Merge pull request #19 from VH-Lab/claude/add-symmetry-tests-vtgOJ

Add cross-language symmetry tests for MATLAB ↔ Python DID databases

Author: stevevanhooser

.github/workflows/symmetry.yml

@@ -0,0 +1,95 @@
+name: Cross-Language Symmetry Tests
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+  workflow_dispatch:
+
+jobs:
+  symmetry:
+    name: MATLAB <-> Python symmetry tests
+    runs-on: ubuntu-latest
+    steps:
+      - name: Check out DID-python
+        uses: actions/checkout@v4
+
+      - name: Check out DID-matlab
+        uses: actions/checkout@v4
+        with:
+          repository: VH-Lab/DID-matlab
+          path: DID-matlab
+
+      # --- MATLAB setup ---
+      - name: Set up MATLAB
+        uses: matlab-actions/setup-matlab@v2
+        with:
+          release: latest
+          cache: true
+          products: Statistics_and_Machine_Learning_Toolbox
+
+      - name: Install MatBox
+        uses: ehennestad/matbox-actions/install-matbox@v1
+
+      - name: Install MATLAB dependencies (mksqlite etc.)
+        uses: matlab-actions/run-command@v2
+        with:
+          command: |
+            addpath(genpath("DID-matlab/src"));
+            addpath(genpath("DID-matlab/tools"));
+            matbox.installRequirements(fullfile(pwd, "DID-matlab"));
+
+      # --- Python setup ---
+      - name: Set up Python 3.12
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Install Python dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install -e ".[dev]"
+
+      # --- Step 1: MATLAB makeArtifacts ---
+      - name: "Step 1: MATLAB makeArtifact tests"
+        uses: matlab-actions/run-command@v2
+        with:
+          command: |
+            addpath(genpath("DID-matlab/src"));
+            addpath(genpath("DID-matlab/tests"));
+            addpath(genpath("DID-matlab/tools"));
+            import matlab.unittest.TestRunner;
+            import matlab.unittest.TestSuite;
+            runner = TestRunner.withTextOutput;
+            makeSuite = TestSuite.fromPackage("did.symmetry.makeArtifacts", "IncludingSubpackages", true);
+            makeResults = runner.run(makeSuite);
+            disp(table(makeResults));
+            assert(all([makeResults.Passed]), "makeArtifacts tests failed");
+
+      # --- Step 2: Python makeArtifacts + readArtifacts ---
+      - name: "Step 2: Python makeArtifact tests"
+        run: pytest -m make_artifacts -v
+
+      - name: "Step 2: Python readArtifact tests"
+        run: pytest -m read_artifacts -v
+
+      # --- Step 3: MATLAB readArtifacts ---
+      - name: "Step 3: MATLAB readArtifact tests"
+        uses: matlab-actions/run-command@v2
+        with:
+          command: |
+            addpath(genpath("DID-matlab/src"));
+            addpath(genpath("DID-matlab/tests"));
+            addpath(genpath("DID-matlab/tools"));
+            import matlab.unittest.TestRunner;
+            import matlab.unittest.TestSuite;
+            runner = TestRunner.withTextOutput;
+            readSuite = TestSuite.fromPackage("did.symmetry.readArtifacts", "IncludingSubpackages", true);
+            readResults = runner.run(readSuite);
+            disp(table(readResults));
+            nFailed = sum([readResults.Failed]);
+            nPassed = sum([readResults.Passed]);
+            nSkipped = sum([readResults.Incomplete]);
+            fprintf("Results: %d passed, %d failed, %d skipped\n", nPassed, nFailed, nSkipped);
+            assert(nFailed == 0, "readArtifacts tests failed");

README.md

@@ -43,17 +43,55 @@ The `did` library provides a framework for managing and querying data that is or
 
 You can run the tests using either `pytest` (if you installed the development dependencies) or the standard `unittest` module.
 
-**Using pytest (Recommended for development):**
+**Run all tests (unit + symmetry):**
 ```bash
 pytest
 ```
 
-**Using unittest (Standard):**
+**Run only the unit tests (excluding symmetry tests):**
+```bash
+pytest tests/ --ignore=tests/symmetry
+```
+
+**Run only the symmetry tests:**
+```bash
+pytest -m symmetry
+```
+
+**Run only the makeArtifact symmetry tests** (generate cross-language artifacts):
+```bash
+pytest -m make_artifacts
+```
+
+**Run only the readArtifact symmetry tests** (validate artifacts from Python and/or MATLAB):
+```bash
+pytest -m read_artifacts
+```
+
+**Using unittest (unit tests only):**
 ```bash
 python -m unittest discover tests
 ```
 
