SOF-8032: prepare CoreJob

schema/job.json

@@ -5,16 +5,101 @@
     "type": "object",
     "allOf": [
         {
-            "$ref": "job/base.json"
+            "description": "in-memory entity",
+            "$ref": "in_memory_entity/named_defaultable_has_metadata.json"
+        },
+        {
+            "description": "Compute Property Schema",
+            "$ref": "job/compute_property.json"
+        },
+        {
+            "$ref": "system/status_track.json"
         }
     ],
     "properties": {
         "workflow": {
             "description": "Arch-workflow for the job",
             "$ref": "workflow.json"
+        },
+        "rmsId": {
+            "description": "Identity used to track jobs originated from command-line",
+            "type": "string"
+        },
+        "status": {
+            "description": "job status",
+            "enum": [
+                "pre-submission",
+                "queued",
+                "submitted",
+                "active",
+                "finished",
+                "terminate-queued",
+                "terminated",
+                "error",
+                "deleted",
+                "timeout"
+            ]
+        },
+        "startTime": {
+            "description": "Approximate start time of the job. e.g. within 10 min",
+            "type": "string"
+        },
+        "workDir": {
+            "description": "The path to the working directory of this job, when the job originates from command-line",
+            "type": "string"
+        },
+        "_project": {
+            "description": "Subset of the full information about the project that this job belongs to.",
+            "$ref": "system/entity_reference.json"
+        },
+        "_material": {
+            "description": "Subset of the full information about the materials used inside this job.",
+            "$ref": "system/entity_reference.json"
+        },
+        "_materials": {
+            "description": "Subset of the full information about the materials used inside a multi-material job.",
+            "type": "array",
+            "items": {
+                "$ref": "system/entity_reference.json"
+            }
+        },
+        "_materialsSet": {
+            "description": "Subset of the full information about the materials set used inside a multi-material job.",
+            "$ref": "system/entity_reference.json"
+        },
+        "parent": {
+            "description": "Subset of the full information about the parent job for the job. Only a job from the same project can be assigned as a parent one.",
+            "$ref": "system/entity_reference.json"
+        },
+        "runtimeContext": {
+            "description": "Context variables that the job will have access to at runtime",
+            "type": "object"
+        },
+        "scopeTrack": {
+            "description": "history of the workflow scope on each update",
+            "type": "array",
+            "items": {
+                "type": "object",
+                "properties": {
+                    "repetition": {
+                        "type": "number"
+                    },
+                    "scope": {
+                        "$ref": "workflow/scope.json"
+                    }
+                }
+            }
+        },
+        "dataset": {
+            "description": "Dataset configuration for a data-driven job",
+            "$ref": "job/dataset.json"
+        },
+        "purged": {
+            "type": "boolean"
+        },
+        "purgedAt": {
+            "type": "number"
         }
     },
-    "required": [
-        "workflow"
-    ]
+    "required": ["workflow", "status", "_project"]
 }

schema/job/dataset.json

@@ -0,0 +1,37 @@
+{
+    "$id": "job/dataset",
+    "$schema": "http://json-schema.org/draft-07/schema#",
+    "title": "job dataset schema",
+    "type": "object",
+    "properties": {
+        "datasetFilepath": {
+            "type": "string"
+        },
+        "datasetBasename": {
+            "type": "string"
+        },
+        "objectStorageContainerData": {
+            "type": "object",
+            "properties": {
+                "CONTAINER": {
+                    "type": "string"
+                },
+                "NAME": {
+                    "type": "string"
+                },
+                "PROVIDER": {
+                    "type": "string"
+                },
+                "REGION": {
+                    "type": "string"
+                },
+                "SIZE": {
+                    "type": "string"
+                },
+                "TIMESTAMP": {
+                    "type": "string"
+                }
+            }
+        }
+    }
+}

schema/workflow.json

@@ -22,7 +22,7 @@
     ],
     "properties": {
         "workflows": {
-            "description": "Array of workflows with the same schema as the current one.",
+            "description": "Array of workflows with the same schema as the current one. NOTE: items are not validated against the workflow schema itself (known limitation - see server/utils.ts's parseIncludeReferenceStatements for why, and consider fixing).",
             "type": "array",
             "items": {
                 "type": "object"

src/js/esse/server/utils.ts

@@ -30,6 +30,21 @@ export function parseIncludeReferenceStatements(filePath: string): JSONSchema {
     const originalId = parsed.$id;
     let dereferenced = deref(parsed, { baseFolder: dirPath, removeIds: true });
     // handle circular references and use non-dereferenced source
+    //
+    // KNOWN LIMITATION: this only catches a schema's own *direct* self-reference (e.g.
+    // workflow.json referencing itself). It does NOT handle a cycle reached transitively
+    // through another schema's own $ref (e.g. job.json -> $ref workflow.json -> $ref back to
+    // workflow.json) - in that case `deref` collapses the nested self-reference to `{}` (losing
+    // its $id) instead of throwing this same "Circular self reference" error, which crashes
+    // `JSONSchemasGenerator.writeResolvedSchemas` ("Schema ID is missing"). This is why
+    // `workflow.json`'s own `workflows` field is declared as `items: { type: "object" }` rather
+    // than a real `$ref` back to itself: consumers (wode's WorkflowSchema, jode's JobEntity,
+    // web-app's CoreJobSchema) hand-override the resulting lossy TS type with a properly
+    // recursive one, but that only fixes the TypeScript side - AJV validation of nested
+    // workflows' own structure is still effectively skipped. Fixing this for real means making
+    // this transitive case hit the same fallback as the direct case (probably by tracking
+    // visited $ids across the whole dereference walk, not just within a single file's own
+    // top-level deref call) - deferred for now.
     if (dereferenced instanceof Error && dereferenced.message === "Circular self reference") {
         dereferenced = parsed;
     }