docs: Move DepthART inference to 'common'

[iwoplaza]

I'd like more examples to be able to use this inference engine, so I think moving it to 'common' until we create a separate library is the best bet for now.

[github-actions]

Resolution Time Benchmark

---
config:
  themeVariables:
    xyChart:
      plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
  title "Random Branching (🔴 PR | 🔵 main | 🟢 release)"
  x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
  y-axis "time (ms)"
  line [0.80, 1.64, 3.69, 5.90, 7.01, 11.06, 20.02, 20.49]
  line [0.83, 1.75, 3.85, 5.80, 6.66, 10.94, 20.27, 20.67]
  line [0.90, 1.76, 3.73, 5.68, 6.37, 11.26, 21.38, 22.79]

Loading

---
config:
  themeVariables:
    xyChart:
      plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
  title "Linear Recursion (🔴 PR | 🔵 main | 🟢 release)"
  x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
  y-axis "time (ms)"
  line [0.32, 0.42, 0.60, 0.69, 0.92, 0.96, 1.19, 1.31]
  line [0.28, 0.48, 0.63, 0.79, 1.07, 1.06, 1.27, 1.43]
  line [0.29, 0.53, 0.65, 0.74, 1.01, 1.08, 1.23, 1.38]

Loading

---
config:
  themeVariables:
    xyChart:
      plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
  title "Full Tree (🔴 PR | 🔵 main | 🟢 release)"
  x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
  y-axis "time (ms)"
  line [0.85, 2.02, 3.10, 5.85, 10.69, 23.32, 49.66, 101.12]
  line [0.88, 1.87, 2.90, 6.06, 10.79, 22.96, 50.81, 100.71]
  line [0.86, 1.92, 3.12, 6.20, 11.02, 23.82, 49.00, 100.04]

Loading

[github-actions]

Bundle size comparison (import * as ... in PR vs import * as ... in target):

🟢 Decreased	➖ Unchanged	🔴 Increased	❔ Unknown
0	325	0	0

import { ... } in PR vs import * as ... in PR (is the library tree-Shakeable?):

Test	tsdown
tgpu_init.ts	263.10 kB ($${\color{green}-3.5\%}$$)
tgpu_initFromDevice.ts	262.56 kB ($${\color{green}-3.7\%}$$)
tgpu_resolve.ts	161.75 kB ($${\color{green}-40.7\%}$$)
tgpu_resolveWithContext.ts	161.68 kB ($${\color{green}-40.7\%}$$)
tgpu_bindGroupLayout.ts	62.32 kB ($${\color{green}-77.1\%}$$)
tgpu_mutableAccessor.ts	57.04 kB ($${\color{green}-79.1\%}$$)
tgpu_accessor.ts	57.04 kB ($${\color{green}-79.1\%}$$)
tgpu_privateVar.ts	55.73 kB ($${\color{green}-79.6\%}$$)
tgpu_workgroupVar.ts	55.73 kB ($${\color{green}-79.6\%}$$)
tgpu_const.ts	55.15 kB ($${\color{green}-79.8\%}$$)
tgpu_lazy.ts	54.95 kB ($${\color{green}-79.8\%}$$)
tgpu_fragmentFn.ts	39.68 kB ($${\color{green}-85.4\%}$$)
tgpu_fn.ts	39.62 kB ($${\color{green}-85.5\%}$$)
tgpu_vertexFn.ts	39.50 kB ($${\color{green}-85.5\%}$$)
tgpu_computeFn.ts	39.20 kB ($${\color{green}-85.6\%}$$)
tgpu_vertexLayout.ts	28.33 kB ($${\color{green}-89.6\%}$$)
tgpu_comptime.ts	15.93 kB ($${\color{green}-94.2\%}$$)
tgpu_unroll.ts	1.75 kB ($${\color{green}-99.4\%}$$)
tgpu_slot.ts	1.70 kB ($${\color{green}-99.4\%}$$)

If you wish to run a comparison for other, slower bundlers, run the 'Tree-shake test' from the GitHub Actions menu.

[pullfrog]

✅ No new issues found.

Reviewed changes

• Moved 33 DepthART-inference files from image-processing/monocular-light-injection/inference/ to common/depthart-inference/ — pure renames, no content change.
• Updated two import paths in monocular-light-injection/index.ts and renderer.ts to point at the relocated ../../common/depthart-inference/... modules.

I verified this is a clean, mechanical move: the old inference/ directory is fully removed, no import inside the moved directory escapes it (all internal relative imports are sibling-level, so they still resolve), and no other references to the old path remain anywhere in the docs app. The example-discovery globs are unaffected — the relocated .ts files match the same usedApis analysis glob as before, and are not picked up as gallery examples (no index.ts in the moved set), so no meta.json/index.html/thumbnail is expected. The docs astro check typechecks the two importing files, whose new import statements resolve to existing files.

