B6: main's CI green

.gitattributes

@@ -0,0 +1,7 @@
+# Tracked files are checked out as their committed bytes on every platform.
+# Benchmark corpora, reference-export sidecars and documentation audits are
+# compared by sha256, which Windows end-of-line conversion would change.
+* -text
+# Windows command scripts keep the platform's line endings.
+*.cmd text
+*.bat text

.github/workflows/integration-tests.yml

@@ -152,6 +152,12 @@ jobs:
     runs-on: ubuntu-latest
     timeout-minutes: 45
     if: github.event_name == 'push' || github.event_name == 'pull_request' || github.event_name == 'workflow_dispatch'
+    # 11.6k tests no longer fit the 45-minute budget in one job; each group
+    # runs its share serially in one process.
+    strategy:
+      fail-fast: false
+      matrix:
+        group: [1, 2, 3]
     env:
       QT_QPA_PLATFORM: xcb
       QT_XCB_GL_INTEGRATION: xcb_glx
@@ -201,7 +207,8 @@ jobs:
           python -m scripts.install_ci_candidate \
             --dependency-source submodules \
             --wheel-directory "$RUNNER_TEMP/openhcs-candidate-wheels" \
-            --extras dev,dev-gui,gui,cellprofiler-compat,fiji,mcp-http,napari
+            --extras dev,dev-gui,gui,cellprofiler-compat,fiji,mcp-http,napari \
+            --requirement pytest-split==0.11.0
 
       - name: Materialize tracked Official30 unit fixture
         run: |
@@ -238,13 +245,14 @@ jobs:
           xvfb-run --auto-servernum --server-args="-screen 0 1920x1080x24 +extension GLX"
           python scripts/run_installed_tests.py --coverage --repo-root "$GITHUB_WORKSPACE" --
           tests/unit tests/core -p no:randomly -vv --tb=short
+          --splits 3 --group ${{ matrix.group }}
           -o faulthandler_timeout=120
           --cov-report=
 
       - name: Upload unit coverage artifact
         uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
         with:
-          name: coverage-unit
+          name: coverage-unit-${{ matrix.group }}
           path: .coverage
           include-hidden-files: true
           if-no-files-found: error
@@ -2443,6 +2451,30 @@ jobs:
             --extras "dev,dev-gui,cellprofiler-compat,$desktop_extras" \
             --requirement pytest-split==0.11.0
 
+      # The test session forbids ImageJ downloads (tests/conftest.py); the
+      # Bio-Formats source tests read the declared Fiji bundle from this cache.
+      - name: Cache managed Java and ImageJ artifacts
+        uses: actions/cache@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
+        with:
+          path: |
+            ~/.cache/polystore/imagej
+            ~/.cache/cjdk
+            ~/.jgo
+            ~/.m2/repository
+          key: ${{ runner.os }}-openhcs-fiji-${{ hashFiles('pyproject.toml', 'external/PolyStore/src/polystore/imagej_distribution.py') }}
+
+      - name: Prewarm managed Fiji runtime
+        run: |
+          source test_wheel/bin/activate
+          python - <<'PY'
+          import imagej
+          import scyjava as sj
+          from polystore.imagej_runtime import FIJI_IMAGEJ_RUNTIME
+
+          ij = FIJI_IMAGEJ_RUNTIME.initialize(imagej, sj, mode="headless")
+          FIJI_IMAGEJ_RUNTIME.shutdown(ij, sj)
+          PY
+
       - name: Smoke installed measured MCP protocol
         run: |
           source test_wheel/bin/activate

.github/workflows/website-pages.yml

@@ -4,49 +4,52 @@ on:
   pull_request:
     paths:
       - "website/**"
-      - "openhcs/resources/assets/openhcs-*.svg"
+      - "openhcs/**"
+      - "external/**"
+      - "scripts/install_ci_candidate.py"
+      - "scripts/validate_local_release_floors.py"
       - "docs/source/_static/ui.png"
       - "scripts/build_website.py"
       - "scripts/capture_media_gallery.py"
       - "scripts/gallery_catalog.py"
       - "scripts/website_gallery_projection.py"
-      - "scripts/validate_local_release_floors.py"
       - "docs/media_gallery_capture.md"
       - "tests/unit/test_build_website.py"
       - "tests/unit/test_capture_media_gallery.py"
-      - "openhcs/__init__.py"
       - "pyproject.toml"
       - ".github/workflows/website-pages.yml"
   push:
     branches: [main]
     paths:
       - "website/**"
-      - "openhcs/resources/assets/openhcs-*.svg"
+      - "openhcs/**"
+      - "external/**"
+      - "scripts/install_ci_candidate.py"
+      - "scripts/validate_local_release_floors.py"
       - "docs/source/_static/ui.png"
       - "scripts/build_website.py"
       - "scripts/capture_media_gallery.py"
       - "scripts/gallery_catalog.py"
       - "scripts/website_gallery_projection.py"
-      - "scripts/validate_local_release_floors.py"
       - "docs/media_gallery_capture.md"
       - "tests/unit/test_build_website.py"
       - "tests/unit/test_capture_media_gallery.py"
-      - "openhcs/__init__.py"
       - "pyproject.toml"
       - ".github/workflows/website-pages.yml"
   workflow_dispatch:
 
 permissions:
   contents: read
 
+# One group per ref: a push to main must not cancel a pull request's validation.
 concurrency:
-  group: pages
+  group: pages-${{ github.ref }}
   cancel-in-progress: true
 
 jobs:
   validate:
     runs-on: ubuntu-latest
-    timeout-minutes: 10
+    timeout-minutes: 30
     steps:
       - name: Check out repository
         uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
@@ -58,57 +61,13 @@ jobs:
         with:
           python-version: "3.12"
 
