From 628a2c84510abfbb5eff67d420d32a9e3ec6b1da Mon Sep 17 00:00:00 2001
From: Alex Goodman <wagoodman@users.noreply.github.com>
Date: Thu, 23 Mar 2023 13:51:06 -0400
Subject: [PATCH] fix workspace schema to be correct relative to actual 1.0
 datashape (#128)

Signed-off-by: Alex Goodman <alex.goodman@anchore.com>
---
 .../schema-1.0.0.json                         | 153 ++----------------
 tests/conftest.py                             |  14 ++
 tests/unit/test_workspace.py                  |  17 ++
 3 files changed, 48 insertions(+), 136 deletions(-)

diff --git a/schema/provider-workspace-state/schema-1.0.0.json b/schema/provider-workspace-state/schema-1.0.0.json
index 3eb0adcf..ef139bb9 100644
--- a/schema/provider-workspace-state/schema-1.0.0.json
+++ b/schema/provider-workspace-state/schema-1.0.0.json
@@ -10,154 +10,34 @@
     "urls": {
       "type": "array",
       "items": [
-        {
-          "type": "string"
-        },
         {
           "type": "string"
         }
       ]
     },
-    "input": {
+    "store": {
+      "type": "string"
+    },
+    "timestamp": {
+      "type": "string"
+    },
+    "listing": {
       "type": "object",
       "properties": {
-        "files": {
-          "type": "array",
-          "items": [
-            {
-              "type": "object",
-              "properties": {
-                "path": {
-                  "type": "string"
-                },
-                "digests": {
-                  "type": "array",
-                  "items": [
-                    {
-                      "type": "string"
-                    },
-                    {
-                      "type": "string"
-                    }
-                  ]
-                },
-                "modified": {
-                  "type": "string"
-                }
-              },
-              "required": [
-                "path",
-                "digests",
-                "modified"
-              ]
-            },
-            {
-              "type": "object",
-              "properties": {
-                "path": {
-                  "type": "string"
-                },
-                "digests": {
-                  "type": "array",
-                  "items": [
-                    {
-                      "type": "string"
-                    },
-                    {
-                      "type": "string"
-                    }
-                  ]
-                },
-                "modified": {
-                  "type": "string"
-                }
-              },
-              "required": [
-                "path",
-                "digests",
-                "modified"
-              ]
-            }
-          ]
+        "digest": {
+          "type": "string"
         },
-        "timestamp": {
+        "path": {
           "type": "string"
-        }
-      },
-      "required": [
-        "files",
-        "timestamp"
-      ]
-    },
-    "results": {
-      "type": "object",
-      "properties": {
-        "files": {
-          "type": "array",
-          "items": [
-            {
-              "type": "object",
-              "properties": {
-                "path": {
-                  "type": "string"
-                },
-                "digests": {
-                  "type": "array",
-                  "items": [
-                    {
-                      "type": "string"
-                    },
-                    {
-                      "type": "string"
-                    }
-                  ]
-                },
-                "modified": {
-                  "type": "string"
-                }
-              },
-              "required": [
-                "path",
-                "digests",
-                "modified"
-              ]
-            },
-            {
-              "type": "object",
-              "properties": {
-                "path": {
-                  "type": "string"
-                },
-                "digests": {
-                  "type": "array",
-                  "items": [
-                    {
-                      "type": "string"
-                    },
-                    {
-                      "type": "string"
-                    }
-                  ]
-                },
-                "modified": {
-                  "type": "string"
-                }
-              },
-              "required": [
-                "path",
-                "digests",
-                "modified"
-              ]
-            }
-          ]
         },
-        "timestamp": {
+        "algorithm": {
           "type": "string"
         }
       },
       "required": [
-        "files",
-        "timestamp"
+        "digest",
+        "path",
+        "algorithm"
       ]
     },
     "schema": {
@@ -179,8 +59,9 @@
   "required": [
     "provider",
     "urls",
-    "input",
-    "results",
+    "store",
+    "timestamp",
+    "listing",
     "schema"
   ]
 }
diff --git a/tests/conftest.py b/tests/conftest.py
index e00aa38f..1bd0761c 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -14,6 +14,10 @@ def __init__(self, root: str, name: str):
         self.root = root
         self.name = name
 
+    @property
+    def metadata_path(self):
+        return self.root / self.name / "metadata.json"
+
     @property
     def input_dir(self):
         return self.root / self.name / "input"
@@ -48,6 +52,16 @@ def result_schemas_valid(self, require_entries: bool = True) -> bool:
 
         return True
 
+    def metadata_schema_valid(self) -> bool:
+        with open(self.metadata_path) as f:
+            item = json.load(f)
+            schema_url = item["schema"]["url"]
+
+            schema_dict = load_json_schema(get_schema_repo_path(schema_url))
+            jsonschema.validate(instance=item, schema=schema_dict)
+
+        return True
+
 
 def load_json_schema(path: str) -> dict:
     with open(path) as f:
diff --git a/tests/unit/test_workspace.py b/tests/unit/test_workspace.py
index fdbf1a8d..57464ed6 100644
--- a/tests/unit/test_workspace.py
+++ b/tests/unit/test_workspace.py
@@ -127,3 +127,20 @@ def test_record_state_urls_persisted_across_runs(tmpdir, dummy_file):
     )
 
     assert current_state == expected_state
+
+
+def test_state_schema(tmpdir, dummy_file, helpers):
+    name = "dummy"
+    ws = workspace.Workspace(root=tmpdir, name=name, create=True)
+
+    # create a dummy files
+    dummy_file(ws.input_path, "dummt-input-1.json")
+    dummy_file(ws.results_path, "dummy-00000.json")
+
+    urls = ["http://localhost:8000/dummy-input-1.json"]
+    store = result.StoreStrategy.FLAT_FILE
+    ws.record_state(urls=urls, store=store.value, timestamp=datetime.datetime(2021, 1, 1))
+
+    ws_helper = helpers.provider_workspace_helper(name=name, create=False)
+
+    assert ws_helper.metadata_schema_valid()