docs: align README + MIGRATION with Isaac Phase 1 in main

[yinsong1986]

Summary

Top-level docs claimed Isaac Sim would "land in v0.3.0+" as future work, but PR #44 / #45 / #46 / #47 / #51 (R6 + R7 Phase 1) all merged to main on 2026-05-26. This PR updates README.md + examples/MIGRATION.md to reflect what's actually shipped vs what's still pending. No code touched.

Inconsistencies fixed

README.md

#	Where	Was	Now
1	Status banner (line 7)	"Backend code lands in v0.3.0+ (Isaac)" — future-tense	Explicit Phase 1 / Phase 2 split. Names what works today (entry-points, IsaacConfig, lifecycle scaffolding, procedural builders, URDF/MJCF/USD loaders) vs what's still no-op (add_object, add_camera, replicate, per-IsaacSimulation _load_*_robot stubs)
2	Quick-start preamble	"become copy-paste-runnable when R6/#13 and R7/#14 (Isaac) ship"	"Isaac Sim Phase 1 has shipped — the snippet below is copy-paste-runnable today on a host with Isaac Sim 2024.x+ installed". Newton snippets remain gated on R11/#18
3	Isaac quick-start snippet	create_simulation("isaac", rtx_mode="path_traced", ...) — rtx_mode is not an IsaacConfig field; would TypeError at construction	create_simulation("isaac", render_mode="rtx_pathtracing", ...) — matches the field actually defined at strands_robots_sim/isaac/config.py:render_mode. Plus a second snippet for the loaders module (load_urdf / load_mjcf / load_usd) since that's the working URDF / MJCF / USD ingestion path today
4	Stage 2 R5 checkbox	[ ]	[x] (#26 merged)
5	Stage 3 R6 checkbox	[ ]	[x] with PR link (#44 merged)
6	Stage 3 R7 checkbox	[ ]	Split into R7 (Phase 1) [x] linking the four merged PRs (#45 / #46 / #47 / #51) and R7 (Phase 2) [ ] for the still-no-op data-plane wiring

examples/MIGRATION.md

#	Where	Was	Now
7	[isaac] install line (~line 28)	Commented-out + "Heavy GPU-only backends ship in later 0.x releases"	Uncommented; inline note "Phase 1 shipped; data-plane in Phase 2". Newton remains commented (Stage 4 pending)
8	"After — same task on Isaac Sim" heading	"Stage 3, future"	"Stage 3, Phase 1 shipped"
9	Isaac snippet kwarg	Same rtx_mode='path_traced' bug as README	Fixed to render_mode='rtx_pathtracing'; added comment noting evaluate_benchmark on a real LIBERO scene still needs Phase 2 data-plane wiring (consistent with the disclosure banner on docs/backends/isaac.md)

Out of scope

• examples/README.md's backend-matrix TBD rows for libero/run_isaac.py / libero/run_isaac_fleet.py. The Isaac BACKEND is in main but those EXAMPLE files (R8/R8: Add examples/libero_isaac.py (Stage 3) #15, R23/R23: Add examples/libero_isaac_fleet.py — IsaacLab fleet-scale LIBERO (Stage 3) #27) haven't shipped yet — the rows are still forward-accurate, no edit needed.
• docs/backends/isaac.md itself. That file is in the in-flight PR #48 (5/5 of feat(isaac): IsaacSimulation backend skeleton + entry-point registration (R7) #31 split); commit 76c3b1f on that PR brings its Phase 1 banner + Architecture file-list + Loaders section in line with the same understanding this PR is documenting at the top level.

Verification

• git diff --stat: 2 files, +28 / -14 LOC. Pure docs.
• README + MIGRATION render cleanly under standard markdown (spot-checked table / heading nesting).
• Both kwarg fixes (render_mode='rtx_pathtracing') verified against current main's IsaacConfig field set (num_envs / device / headless / physics_dt / rendering_dt / render_mode / gravity / ground_plane / stage_path / nucleus_url / camera_width / camera_height / enable_rtx_sensors / verbose / extra).

Cross-references

• Same kind of post-merge alignment as PR #36 did for examples/README.md after R5: Add LIBERO-on-MuJoCo baseline example (Stage 2) #26's R5 work merged. This is its R7 Phase 1 sibling.
• Forward-pointing complement to in-flight PR #48 (docs/backends/isaac.md's Phase 1 banner + Loaders section + Architecture file-list — same framing, three doc surfaces in sync).

[cagataycali]

Approve — verified all factual claims against main @ 06fce4d

Local verification on a fresh clone:

Field-name fix is correct

strands_robots_sim/isaac/config.py lines 62-76 enumerate the exact IsaacConfig field set the PR description claims. render_mode exists (line 67, default "headless"); rtx_mode does not. The pre-PR snippet create_simulation("isaac", rtx_mode="path_traced", ...) would raise TypeError: __init__() got an unexpected keyword argument 'rtx_mode' at construction. rtx_pathtracing is in RENDER_MODES (verified via the __post_init__ check on line 81 + the auto-promotion path on line 116). Fix is exactly right.

Phase 1 surface enumeration matches code

• Procedural builders: _build_so100 / _build_panda / _build_unitree_g1 exist in strands_robots_sim/isaac/procedural.py (line 98 onwards). Aliases so-100 / so_100 / so101 resolve to so100 (lines 274-277). Registered keys returned by list_procedural_robots() are ["so100", "panda", "unitree_g1"] (line 311).
• Loaders: load_urdf (line 157), load_mjcf (line 367), load_usd (line 576) all defined in strands_robots_sim/isaac/loaders.py, all returning ProceduralRobot.
• Lifecycle: IsaacSimulation.create_world / destroy / cleanup are non-stub (PR #47's commit 57929b3).

Phase 2 still-no-op enumeration matches the in-flight banner on PR #48

The add_object / add_camera / replicate / per-IsaacSimulation _load_*_robot stubs / articulation-touching paths under get_observation / send_action / render list in this PR's status banner is identical to the rewritten banner on docs/backends/isaac.md (PR #48 commit 76c3b1f). All three doc surfaces (README.md / examples/MIGRATION.md / docs/backends/isaac.md) will be in sync once both PRs land.

Roadmap checkbox accuracy

• R5 / #12 → checked: closed by PR #26 (4a89396 on main).
• R6 / #13 → checked, linked to PR #44.
• R7 / #14 split into Phase 1 (checked, linked to #45 / #46 / #47 / #51) + Phase 2 (unchecked, data-plane). Preserves the historical issue reference while exposing next-up scope. Good split.
• R8 / #15 corrected from examples/libero_isaac.py to examples/libero/run_isaac.py — matches the actual examples/libero/run_*.py convention established by #26.

MIGRATION.md kwarg parity

Same rtx_mode → render_mode fix as the README. The added 2-line comment about evaluate_benchmark needing Phase 2 data-plane wiring is consistent with the disclosure banner on the in-flight docs/backends/isaac.md rewrite.

Out-of-scope acknowledgment is correct

examples/README.md's TBD rows for libero/run_isaac.py / libero/run_isaac_fleet.py are still forward-accurate (those example files haven't shipped — R8/#15 / R23/#27 are open). Leaving them as-is is the right call.

No code touched, no tests to run

2 files, +28 / -14 LOC, pure docs. Markdown table / heading nesting renders cleanly under a standard parser.

Mergeability

MERGEABLE against current main. No code dependency on PR #48 — the two PRs can land independently in either order, as the description correctly states. Approving.