rewrite parent registry & repo init

python/lsst/daf/persistence/butler.py

@@ -532,56 +532,24 @@ def __init__(self, root=None, mapper=None, inputs=None, outputs=None, **mapperAr
 
         self._setRepoDataTags()
 
-        for repoData in reversed(repoDataList):
-            self._setParentRegistry(repoData)
-            repoData.repo = Repository(repoData)
-
-    def _setParentRegistry(self, repoData):
-        """Try to get a parent registry that can be used by this repository. To be usable the repository must
-        "match", meaning the mapper in the passed-in repo is the same type as the mapper in the parent.
-        """
-
-        def getParentRegsitry(repoData, context):
-            """Get the first found registry that matches the the passed-in repo.
-
-            Parameters
-            ----------
-            repoData : RepoData
-                The RepoData for the repository for which we are searching for a
-                parent registry.
-
-            Returns
-            -------
-            Registry or None
-                A registry from a parent if one can be found, or None.
-
-            Raises
-            ------
-            RuntimeError
-                Indicates a butler init order problem, all parents should be initialized before child
-                repositories, so this function should be able to get any parent of any child repo.
-            """
-            if id(self) in context:
-                return None
-            else:
-                context.add(id(self))
-            for parentRepoData in repoData.getParentRepoDatas():
-                if parentRepoData.cfg.mapper == repoData.cfg.mapper:
-                    if parentRepoData.repo is None:
-                        self.log.debug(
-                            "_getParentRegistry: Parent {} of new repo {} not yet created, ignoring.".format(
-                                parentRepoData, repoData))
-                    else:
-                        parentRegistry = parentRepoData.repo.getRegistry()
-                        if parentRegistry:
-                            return parentRegistry
-                        else:
-                            parentRegistry = getParentRegsitry(parentRepoData, context)
-                            if parentRegistry:
-                                return parentRegistry
-            return None
+        for repoData in repoDataList:
+            self._initRepo(repoData)
 
-        repoData.repoData.parentRegistry = getParentRegsitry(repoData.repoData, set())
+    def _initRepo(self, repoData):
+        if repoData.repo is not None:
+            # this repository may have already been initialized by its children, in which case there is
+            # nothing more to do.
+            return
+        for parentRepoData in repoData.parentRepoDatas:
+            if parentRepoData.cfg.mapper != repoData.cfg.mapper:
+                continue
+            if parentRepoData.repo is None:
+                self._initRepo(parentRepoData)
+            parentRegistry = parentRepoData.repo.getRegistry()
+            repoData.parentRegistry = parentRegistry if parentRegistry else parentRepoData.parentRegistry
+            if repoData.parentRegistry:
+                break
+        repoData.repo = Repository(repoData)
 
     def _processInputArguments(self, root=None, mapper=None, inputs=None, outputs=None, **mapperArgs):
         """Process, verify, and standardize the input arguments.

python/lsst/daf/persistence/registries.py

@@ -418,6 +418,7 @@ def __init__(self, location):
         if os.path.exists(location):
             conn = sqlite3.connect(location)
             conn.text_factory = str
+            self.root = location
         else:
             conn = None
         SqlRegistry.__init__(self, conn)
@@ -441,6 +442,7 @@ def __init__(self, location):
         self._config = config
         conn = pgsql.connect(host=config["host"], port=config["port"], database=config["database"],
                              user=config["user"], password=config["password"])
+        self.root = location
         SqlRegistry.__init__(self, conn)
 
     @staticmethod

tests/test_repository.py

@@ -330,6 +330,46 @@ def testCreateAggregateAndLoadingAChild(self):
         self.assertEqual(objB, butlerD.get('foo', {'val': 2}))
 
 
+class TestDiamondPattern(unittest.TestCase):
+    """A test case for when a butler is created with an output repository and two input repositories that have
+    a common parent; verifies that the parent's registry is loaded as the parentRegistry of the output
+    repository.
+    """
+    @staticmethod
+    def makeRegistry(location, name):
+        conn = sqlite3.connect(location)
+        conn.execute("CREATE TABLE {name} (real)".format(name=name))
+        conn.commit()
+        conn.close()
+
+    def setUp(self):
+        self.testDir = tempfile.mkdtemp(dir=ROOT, prefix="TestDiamondPattern-")
+
+    def tearDown(self):
+        if os.path.exists(self.testDir):
+            shutil.rmtree(self.testDir)
+
+    def test(self):
+        repoARoot = os.path.join(self.testDir, 'repoA')
+        butler = dp.Butler(outputs={'root': repoARoot, 'mapper': ParentRepoTestMapper})
+        self.makeRegistry(os.path.join(repoARoot, 'registry.sqlite3'), 'repoA')
+        del butler
+
+        repoBRoot = os.path.join(self.testDir, 'repoB')
+        butlerB = dp.Butler(inputs=repoARoot, outputs=repoBRoot)
+        del butlerB
+
+        repoCRoot = os.path.join(self.testDir, 'repoC')
+        butlerC = dp.Butler(inputs=repoARoot, outputs=repoCRoot)
+        del butlerC
+
+        repoDRoot = os.path.join(self.testDir, 'repoD')
+        butlerD = dp.Butler(inputs=(repoBRoot, repoCRoot), outputs=repoDRoot)
+
+        self.assertEqual(1, len(butlerD._repos.outputs()))
+        self.assertEqual(os.path.dirname(butlerD._repos.outputs()[0].parentRegistry.root), repoARoot)
+
+
 class TestMasking(unittest.TestCase):
     """A test case for the repository classes.