-Both commands will discover and run all the tests in the `tests` directory.
+#### Symmetry Tests
+
+The `tests/symmetry/` directory contains cross-language symmetry tests that verify
+DID databases created in Python can be read by MATLAB and vice versa:
+
+*   **`make_artifacts/`** — Creates a DID database with multiple branches and
+    documents, then writes the database file and JSON summary artifacts to a
+    well-known temporary directory
+    (`<tempdir>/DID/symmetryTest/pythonArtifacts/`).
+*   **`read_artifacts/`** — Reads artifacts produced by either the Python or
+    MATLAB test suite, re-summarizes the live database, and compares the
+    result against the saved summary. Tests are parameterized over
+    `matlabArtifacts` and `pythonArtifacts` and skip gracefully when
+    artifacts from a given source are not available.
+
+The CI workflow runs the full cross-language cycle:
+1. MATLAB `makeArtifact` tests create artifacts
+2. Python `makeArtifact` and `readArtifact` tests run (reading MATLAB artifacts)
+3. MATLAB `readArtifact` tests run (reading Python artifacts)
 
 ## Documentation
 

pyproject.toml

@@ -27,6 +27,14 @@ dev = [
     "pytest",
 ]
 
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+markers = [
+    "symmetry: cross-language symmetry tests (MATLAB <-> Python)",
+    "make_artifacts: tests that generate artifacts for symmetry testing",
+    "read_artifacts: tests that read and validate artifacts from another implementation",
+]
+
 [tool.setuptools.packages.find]
 where = ["src"]
 

src/did/datastructures.py

@@ -267,8 +267,7 @@ def field_search(a, search_struct):
                         b = True
                         break
     elif op_lower == "or":
-        if isinstance(param1, dict) and isinstance(param2, dict):
-            b = field_search(a, param1) or field_search(a, param2)
+        b = field_search(a, param1) or field_search(a, param2)
     elif op_lower == "depends_on":
         # param1 = dependency name, param2 = dependency value
         if "depends_on" in a:

src/did/document.py

@@ -16,10 +16,18 @@ def __init__(self, document_type="base", **kwargs):
             self.document_properties["base"]["datestamp"] = str(datetime.utcnow())
 
             for key, value in kwargs.items():
-                # This is a simplified way to set properties. A full implementation
-                # would need to handle nested properties like 'base.name'.
-                if key in self.document_properties:
-                    self.document_properties[key] = value
+                path = key.split(".")
+                if len(path) == 1:
+                    if key in self.document_properties:
+                        self.document_properties[key] = value
+                else:
+                    d = self.document_properties
+                    for p in path[:-1]:
+                        existing = d.get(p)
+                        if not isinstance(existing, dict):
+                            d[p] = {}
+                        d = d[p]
+                    d[path[-1]] = value
 
             self._reset_file_info()
 
@@ -95,6 +103,8 @@ def read_blank_definition(json_file_location_string):
                 # Ensure the 'base' key exists
                 if "base" not in data:
                     data["base"] = {}
+                # Convert flat classname/superclasses to document_class format
+                data = Document._normalize_to_document_class(data)
                 return data
 
         # Fallback for base
@@ -113,6 +123,21 @@ def read_blank_definition(json_file_location_string):
             f"Could not find definition for {json_file_location_string}"
         )
 
+    @staticmethod
+    def _normalize_to_document_class(data):
+        """Convert flat schema format to MATLAB-compatible document_class format."""
+        if "document_class" in data:
+            return data
+        class_name = data.pop("classname", "")
+        superclasses = data.pop("superclasses", [])
+        data["document_class"] = {
+            "class_name": class_name,
+            "property_list_name": class_name,
+            "class_version": 1,
+            "superclasses": superclasses,
+        }
+        return data
+
     def dependency_value(self, dependency_name, error_if_not_found=True):
         if "depends_on" in self.document_properties:
             for dep in self.document_properties["depends_on"]:

src/did/implementations/sqlitedb.py

@@ -261,11 +261,11 @@ def _normalize_loaded_props(props):
         """
         dc = props.get("document_class", {})
         sc = dc.get("superclasses")
-        if isinstance(sc, dict):
+        if sc is not None and not isinstance(sc, list):
             dc["superclasses"] = [sc]
 
         dep = props.get("depends_on")
-        if isinstance(dep, dict):
+        if dep is not None and not isinstance(dep, list):
             props["depends_on"] = [dep]
 
         return props

src/did/util/__init__.py

@@ -0,0 +1,4 @@
+from .database_summary import database_summary as database_summary
+from .compare_database_summary import (
+    compare_database_summary as compare_database_summary,
+)

