualization - Walkable area boundaries and obstacles displayed as curve objects
• Built-in dependency installer - Easy setup with automatic Python package installation
• Temporal downsampling - Optional frame skipping for faster loading of large simulations

Requirements

• Blender 4.0+
• Python packages: pedpy, numpy<2.0 (installed automatically)

Installation

Install via ZIP or create a symbolic link to the blender_jps folder. See the README for detailed instructions.