Run the truth report in CI as an artifact; drop the test that could never run

.github/workflows/ci.yml

@@ -87,6 +87,19 @@ jobs:
         if: matrix.os == 'ubuntu-latest' && matrix.python-version == '3.13'
         run: git diff --exit-code -- docs/reference/validation.md
 
+      - name: Truth report (ubuntu / 3.13 only; a report, not a gate)
+        # How closely the methods recover answers known by construction. Kept as an
+        # artifact on every run so the figures behind validation.md's last section are at hand.
+        if: matrix.os == 'ubuntu-latest' && matrix.python-version == '3.13'
+        run: python validation/truth_report.py --out truth_report.md
+
+      - name: Upload the truth report
+        if: matrix.os == 'ubuntu-latest' && matrix.python-version == '3.13'
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
+        with:
+          name: truth-report
+          path: truth_report.md
+
       - name: Upload coverage to Codecov
         if: matrix.os == 'ubuntu-latest' && matrix.python-version == '3.11'
         uses: codecov/codecov-action@fb8b3582c8e4def4969c97caa2f19720cb33a72f # v7.0.0

CHANGELOG.md

@@ -73,10 +73,15 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - **`examples/` and `validation/effects_visualization.ipynb`.** These demos predated the 0.2.0
   API (`ProcessDataFrame`, `study.sds`, moved data files) and no longer ran; nothing linked to
   them. The maintained, CI-executed tutorials are in `docs/tutorials/`.
+- **`tests/test_bishop_reference.py`.** Its one test compared residual values with a workbook
+  of Dr. Bishop's that is not in the repository, so it skipped on every run. Residual charts are
+  still compared with VAS through their limits.
 
 ### CI / Infrastructure
 - **Python 3.14 is tested** on Ubuntu and Windows, in CI and in the release gate, and listed in
   the package classifiers.
+- **The truth report runs in CI.** `validation/truth_report.py` runs on every push (Ubuntu,
+  Python 3.13) and its output is kept as the `truth-report` artifact.
 
 ## [0.5.0] - 2026-10-29
 

CONTRIBUTING.md

@@ -74,13 +74,7 @@ The full test suite is the gate: if it fails, the change is not ready.
 The file `validation/PBTESTDATABASE_T100.csv` is ground-truth Minitab output
 for DS 1–6 from Bishop's reference materials. Numerical changes that touch
 residuals (R1–R5), DS detection, limit calculations, or variance
-decomposition must keep the relevant Bishop-reference tests green:
-
-```bash
-pytest tests/test_bishop_reference.py -v   # skips unless Bishop's residual workbooks are present
-```
-
-There is also a full end-to-end check that compares every chart center, control
+decomposition must keep the end-to-end check green. It compares every chart center, control
 limit, capability index and loss component against Bishop's VAS runs for the
 10-12-2026 manual (PM SDS 1–6, PM INERT, his known-effects data and the Medicare data):
 

docs/reference/validation.md

@@ -52,8 +52,7 @@ These outputs have no VAS reference in this run. They are covered by unit tests,
 other references, but no VAS value is checked:
 
 - **Residual values one by one.** Charts of R1-R6 are compared through their limits above;
-  `tests/test_bishop_reference.py` compares the values themselves when the reference
-  spreadsheets are present, which they are not in this repository, so it skips.
+  the values themselves are not compared.
 - **Signals** (run rules and points beyond the limits).
 - **Staged (phased) Xbar and S charts.** These are checked against Wheeler's cavity data and
   Minitab's per-stage limits in `tests/test_phased_xbar_s.py`.
@@ -64,7 +63,8 @@ other references, but no VAS value is checked:
 
 `validation/truth_report.py` asks a different question: how closely the methods, as
 implemented, recover answers known by construction (pure noise, and data built from known
-effects). It is a report, not a gate.
+effects). It is a report, not a gate. CI runs it on every push and keeps its output as the
+`truth-report` artifact.
 
 ## Values with no reference yet
 

validation/e2e_bishop_report.py

@@ -870,8 +870,7 @@ def generate_myst_summary(all_results, aux_by_run, runs):
         'other references, but no VAS value is checked:',
         '',
         '- **Residual values one by one.** Charts of R1-R6 are compared through their limits above;',
-        '  `tests/test_bishop_reference.py` compares the values themselves when the reference',
-        '  spreadsheets are present, which they are not in this repository, so it skips.',
+        '  the values themselves are not compared.',
         '- **Signals** (run rules and points beyond the limits).',
         "- **Staged (phased) Xbar and S charts.** These are checked against Wheeler's cavity data and",
         "  Minitab's per-stage limits in `tests/test_phased_xbar_s.py`.",
@@ -882,7 +881,8 @@ def generate_myst_summary(all_results, aux_by_run, runs):
         '',
         '`validation/truth_report.py` asks a different question: how closely the methods, as',
         'implemented, recover answers known by construction (pure noise, and data built from known',
-        'effects). It is a report, not a gate.',
+        'effects). It is a report, not a gate. CI runs it on every push and keeps its output as the',
+        '`truth-report` artifact.',
         '',
         '## Values with no reference yet',
         '',