BUG: Fix ASV harness against real ITK build + CLI reality

Changes surfaced during end-to-end smoke-test against a local ITK build
with Module_PerformanceBenchmarking=ON:

asv.conf.json
  - repo: '.' → 'itk-repo' (symlink the caller places into the harness
    pointing at the ITK source tree). Required so --set-commit-hash
    can validate SHAs against git. The consuming CI workflow and the
    local smoke-test recipe in README-asv.md both create this symlink.
  - branches: adjusted to 'origin/master' (the ITK branch whose tip we
    track for publish's time-series graph).
  - install_command: installs from relative 'python/' path rather than
    {build_dir}, since {build_dir} now refers to ITK checkouts.

python/itk_perf_shim/registry.py
  - Executable names: itkCopyIterationBenchmark → CopyIterationBenchmark
    and the same for VectorIterationBenchmark. The add_executable()
    target name drops the 'itk' prefix present on the source file.
  - CLI: Core benchmarks take 'timingsFile iterations imageSize', not
    just 'timingsFile iterations'. Fixed imageSize to 128 (matching
    the in-tree CTest registrations) and iteration counts to 25 / 50.

python/itk_perf_shim/runner.py
  - Input fixture: brainweb165a10f17.mha (the real ExternalData
    object) replaces the placeholder OAS1 filename I guessed
    pre-build-validation.

.gitignore
  - Ignore the 'itk-repo' symlink.

Smoke-test result (single-machine, same-binary, labeled as two
different ITK SHAs to exercise asv compare):

  core.track_copy_iteration         10.5ms →  10.2ms   ratio 0.98
  core.track_vector_iteration       4.98ms →  4.61ms   ratio 0.93
  filtering.track_binary_add        5.39ms →  5.09ms   ratio 0.94
  filtering.track_gradient_magnitude 48.3ms →  47.3ms  ratio 0.98
  filtering.track_median             321ms →   312ms   ratio 0.97

All within expected per-run jitter; exercise validates discovery,
shim invocation, probe-JSON parsing, and asv compare output.

Author: hjmjohnson

.gitignore

@@ -47,3 +47,4 @@ doc/_minted-ITKPerformanceBenchmarking/*
 *.egg-info/
 __pycache__/
 .asv/
+/itk-repo

asv.conf.json

@@ -2,9 +2,9 @@
     "version": 1,
     "project": "ITK",
     "project_url": "https://itk.org/",
-    "repo": ".",
-    "repo_subdir": "python",
-    "branches": ["asv-native-harness", "master"],
+    "_comment_repo": "'repo' points at the ITK source tree whose commits label the time series. The caller (CI workflow or local smoke test) must create a symlink named 'itk-repo' at the root of this harness pointing to the ITK clone before invoking asv.",
+    "repo": "itk-repo",
+    "branches": ["origin/master"],
     "environment_type": "existing",
     "show_commit_url": "https://github.com/InsightSoftwareConsortium/ITK/commit/",
     "benchmark_dir": "benchmarks",
@@ -13,7 +13,7 @@
     "html_dir": ".asv/html",
     "build_command": [],
     "install_command": [
-        "python -m pip install --no-build-isolation -e {build_dir}"
+        "python -m pip install --no-build-isolation -e python"
     ],
     "uninstall_command": [
         "return-code=any python -m pip uninstall -y itk_perf_shim"

python/itk_perf_shim/registry.py

@@ -13,14 +13,14 @@
 
 BENCHMARKS = {
     "core.copy_iteration": {
-        "exe": "itkCopyIterationBenchmark",
-        "args": ["{timings_json}", "{iterations}"],
-        "iterations": 5,
+        "exe": "CopyIterationBenchmark",
+        "args": ["{timings_json}", "{iterations}", "128"],
+        "iterations": 25,
     },
     "core.vector_iteration": {
-        "exe": "itkVectorIterationBenchmark",
-        "args": ["{timings_json}", "{iterations}"],
-        "iterations": 5,
+        "exe": "VectorIterationBenchmark",
+        "args": ["{timings_json}", "{iterations}", "128"],
+        "iterations": 50,
     },
     "filtering.median": {
         "exe": "MedianBenchmark",

python/itk_perf_shim/runner.py

@@ -81,7 +81,7 @@ def run_benchmark(name: str) -> float:
     ) as tf:
         timings_json = Path(tf.name)
 
-    brain_image = data_dir / "OAS1_0001_MR1_mpr-1_anon.nrrd"
+    brain_image = data_dir / "brainweb165a10f17.mha"
     subs = {
         "timings_json": str(timings_json),
         "iterations": str(spec["iterations"]),