  ^{｜ View workflow run ｜ Using DeepSeek Flash (free via Pullfrog for OSS) ｜ 𝕏}

[github-actions]

pkg.pr.new

No packages changed since release

[Copilot]

Pull request overview

This PR moves the DepthART inference implementation used by the monocular-light-injection docs example into the shared examples/common area, so the inference engine can be reused by additional docs examples without duplicating the implementation.

Changes:

• Updated the monocular-light-injection example to import DepthART inference utilities from src/examples/common/depthart-inference.
• Added the DepthART bundle parser, inference plan orchestration, and GPU resource management under examples/common/depthart-inference.
• Added/relocated the compute kernels and dispatch builders needed to execute DepthART on the GPU (including Winograd, depthwise, pointwise/spatial conv, scan-project, selective scan, etc.).

Reviewed changes

Copilot reviewed 2 out of 35 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
apps/typegpu-docs/src/examples/image-processing/monocular-light-injection/renderer.ts	Switches DepthART estimator/plan imports to the shared common/depthart-inference location.
apps/typegpu-docs/src/examples/image-processing/monocular-light-injection/index.ts	Switches DepthART bundle parsing and plan construction imports to common/depthart-inference.
apps/typegpu-docs/src/examples/common/depthart-inference/bundle.ts	Adds bundle format parsing (DARTBND) into a reusable helper.
apps/typegpu-docs/src/examples/common/depthart-inference/conv-dispatches.ts	Adds convolution dispatch building logic for multiple precision/layout variants.
apps/typegpu-docs/src/examples/common/depthart-inference/depthart.ts	Adds the shared DepthInferencePlan orchestration class (preprocess + dispatch execution).
apps/typegpu-docs/src/examples/common/depthart-inference/dispatch-helpers.ts	Adds shared helpers for shapes, section bindings, pipeline caching, and uniform/storage creation.
apps/typegpu-docs/src/examples/common/depthart-inference/dispatches.ts	Adds construction of the full execution dispatch list from the bundle’s recorded ops.
apps/typegpu-docs/src/examples/common/depthart-inference/disparity-range.ts	Adds a reusable disparity range estimator implementation.
apps/typegpu-docs/src/examples/common/depthart-inference/execution-plan.ts	Adds the PreparedDispatch representation and a helper to record dispatches.
apps/typegpu-docs/src/examples/common/depthart-inference/gpu-resources.ts	Adds immutable weight upload + optional lane transpose handling.
apps/typegpu-docs/src/examples/common/depthart-inference/half-weights.ts	Adds conversion utility for FP32 weights to FP16.
apps/typegpu-docs/src/examples/common/depthart-inference/preprocess.ts	Adds shared preprocessing kernel + wrapper for input frame normalization/cropping.
apps/typegpu-docs/src/examples/common/depthart-inference/tensor-arena.ts	Adds tensor arena allocation and typed buffer views for IO/slots.
apps/typegpu-docs/src/examples/common/depthart-inference/types.ts	Adds the shared model/bundle/dispatch/tensor type definitions.
apps/typegpu-docs/src/examples/common/depthart-inference/winograd-dispatches.ts	Adds the Winograd dispatcher (scratch sizing, pipeline selection, dispatch emission).
apps/typegpu-docs/src/examples/common/depthart-inference/winograd-weight.ts	Adds weight section slicing + Winograd F(4x4,3x3) weight transform logic.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/channel-views.ts	Adds kernels for channel split/concat used by the graph.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/cross-scan.ts	Adds cross-scan indexing helpers and merge kernel.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/depthwise.ts	Adds depthwise kernel family and variant generator.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/elementwise.ts	Adds binary/unary/affine kernel builders and broadcast logic.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/helpers.ts	Adds activation helpers, indexing helpers, dot-product helpers, and softplus/component helpers.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/layouts.ts	Adds bind group layouts and uniform schemas for all kernels.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/native-f16-conv.ts	Adds native-FP16 convolution kernels and IO specialization slots.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/normalization.ts	Adds specialized layer norm kernel generator.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/packed-f16-conv.ts	Adds packed-FP16 depthwise conv decoding + kernels.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/pointwise.ts	Adds pointwise (1x1) kernels and specialized kernel generator.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/pooling.ts	Adds average pooling kernel.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/resize.ts	Adds nearest/bilinear resize kernels matching ONNX modes.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/scan-project.ts	Adds specialized scan-project kernel generator.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/selective-scan.ts	Adds sequential selective scan kernel.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/spatial-conv.ts	Adds spatial (3x3) kernel + specialized generator hookup.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/spatial-specialized.ts	Adds shape-specialized 3x3 convolution generator used by multiple variants.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/types.ts	Adds shared kernel constants and planning helpers for specialized launches.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/winograd-f4.ts	Adds Winograd F(4x4,3x3) input/output transform kernels.
apps/typegpu-docs/src/examples/common/depthart-inference/kernels/winograd-gemm.ts	Adds Winograd GEMM kernels and a shape-specialized GEMM generator.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.