From ce24dda992c33f9a6190d8693f0a49979db2d4e6 Mon Sep 17 00:00:00 2001
From: Nikolai Kondrashov <Nikolai.Kondrashov@redhat.com>
Date: Mon, 6 Nov 2023 14:26:49 +0200
Subject: [PATCH] Support checking for and stripping metadata

---
 kcidb_io/schema/abstract.py    | 62 +++++++++++++++++++++++++++++
 kcidb_io/schema/test_v04_03.py | 72 ++++++++++++++++++++++++++++++++++
 2 files changed, 134 insertions(+)
 create mode 100644 kcidb_io/schema/test_v04_03.py

diff --git a/kcidb_io/schema/abstract.py b/kcidb_io/schema/abstract.py
index b0efe6e..7afe1c1 100644
--- a/kcidb_io/schema/abstract.py
+++ b/kcidb_io/schema/abstract.py
@@ -365,6 +365,68 @@ def new(cls):
         assert LIGHT_ASSERTS or cls.is_valid_exactly(data)
         return data
 
+    @classmethod
+    def has_metadata(cls, data):
+        """
+        Check if a dataset has metadata.
+
+        Args:
+            data:   The dataset to check.
+
+        Returns:
+            True if the dataset has metadata fields.
+        """
+        assert cls.is_compatible_exactly(data)
+        assert LIGHT_ASSERTS or cls.is_valid_exactly(data)
+
+        def node_has_metadata(node):
+            """Check if a dataset node has metadata"""
+            if isinstance(node, dict):
+                return any(
+                    k.startswith("_") or k != "misc" and node_has_metadata(v)
+                    for k, v in node.items()
+                )
+            if isinstance(node, list):
+                return any(node_has_metadata(v) for v in node)
+            return False
+
+        return node_has_metadata(data)
+
+    @classmethod
+    def strip_metadata(cls, data, copy=True):
+        """
+        Remove metadata from a dataset, if any.
+
+        Args:
+            data:   The dataset to remove metadata from.
+            copy:   True, if the data should be copied before handling.
+                    False, if the metadata should be removed in-place.
+
+        Returns:
+            The (copy of the) dataset with metadata removed.
+        """
+        assert cls.is_compatible_exactly(data)
+        assert LIGHT_ASSERTS or cls.is_valid_exactly(data)
+
+        # Copy the data, if requested
+        if copy:
+            data = deepcopy(data)
+
+        def node_strip_metadata(node):
+            """Strip metadata from a node in a dataset"""
+            if isinstance(node, dict):
+                for k, v in list(node.items()):
+                    if k.startswith("_"):
+                        del node[k]
+                    elif k != "misc":
+                        node_strip_metadata(v)
+            elif isinstance(node, list):
+                for v in node:
+                    node_strip_metadata(v)
+
+        node_strip_metadata(data)
+        return data
+
     @staticmethod
     @abstractmethod
     def _inherit(data):
diff --git a/kcidb_io/schema/test_v04_03.py b/kcidb_io/schema/test_v04_03.py
new file mode 100644
index 0000000..70ad2e7
--- /dev/null
+++ b/kcidb_io/schema/test_v04_03.py
@@ -0,0 +1,72 @@
+"""v04_03 module tests"""
+
+from kcidb_io.schema.v04_03 import Version
+
+
+def test_has_metadata():
+    """Check has_metadata() works correctly"""
+    assert not Version.has_metadata(Version.new())
+    assert not Version.has_metadata(dict(
+        version=dict(major=Version.major, minor=Version.minor),
+        checkouts=[
+            dict(id="_:origin1:5e29d1443c46b6ca70a4c940a67e8c09f05dcb7e",
+                 origin="origin1",
+                 patchset_hash="")
+        ]
+    ))
+    assert not Version.has_metadata(dict(
+        version=dict(major=Version.major, minor=Version.minor),
+        checkouts=[
+            dict(id="_:origin1:5e29d1443c46b6ca70a4c940a67e8c09f05dcb7e",
+                 origin="origin1",
+                 patchset_hash="",
+                 misc=dict(_timestamp="2023-11-06T11:58:15.163000+00:00"))
+        ]
+    ))
+    assert Version.has_metadata(dict(
+        version=dict(major=Version.major, minor=Version.minor),
+        checkouts=[
+            dict(id="_:origin1:5e29d1443c46b6ca70a4c940a67e8c09f05dcb7e",
+                 origin="origin1",
+                 patchset_hash="",
+                 _timestamp="2023-11-06T11:58:15.163000+00:00"),
+        ]
+    ))
+
+
+def test_strip_metadata():
+    """Check strip_metadata() works correctly"""
+    assert Version.strip_metadata(Version.new()) == Version.new()
+
+    io_without_metadata = dict(
+        version=dict(major=Version.major, minor=Version.minor),
+        checkouts=[
+            dict(id="_:origin1:5e29d1443c46b6ca70a4c940a67e8c09f05dcb7e",
+                 origin="origin1",
+                 patchset_hash="")
+        ]
+    )
+    assert Version.strip_metadata(io_without_metadata) == io_without_metadata
+
+    io_with_misc_metadata = dict(
+        version=dict(major=Version.major, minor=Version.minor),
+        checkouts=[
+            dict(id="_:origin1:5e29d1443c46b6ca70a4c940a67e8c09f05dcb7e",
+                 origin="origin1",
+                 patchset_hash="",
+                 misc=dict(_timestamp="2023-11-06T11:58:15.163000+00:00"))
+        ]
+    )
+    assert Version.strip_metadata(io_with_misc_metadata) == \
+        io_with_misc_metadata
+
+    io_with_metadata = dict(
+        version=dict(major=Version.major, minor=Version.minor),
+        checkouts=[
+            dict(id="_:origin1:5e29d1443c46b6ca70a4c940a67e8c09f05dcb7e",
+                 origin="origin1",
+                 patchset_hash="",
+                 _timestamp="2023-11-06T11:58:15.163000+00:00"),
+        ]
+    )
+    assert Version.strip_metadata(io_with_metadata) == io_without_metadata