-      - name: Install website test dependencies
+      - name: Install the OpenHCS candidate and the repository's test extra
         run: |
-          python -m pip install pytest
-          mapfile -t local_project_paths < <(
-            python - <<'PY'
-          from pathlib import Path
-
-          from packaging.utils import canonicalize_name
-
-          from scripts.validate_local_release_floors import (
-              discover_local_projects,
-              read_project,
-          )
-
-          project = read_project(Path("pyproject.toml"))
-          required_names = {
-              canonicalize_name(name)
-              for name in (
-                  "metaclass-registry",
-                  "polystore",
-                  "python-introspect",
-                  "zmqruntime",
-              )
-          }
-          declared_names = {
-              canonicalize_name(requirement.name)
-              for requirement in project.dependencies
-          }
-          undeclared_names = required_names - declared_names
-          if undeclared_names:
-              raise SystemExit(
-                  "Missing website dependencies in project metadata: "
-                  + ", ".join(sorted(undeclared_names))
-              )
-          local_projects = {
-              candidate.canonical_name: candidate
-              for candidate in discover_local_projects()
-          }
-          missing_sources = required_names - local_projects.keys()
-          if missing_sources:
-              raise SystemExit(
-                  "Missing website dependency submodules: "
-                  + ", ".join(sorted(missing_sources))
-              )
-          print(
-              *(local_projects[name].path.parent for name in sorted(required_names)),
-              sep="\n",
-          )
-          PY
-          )
-          python -m pip install "${local_project_paths[@]}"
+          python -m pip install --upgrade pip build
+          python -m scripts.install_ci_candidate \
+            --dependency-source submodules \
+            --wheel-directory "$RUNNER_TEMP/openhcs-candidate-wheels" \
+            --extras dev
 
       - name: Test, build, and validate landing page
         run: |

README.md

@@ -98,7 +98,7 @@ graph LR
 <td width="50%" valign="top">
 
 ### 🛡️ Compile-Time Validation
-Configuration is resolved once into step snapshots and a compilation session. Typed plans then validate sources, artifacts, materialization, memory contracts, and worker requirements before execution begins. Errors surface immediately, not after hours of processing.
+Configuration is resolved once into a resolved pipeline definition and per-context compilation sessions. Typed plans then validate sources, artifacts, materialization, memory contracts, and worker requirements before execution begins. Errors surface immediately, not after hours of processing.
 
 </td>
 <td width="50%" valign="top">
@@ -199,7 +199,7 @@ graph TD
 | [**PyQT-reactive**](https://github.com/OpenHCSDev/PyQT-reactive) | UI form generation | React-style reactive forms from dataclasses with cross-window sync and flash animations |
 | [**pycodify**](https://github.com/OpenHCSDev/pycodify) | Code ↔ object conversion | Python source as serialization format — type-preserving, diffable, editable, with collision handling |
 | [**python-introspect**](https://github.com/OpenHCSDev/python-introspect) | Signature analysis | Pure-Python function/dataclass introspection for automatic UI generation and contract analysis |
-| [**metaclass-registry**](https://github.com/OpenHCSDev/metaclass-registry) | Plugin discovery | Zero-boilerplate registry system powering microscope handler and storage backend auto-discovery |
+| [**metaclass-registry**](https://github.com/OpenHCSDev/metaclass-registry) | Plugin discovery | Zero-boilerplate registry system powering dataset source and storage backend auto-discovery |
 
 ---
 
@@ -219,7 +219,7 @@ graph TD
 | OMERO | Remote image and metadata access |
 | OpenHCS format | Native generated and materialized plates |
 
-Source handlers are auto-detected and extensible through `metaclass-registry`.
+Dataset sources are auto-detected and extensible through `metaclass-registry`.
 
 </td>
 <td width="60%" valign="top">
@@ -309,24 +309,27 @@ python -m pip install "openhcs[gui,mcp,viz]"
 
 ```python
 # Or lower a CellProfiler pipeline into public OpenHCS declarations
+import queue
 from pathlib import Path
 
 from objectstate import ensure_global_config_context
 from openhcs.core.config import GlobalPipelineConfig
 from openhcs.core.orchestrator.orchestrator import PipelineOrchestrator
+from openhcs.core.progress import set_progress_queue
 from openhcs.interop.cellprofiler.pipeline_import import import_cellprofiler_pipeline
 
-plate_path = Path("/data/plate").resolve()
+dataset_root = Path("/data/plate").resolve()
 ensure_global_config_context(GlobalPipelineConfig, GlobalPipelineConfig())
 steps, pipeline_config = import_cellprofiler_pipeline(
     "analysis.cppipe",
-    source_root=plate_path,
+    source_root=dataset_root,
 )
 
 orchestrator = PipelineOrchestrator(
-    plate_path,
+    dataset_root,
     pipeline_config=pipeline_config,
 ).initialize()
+set_progress_queue(queue.Queue())  # compilation emits progress events
 execution_bundle = orchestrator.compile_pipelines(steps)
 ```
 
