Unify benchmark documentation navigation

Integrate the micro-benchmark guide with the CLI-first user guide and advanced framework reference, and make the intended reader path explicit.

Author: AntoineRichard

.agents/skills/isaaclab-updating-environment-docs

@@ -0,0 +1 @@
+../../skills/developer/isaaclab-updating-environment-docs

.github/actions/multi-gpu/multi_gpu_shard_runner.sh

@@ -49,7 +49,7 @@ mkdir -p /tmp/mgpu-base-home /tmp/mgpu-pyuserbase
 
 # Pytest deps (same as run-tests action). junitparser is imported at
 # tools/conftest.py load time, so it must be present first.
-./isaaclab.sh -p -m pip install pytest pytest-mock junitparser flatdict flaky "coverage>=7.6.1"
+./isaaclab.sh -p -m pip install pytest pytest-mock junitparser flaky "coverage>=7.6.1"
 
 # Shard count from nvidia-smi -L (truth; torch under-counts MIG).
 MIG_COUNT=$(nvidia-smi -L | grep -c "^  MIG ")  # grep -c = count of matching lines (MIG slices)

.github/actions/run-tests/action.yml

@@ -396,7 +396,7 @@ runs:
               # set this detect-only flag, which makes cold asset downloads
               # fall back to slow repeated retries.
               unset HUB__ARGS__DETECT_ONLY
-              ./isaaclab.sh -p -m pip install pytest pytest-mock junitparser flatdict flaky \"coverage>=7.6.1\"
+              ./isaaclab.sh -p -m pip install pytest pytest-mock junitparser flaky \"coverage>=7.6.1\"
               if [ -n \"\${TEST_WHEELHOUSE_PACKAGES:-}\" ]; then
                 if [ ! -d \"\${TEST_WHEELHOUSE_PATH:-}\" ]; then
                   echo \"Wheelhouse path is missing: \${TEST_WHEELHOUSE_PATH:-}\"

.github/workflows/docs.yaml

@@ -118,10 +118,9 @@ jobs:
         git for-each-ref --format="%(refname:short)" refs/heads/ | xargs -r git branch -D
         make multi-docs
 
-    - name: Upload docs artifact
-      uses: actions/upload-artifact@v7
+    - name: Upload GitHub Pages artifact
+      uses: actions/upload-pages-artifact@v4
       with:
-        name: docs-html
         path: ./docs/_build
 
   deploy-docs:
@@ -131,16 +130,15 @@ jobs:
     # deploy only on "deploy" branches
     if: needs.doc-build-type.outputs.trigger-deploy == 'true'
 
-    steps:
-    - name: Download docs artifact
-      uses: actions/download-artifact@v8
-      with:
-        name: docs-html
-        path: ./docs/_build
+    permissions:
+      pages: write
+      id-token: write
 
-    - name: Deploy to gh-pages
-      uses: peaceiris/actions-gh-pages@v4
-      with:
-        github_token: ${{ secrets.GITHUB_TOKEN }}
-        publish_dir: ./docs/_build
-        keep_files: false
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+
+    steps:
+    - name: Deploy GitHub Pages
+      id: deployment
+      uses: actions/deploy-pages@v4

.github/workflows/install-ci.yml

@@ -60,6 +60,7 @@ jobs:
             $'^\\.github/workflows/install-ci\\.yml$\tThis workflow file'
             $'^VERSION$\tVersion file'
             $'(^|/)pyproject\\.toml$\tPython project metadata'
+            $'^uv\\.lock$\tuv lockfile'
             $'(^|/)environment\\.ya?ml$\tConda environment file'
           )
 

docs/source/_static/tasks/classic/fourbar_pole.jpg

@@ -464,7 +464,9 @@ Isolate One Method Or Sensor
 :class:`~isaaclab.benchmark.MethodBenchmarkRunner` builds on the base class to
 repeat one method or property across input modes, warm-up steps, and instance
 counts. It is the common bridge used by the asset micro-benchmarks. Sensor
-benchmarks use live scenes and their own correctness and timing boundaries.
+benchmarks use live scenes, :func:`~isaaclab.benchmark.measure_latency` for
+paired timing boundaries, and
+:class:`~isaaclab.benchmark.LatencyBenchmarkRunner` for structured output.
 
 Do not infer end-to-end environment or training throughput from either kind of
 isolated result. The command matrix, backend prerequisites, mock-versus-live