EdgeEnv v0.1.5

Summary

• Released InferEdgeEnv v0.1.5 as the v1-complete local-first run evidence registry and comparability checker baseline.
• Aligned README, Korean README, portfolio summary, and release follow-up wording around the same completion judgement.
• Kept SSH, remote execution, cloud services, dashboards, and leaderboard features explicitly out of scope for v1.

Validation

• git diff --check
• python -m inferedge_env.cli doctor
• edgeenv doctor
• python -m pytest -q
• scripts/smoke_release_quality_gate.sh
• GitHub Actions python-3.10, python-3.11

Impact

• Establishes the stable v1 submission baseline for local benchmark evidence recording, registry lookup, export/import, comparability reports, failed-run diagnostics, resource metrics, sampler evidence, and read-only bundle summaries.
• Makes future target expansion or remote execution work v1.1+ extension scope rather than missing MVP scope.

Non-goals

• No OS/VM/Docker/WSL/SSH/cloud/auth/dashboard/leaderboard/upload/composite ranking support added.
• No deployment decision logic added; InferEdgeLab remains the validation / decision layer.

EdgeEnv v0.1.4

Summary

• Bumped InferEdgeEnv to 0.1.4.
• Added the v0.1.4 release quality follow-up note and promoted it from README.
• Closed the six-month quality roadmap into a repeatable release quality gate covering evidence conformance, adapter templates, schema policy, demo path, error guidance, and smoke automation.

Validation

• scripts/smoke_release_quality_gate.sh
• git diff --check
• python -m inferedge_env.cli doctor
• edgeenv doctor
• GitHub Actions python-3.10 and python-3.11 passed on PR #80

Impact

• Freezes the current local-first evidence loop as the v0.1.4 baseline.
• Confirms successful-run and failed-run portability, resource query lookup, compare output, and bundle-summary handoff remain covered by the release quality smoke.
• Keeps Jetson sampled evidence optional and hardware-gated.

Non-goals

• No OS, VM, Docker, WSL, SSH, cloud, auth, dashboard, leaderboard, upload server, or composite ranking behavior.
• No schema migration for unknown future artifact versions.
• No change to the protocol-first comparability contract.

Notes

• Optional clean-room and Jetson gates were not rerun for this release branch.
• The v0.1.4 tag points to main commit 268e082.

EdgeEnv v0.1.3

Summary

• EdgeEnv v0.1.3 is a first-user polish release on top of the MVP v1 baseline.
• Bumped the package and CLI version to 0.1.3.
• Tightened README entry flow and added a v0.1.3 follow-up note.
• Includes install/quickstart resilience, README command consistency, resource query UX refinement, and release maintenance checklist work from the v0.1.3 candidate sequence.

Validation

• python -m pytest -q
• git diff --check
• python -m inferedge_env.cli doctor
• edgeenv doctor
• v0.1.3 release rehearsal local gate
• Version-bump smoke with temporary edgeenv root: fake run, runs show, resource query JSON, successful-run export/import, imported resource query JSON
• GitHub Actions readiness: python-3.10, python-3.11

Impact

• Improves first-run confidence for new users.
• Makes resource metric lookup easier to inspect and script through text and JSON output.
• Leaves a repeatable release maintenance gate for future maintainers.
• Keeps successful and failed evidence portability workflows intact.

Non-goals

• No OS, bootloader, VM, Docker, WSL, SSH, cloud, auth, dashboard, leaderboard, upload server, or composite ranking support.
• Resource metrics remain supplemental lookup evidence and do not affect comparability gates.
• Jetson validation remains local execution on the device, not remote target support.

Notes

• See docs/release-follow-up-v0.1.3.md for the short user-facing starting point.
• See docs/v0.1.3-release-rehearsal.md for the release rehearsal record.

EdgeEnv MVP v1

Summary

• EdgeEnv MVP v1 provides config-driven fake/local benchmark runs, local artifact storage, SQLite registry lookup, resource metric lookup, comparability judgement, read-only handoff summaries, and portable evidence export/import.
• Successful runs are stored under .edgeenv/runs/<run_id>/ and failed diagnostics under .edgeenv/failed-runs/<run_id>/.
• Compare reports prioritize comparability mode before optional metric deltas.
• Resource metrics remain optional secondary evidence; result.json is canonical and resource_metric_index is rebuildable local lookup state.
• Bundle summaries are human-readable handoff reports and do not replace result artifacts, manifests, sampler metadata, or raw logs.

Validation

• python -m pytest -q
• python -m inferedge_env.cli doctor
• edgeenv doctor
• README user-flow rehearsal with temporary --edgeenv-root
• runs resources list before and after successful-run import
• report bundle-summary or Jetson sampled bundle handoff smoke with --bundle-summary-output
• GitHub Actions readiness: python-3.10, python-3.11

Non-goals

• No OS, VM, Docker, WSL, SSH, cloud, auth, dashboard, leaderboard, upload server, or composite ranking support.
• No remote target execution semantics; Jetson validation runs execute locally on the Jetson.

EdgeEnv v0.1.1 - Jetson tegrastats wrapper

Summary

• Added a Jetson tegrastats wrapper command that emits optional EdgeEnv resource metrics through the existing local runner stdout contract.
• Added Jetson local benchmark/profile examples and a Jetson wrapper guide.
• Bumped the package version to 0.1.1.

Validation

• python -m pytest tests/test_jetson_tegrastats_wrapper.py tests/test_config_validation.py -q
• python -m pytest -q
• git diff --check
• python -m inferedge_env.cli doctor
• edgeenv doctor
• Jetson nano01 wrapper smoke: run_with_tegrastats.py emitted EDGEENV_RESOURCE_METRICS_JSON and EDGEENV_METRICS_JSON
• Jetson nano01 EdgeEnv smoke: edgeenv bench run stored resource_metrics from jetson-tegrastats
• GitHub Actions readiness: python-3.10, python-3.11

Non-goals

• This is not an SSH runner and does not add SSH target execution.
• LocalRunner still only executes commands and parses explicit stdout JSON lines.
• Resource metrics remain optional artifact evidence and are not a comparability gate.
• Platform sampler adapter APIs remain future work.

EdgeEnv MVP v1

Summary

• EdgeEnv MVP v1 provides config-driven fake/local benchmark runs, local artifact storage, SQLite registry lookup, comparability judgement, and portable evidence export/import.
• Successful runs are stored under .edgeenv/runs/<run_id>/ and failed diagnostics under .edgeenv/failed-runs/<run_id>/.
• Compare reports prioritize comparability mode before optional metric deltas.

Validation

• python -m pytest -q
• python -m inferedge_env.cli doctor
• edgeenv doctor
• README user-flow rehearsal with temporary --edgeenv-root
• GitHub Actions readiness on main: python-3.10, python-3.11

Non-goals

• No OS, VM, Docker, WSL, SSH, cloud, auth, dashboard, leaderboard, upload server, or composite ranking support.
• Jetson/platform-native sampler adapters remain future work.