fix: enforce Phase 2 request contracts

Author: ayush-shah

platform-api/src/openmetadata_demo_api/catalog.py

@@ -82,6 +82,19 @@
     "profile",
     "status",
 )
+DEPENDENT_RECORD_TERMS = (
+    "extension",
+    "follower",
+    "lineage",
+    "profile",
+    "relationship",
+    "sample",
+    "status",
+    "tag",
+    "testresult",
+    "usage",
+    "vote",
+)
 BULK_ASSET_PREREQUISITES = CORE_SERVICE_PREREQUISITES | {
     "createAPICollection",
     "createAPIEndpoint",
@@ -172,6 +185,7 @@ def _is_bulk_drive_operation(operation: Mapping[str, Any]) -> bool:
     return (
         operation_id in BULK_ASSET_PREREQUISITES
         or source_file.startswith(("drive/", "drives/", "csv/"))
+        or "bulk" in haystack
         or (
             source_file.startswith(CORE_PREFIXES)
             and (
@@ -190,6 +204,17 @@ def _phase(operation: Mapping[str, Any]) -> str:
     source_file = operation["source"]["file"]
     if source_file.startswith("services/"):
         return "service-management"
+    operation_text = " ".join(
+        (
+            str(operation["operation_id"]),
+            str(operation["java_method"]),
+            str(operation["path"]),
+        )
+    ).lower()
+    if source_file.startswith(CORE_PREFIXES) and any(
+        term in operation_text for term in DEPENDENT_RECORD_TERMS
+    ):
+        return "enrichment"
     if source_file in CORE_PARENT_FILES:
         return "parent-assets"
     if source_file in CORE_CHILD_FILES:

platform-api/src/openmetadata_demo_api/inventory.py

@@ -39,6 +39,7 @@
 CAPITALIZED_TYPE = re.compile(r"\b(?:Create|Add|Update|Restore|Bulk|Search|Move|Set|Vote)\w+\b")
 IMPORT = re.compile(r"(?m)^import\s+(?!static\s)([\w.]+);")
 TYPE_TOKEN = re.compile(r"\b[A-Z][A-Za-z0-9_]+\b")
+JAVA_IDENTIFIER = re.compile(r"\b[A-Za-z_$][A-Za-z0-9_$]*\b")
 PARAMETER_LOCATION = re.compile(
     r"@(PathParam|QueryParam|HeaderParam|CookieParam|FormParam|FormDataParam)"
     r"\(\s*\"([^\"]+)\"\s*\)"
@@ -91,6 +92,7 @@ class ParameterContract:
     name: str
     location: str
     java_type: str
+    cardinality: str
     required: bool
     default: str | None
     model_hints: tuple[str, ...]
@@ -179,6 +181,46 @@ def _split_java_parameters(signature: str) -> list[str]:
     return chunks
 
 
+def _without_java_annotations(text: str) -> str:
+    """Remove annotations, including nested arguments, from a parameter declaration."""
+
+    output: list[str] = []
+    offset = 0
+    while offset < len(text):
+        if text[offset] != "@":
+            output.append(text[offset])
+            offset += 1
+            continue
+        offset += 1
+        while offset < len(text) and (text[offset].isalnum() or text[offset] in "_.$"):
+            offset += 1
+        while offset < len(text) and text[offset].isspace():
+            offset += 1
+        if offset < len(text) and text[offset] == "(":
+            depth = 1
+            offset += 1
+            in_string = False
+            escaped = False
+            while offset < len(text) and depth:
+                character = text[offset]
+                if in_string:
+                    if escaped:
+                        escaped = False
+                    elif character == "\\":
+                        escaped = True
+                    elif character == '"':
+                        in_string = False
+                elif character == '"':
+                    in_string = True
+                elif character == "(":
+                    depth += 1
+                elif character == ")":
+                    depth -= 1
+                offset += 1
+        output.append(" ")
+    return "".join(output)
+
+
 def _parameter_contracts(signature: str, imports: dict[str, str]) -> tuple[ParameterContract, ...]:
     contracts: list[ParameterContract] = []
     locations = {
@@ -200,14 +242,23 @@ def _parameter_contracts(signature: str, imports: dict[str, str]) -> tuple[Param
         name = annotation.group(2) if annotation else variable.group(1)
         default_match = DEFAULT_VALUE.search(parameter)
         default = default_match.group(1) if default_match else None
-        type_tokens = TYPE_TOKEN.findall(parameter[: variable.start()])
+        declaration = _without_java_annotations(parameter[: variable.start()])
+        type_tokens = [
+            token for token in JAVA_IDENTIFIER.findall(declaration) if token not in {"final"}
+        ]
         java_type = type_tokens[-1] if type_tokens else "unknown"
+        cardinality = (
+            "array"
+            if re.search(r"\b(?:Collection|List|Set)\s*<|\[\]", parameter[: variable.start()])
+            else "single"
+        )
         model_hints = tuple(sorted(_schema_type_hints(parameter, imports)))
         contracts.append(
             ParameterContract(
                 name=name,
                 location=location,
                 java_type=java_type,
+                cardinality=cardinality,
                 required=location in {"path", "body"} or "@NotNull" in parameter,
                 default=default,
                 model_hints=model_hints,
@@ -381,6 +432,7 @@ def parse_resource_source(source_file: Path, resource_root: Path) -> list[Source
                 name="inherited_collection",
                 location="path",
                 java_type="String",
+                cardinality="single",
                 required=True,
                 default="entities",
                 model_hints=(),