@@ -360,16 +363,18 @@ digests. Archived receipts without the latter two digests remain readable but
 cannot verify those files. These routes do not create a
 second pipeline runner or job-status system.
 `openhcs-benchmark run-measured --plate PLATE --pipeline-source-file PIPELINE.py
---output-dir EVIDENCE --run-id ID --wait-timeout-ms 120000` submits that
-pipeline through ordinary headless job control, then calls the same receipt
-finalizer as MCP. Use `--execution-plate PREPARED_PLATE` when the input workspace
-has been prepared separately; the evidence directory must be empty.
+--output-dir EVIDENCE --run-id ID --wait-timeout-ms 120000` runs that
+pipeline through the OpenHCS session, then calls the same receipt finalizer as
+MCP. The source plate is never written: the run mirrors it into an execution
+workspace under `EVIDENCE/workspace` (or `--workspace-root`), or uses
+`--execution-plate PREPARED_PLATE` when the input workspace has been prepared
+separately; the evidence directory must be empty.
 Use `--observation-scope outcomes` to retain per-axis success and provenance
 without transferring runtime arrays; this does not compare output values.
-After an ordinary headless job completes, the expert
-`openhcs_finalize_measured_pipeline_run` tool can validate its runtime
-observation and retain that receipt from the exact submitted pipeline and
-server completion record. It does not submit another job.
+After the OpenHCS session finishes an ordinary run that exported a runtime
+observation, the expert `openhcs_finalize_measured_pipeline_run` tool can
+validate that observation and retain the receipt; the submission and server
+record come from the session. It does not submit another run.
 
 <details>
 <summary><b>📦 All installation options</b></summary>
@@ -432,7 +437,7 @@ Supported on Python 3.11 and 3.12. See [Glencoe Software](https://www.glencoesof
 ```
 PipelineConfig + list[FunctionStep]
         ↓ resolve once
-StepSnapshot + CompilationSession
+ResolvedPipelineDefinition + CompilationSession
         ↓ derive and validate
 typed CompiledStepPlan objects
         ↓ package

benchmark/adapters/openhcs.py

@@ -18,6 +18,8 @@
 from pathlib import Path
 from typing import Any
 
+from objectstate import replace_raw
+
 from benchmark.adapters.cppipe_source import (
     CPPipeSourceRequest,
     CPPipeSourceResolution,
@@ -308,7 +310,6 @@ def _execution_global_config(
         request: OpenHCSRunRequest,
         *,
         output_suffix: str,
-        output_dataset_root: Path,
         compilation_debug_config: CompilationDebugConfig,
     ) -> GlobalPipelineConfig:
         """Select benchmark output paths without replacing declared path policy."""
@@ -330,7 +331,6 @@ def _execution_global_config(
             ),
             compilation_debug_config=compilation_debug_config,
             materialize_runtime_artifacts=request.materialize_runtime_artifacts,
-            materialization_results_path=output_dataset_root / "results",
         )
 
     def _run_converted_cppipe_pipeline(
@@ -361,9 +361,6 @@ def _run_converted_cppipe_pipeline(
             ) from exc
 
         output_suffix = f"_{request.pipeline_name}_converted_cppipe"
-        output_dataset_root = (
-            request.output_dir / f"{request.dataset_path.name}{output_suffix}"
-        )
         generated_source_path = request.output_dir / f"{cppipe_path.stem}_openhcs.py"
         source_workspace_path = (
             request.output_dir
@@ -417,7 +414,9 @@ def _run_converted_cppipe_pipeline(
                 "UI/ZMQ FunctionStep source before execution."
             )
         if compilation_debug_config.enabled:
-            pipeline_config = replace(
+            # replace_raw keeps the imported config's unset fields inheriting
+            # from the global config; dataclasses.replace would freeze them.
+            pipeline_config = replace_raw(
                 pipeline_config,
                 compilation_debug_config=LazyCompilationDebugConfig(
                     enabled=compilation_debug_config.enabled,
@@ -430,7 +429,6 @@ def _run_converted_cppipe_pipeline(
         global_config = self._execution_global_config(
             request,
             output_suffix=output_suffix,
-            output_dataset_root=output_dataset_root,
             compilation_debug_config=compilation_debug_config,
         )
         ensure_global_config_context(GlobalPipelineConfig, global_config)
@@ -550,7 +548,6 @@ def _run_converted_cppipe_pipeline(
                     raise ToolExecutionError(equivalence_failure_message)
 
         metric_results = self._metric_results(request.metrics)
-        output_dataset_root.mkdir(parents=True, exist_ok=True)
         execution_output_root.mkdir(parents=True, exist_ok=True)
 
         provenance = {