src/did/util/compare_database_summary.py

@@ -0,0 +1,132 @@
+"""Compare two database summaries and return a list of discrepancy messages.
+
+Mirrors MATLAB's did.util.compareDatabaseSummary for cross-language symmetry testing.
+"""
+
+
+def compare_database_summary(summary_a, summary_b):
+    """Compare two summary dicts and return a list of mismatch descriptions.
+
+    Returns an empty list when the summaries are equivalent.
+    """
+    report = []
+
+    branches_a = summary_a.get("branchNames", [])
+    branches_b = summary_b.get("branchNames", [])
+
+    only_in_a = set(branches_a) - set(branches_b)
+    only_in_b = set(branches_b) - set(branches_a)
+
+    for name in sorted(only_in_a):
+        report.append(f'Branch "{name}" exists only in summary A.')
+    for name in sorted(only_in_b):
+        report.append(f'Branch "{name}" exists only in summary B.')
+
+    # Compare branch hierarchy
+    hier_a = summary_a.get("branchHierarchy", {})
+    hier_b = summary_b.get("branchHierarchy", {})
+    common_branches = sorted(set(branches_a) & set(branches_b))
+
+    for branch_name in common_branches:
+        if branch_name in hier_a and branch_name in hier_b:
+            parent_a = hier_a[branch_name].get("parent", "")
+            parent_b = hier_b[branch_name].get("parent", "")
+            if parent_a != parent_b:
+                report.append(
+                    f'Branch "{branch_name}": parent mismatch '
+                    f'("{parent_a}" vs "{parent_b}").'
+                )
+
+    # Compare per-branch documents
+    br_a = summary_a.get("branches", {})
+    br_b = summary_b.get("branches", {})
+
+    for branch_name in common_branches:
+        if branch_name not in br_a or branch_name not in br_b:
+            continue
+
+        branch_a = br_a[branch_name]
+        branch_b = br_b[branch_name]
+
+        if branch_a["docCount"] != branch_b["docCount"]:
+            report.append(
+                f'Branch "{branch_name}": doc count mismatch '
+                f'({branch_a["docCount"]} vs {branch_b["docCount"]}).'
+            )
+
+        map_a = {d["id"]: d for d in branch_a.get("documents", [])}
+        map_b = {d["id"]: d for d in branch_b.get("documents", [])}
+
+        missing_in_a = sorted(set(map_b) - set(map_a))
+        missing_in_b = sorted(set(map_a) - set(map_b))
+
+        for doc_id in missing_in_a:
+            report.append(
+                f'Branch "{branch_name}": doc "{doc_id}" missing in summary A.'
+            )
+        for doc_id in missing_in_b:
+            report.append(
+                f'Branch "{branch_name}": doc "{doc_id}" missing in summary B.'
+            )
+
+        for doc_id in sorted(set(map_a) & set(map_b)):
+            doc_a = map_a[doc_id]
+            doc_b = map_b[doc_id]
+
+            if doc_a.get("className", "") != doc_b.get("className", ""):
+                report.append(
+                    f'Branch "{branch_name}", doc "{doc_id}": class name mismatch '
+                    f'("{doc_a["className"]}" vs "{doc_b["className"]}").'
+                )
+
+            props_a = doc_a.get("properties", {})
+            props_b = doc_b.get("properties", {})
+
+            for field in ("demoA", "demoB", "demoC"):
+                has_a = field in props_a
+                has_b = field in props_b
+                if has_a and has_b:
+                    val_a = (
+                        props_a[field].get("value")
+                        if isinstance(props_a[field], dict)
+                        else props_a[field]
+                    )
+                    val_b = (
+                        props_b[field].get("value")
+                        if isinstance(props_b[field], dict)
+                        else props_b[field]
+                    )
+                    if val_a != val_b:
+                        report.append(
+                            f'Branch "{branch_name}", doc "{doc_id}": '
+                            f"{field}.value mismatch ({val_a} vs {val_b})."
+                        )
+                elif has_a != has_b:
+                    report.append(
+                        f'Branch "{branch_name}", doc "{doc_id}": '
+                        f'field "{field}" present in one summary but not the other.'
+                    )
+
+            # Compare depends_on
+            deps_a = props_a.get("depends_on", [])
+            deps_b = props_b.get("depends_on", [])
+            if isinstance(deps_a, dict):
+                deps_a = [deps_a]
+            if isinstance(deps_b, dict):
+                deps_b = [deps_b]
+            norm_a = [
+                (d.get("name", ""), d.get("value", ""))
+                for d in deps_a
+                if isinstance(d, dict)
+            ]
+            norm_b = [
+                (d.get("name", ""), d.get("value", ""))
+                for d in deps_b
+                if isinstance(d, dict)
+            ]
+            if norm_a != norm_b:
+                report.append(
+                    f'Branch "{branch_name}", doc "{doc_id}": depends_on mismatch.'
+                )
+
+    return report