F merge entities

A function to merge entities, which is needed by different modules, e.g. the crawler 2.0.

There are some functions dependent on this MR, e.g. !45 (merged) . If possible, review should be done simultaneously.

There is a basic unit test.

src/caosdb/apiutils.py

@@ -557,7 +557,7 @@ def getCommitIn(folder):
         return t.readline().strip()
 
 
-COMPARED = ["name", "role", "datatype", "description", "importance",
+COMPARED = ["name", "role", "datatype", "description",
             "id", "path", "checksum", "size"]
 
 
@@ -673,10 +673,94 @@ def compare_entities(old_entity: Entity, new_entity: Entity):
     return (olddiff, newdiff)
 
 
+def copy_entity(entity: Entity):
+    """
+    Return a copy of entity.
+
+    If deep == True return a deep copy, recursively copying all sub entities.
+    """
+    print(entity)
+    if entity.role == "File":
+        new = File()
+    elif entity.role == "Property":
+        new = Property()
+    elif entity.role == "RecordType":
+        new = RecordType()
+    elif entity.role == "Record":
+        new = Record()
+    elif entity.role == "Entity":
+        new = Entity()
+    else:
+        raise RuntimeError("Unkonwn role.")
+
+    # Copy special attributes:
+    # TODO: this might rise an exception when copying
+    #       special file attributes like checksum and size.
+    for attribute in COMPARED + ["value"]:
+        val = getattr(entity, attribute)
+        if val is not None:
+            setattr(new, attribute, val)
+
+    # Copy parents:
+    for p in entity.parents:
+        new.add_parent(p)
+
+    # Copy properties:
+    for p in entity.properties:
+        new.add_property(p, importance=entity.get_importance(p))
+
+    return new
+
+
+def merge_entities(entity_a: Entity, entity_b: Entity):
+    """
+    Merge entity_b into entity_a such that they have the same parents and properties.
+    """
+
+    # Compare both entities:
+    diff_r1, diff_r2 = compare_entities(entity_a, entity_b)
+
+    # Go through the comparison and try to apply changes to entity_a:
+    for key in diff_r2["parents"]:
+        entity_a.add_parent(entity_b.get_parent(key))
+
+    for key in diff_r2["properties"]:
+        if key in diff_r1["properties"]:
+            if ("importance" in diff_r1["properties"][key] and
+                    "importance" in diff_r2["properties"][key]):
+                if (diff_r1["properties"][key]["importance"] !=
+                        diff_r2["properties"][key]["importance"]):
+                    raise NotImplementedError()
+            elif ("importance" in diff_r1["properties"][key] or
+                  "importance" in diff_r2["properties"][key]):
+                raise NotImplementedError()
+
+            for attribute in ("datatype", "unit", "value"):
+                if diff_r1["properties"][key][attribute] is None:
+                    setattr(entity_a.get_property(key), attribute,
+                            diff_r2["properties"][key][attribute])
+                else:
+                    raise RuntimeError("Merge conflict.")
+        else:
+            entity_a.add_property(
+                entity_b.get_property(key),
+                importance=entity_b.get_importance(key))
+
+    for special_attribute in ("name", "description"):
+        sa_a = getattr(entity_a, special_attribute)
+        sa_b = getattr(entity_b, special_attribute)
+        if sa_a != sa_b:
+            if sa_a is None:
+                setattr(entity_a, special_attribute, sa_b)
+            else:
+                raise RuntimeError("Merge conflict.")
+    return entity_a
+
+
 def describe_diff(olddiff, newdiff, name=None, as_update=True):
     description = ""
 
-    for attr in list(set(list(olddiff.keys())+list(newdiff.keys()))):
+    for attr in list(set(list(olddiff.keys()) + list(newdiff.keys()))):
         if attr == "parents" or attr == "properties":
             continue
         description += "{} differs:\n".format(attr)