Last refreshed: 2025-10-17 12:13 UTC via gh issue list --limit 20 --json number,title,state,labels,assignees.
@cappelletto- project maintainer; currently driving performance profiling, exporter multithreading, and SIMD exploration acrossvideostrip_core(#55, #56, #41, #42, #43).- Unassigned work - issue backlog is open for contributors; see tracker snapshot below for entry points.
- C++17 frame extraction engine (OpenCV capture, keyframe selection, feature pipelines) plus metadata emitters.
- Current focus: throughput work (queue-based exporter, SIMD hot paths) and profiling instrumentation.
- CLI wrapper around
VideoFrameExtractorwith args.hxx parsing, YAML config, schema printing. - Keep configs in
configs/; use--configto point at presets.
- Experimental viewer; not built by default but kept for prototyping.
- YAML presets demonstrating minimal and enhanced runs; safe starting point when debugging.
- Generated manpages and supporting docs.
- Catch2 suites (
test_core_smoke, config/enhance coverage). Enable with-DBUILD_TESTS=ON; CI mirrorstests/build_with_tests.sh.
- CMake artifacts; ignored by git. Reconfigure with
cmake -S . -B build ...as needed.
- Vendored libs (e.g. args.hxx) pinned for reproducible builds.
- Configure:
cmake -S . -B build -DCMAKE_BUILD_TYPE=Release - Build CLI:
cmake --build build --target videostrip_cli -j - Enable tests: add
-DBUILD_TESTS=ON, thencmake --build build --target test_core_smoke test_config_yaml -j - Run suites:
ctest --test-dir build --output-on-failure
| # | Title | Labels | Assignee |
|---|---|---|---|
| #74 | Automated plugin self-test on Meshroom | - | Unassigned |
| #73 | Preflight binary check tool | - | Unassigned |
| #59 | Record enhancement & normalization config in summary.yaml | documentation | Unassigned |
| #58 | Performance regression smoke test in CI | performance | Unassigned |
| #57 | SIMD optimization for enhancement hot paths | performance | Unassigned |
| #56 | Queue-based Multithreaded Frame Export | performance | Unassigned |
| #55 | Profiling Report with VTune & perf | performance | Unassigned |
| #43 | Explore SIMD intrinsics acceleration for specific target-arch | enhancement | @cappelletto |
| #42 | Implement queue based multithreading write when exporting selected frame | enhancement | @cappelletto |
| #41 | Run in-depth VTune profiling and identify/prioritize optimization strategy | enhancement | @cappelletto |
| #32 | Extend frame selection overlap modes (Optical Flow & ECC) | enhancement | @cappelletto |
- Use 4-space indentation, Allman braces for types/namespaces, inline braces for control flow.
- Format with
clang-format -i(LLVM style, 120 cols fallback). - Public headers carry brief Doxygen docs.
- Commit style:
feat(core): add grid normalization; link relevant issues (Fixes #42). Target PRs atdevelopand ensurecmake --build buildplusctestpass on Linux/Windows.