Add canonical chunk plan source model

2026-07-17 23:47:22 +00:00
parent 1c13e1d64a
commit 3bfac14397
16 changed files with 755 additions and 73 deletions
--- a/internal/framework/pipeline/chunk_annotations_test.go
+++ b/internal/framework/pipeline/chunk_annotations_test.go
@@ -0,0 +1,70 @@
+package pipeline
+
+import (
+	"encoding/json"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+)
+
+func TestChunkCanonicalizationAndClonePreserveAnnotationScopes(t *testing.T) {
+	doc := validSourceDocument()
+	chunk := source.Chunk{
+		ID: "chunk-1", SourceID: doc.ID, Index: 0,
+		Ref: doc.Units[0].Ref, Content: []byte(`{"units":[1]}`), MediaType: "application/json", Units: doc.Units[:1],
+		Annotations:     source.ChunkAnnotations{"shared": json.RawMessage(` {"range": true} `)},
+		PlanAnnotations: source.ChunkAnnotations{"shared": json.RawMessage(` {"plan": true} `)},
+	}
+	canonical, err := validateAndCanonicalizeChunkResult(doc, []source.Chunk{chunk})
+	if err != nil {
+		t.Fatalf("validateAndCanonicalizeChunkResult() error = %v", err)
+	}
+	if got := string(canonical[0].Annotations["shared"]); got != `{"range":true}` {
+		t.Fatalf("range annotation = %q", got)
+	}
+	if got := string(canonical[0].PlanAnnotations["shared"]); got != `{"plan":true}` {
+		t.Fatalf("plan annotation = %q", got)
+	}
+
+	cloned := cloneSourceChunk(canonical[0])
+	cloned.Annotations["shared"][0] = '['
+	cloned.PlanAnnotations["shared"][0] = '['
+	if string(canonical[0].Annotations["shared"]) != `{"range":true}` || string(canonical[0].PlanAnnotations["shared"]) != `{"plan":true}` {
+		t.Fatal("cloneSourceChunk() shares annotation bytes")
+	}
+}
+
+func TestSerializedChunkValidationRepresentationIncludesAnnotationScopes(t *testing.T) {
+	doc := validSourceDocument()
+	chunks := []source.Chunk{{
+		ID: "chunk-1", SourceID: doc.ID, Ref: doc.Units[0].Ref, MediaType: "application/json", Units: doc.Units[:1],
+		Annotations:     source.ChunkAnnotations{"same": json.RawMessage(`{"range":1}`)},
+		PlanAnnotations: source.ChunkAnnotations{"same": json.RawMessage(`{"plan":2}`)},
+	}}
+	encoded, err := json.Marshal(chunks)
+	if err != nil {
+		t.Fatalf("json.Marshal(chunks) error = %v", err)
+	}
+	got := string(encoded)
+	if !strings.Contains(got, `"annotations":{"same":{"range":1}}`) || !strings.Contains(got, `"plan_annotations":{"same":{"plan":2}}`) {
+		t.Fatalf("serialized chunks = %s, want both annotation scopes", got)
+	}
+}
+
+func TestDebugChunkEnvelopeClonesAnnotationScopes(t *testing.T) {
+	chunk := source.Chunk{
+		ID: "chunk-1", SourceID: "source-1",
+		Annotations:     source.ChunkAnnotations{"same": json.RawMessage(`{"range":1}`)},
+		PlanAnnotations: source.ChunkAnnotations{"same": json.RawMessage(`{"plan":2}`)},
+	}
+	envelope := debugSourceChunkEnvelope(chunk)
+	if string(envelope.Annotations["same"]) != `{"range":1}` || string(envelope.PlanAnnotations["same"]) != `{"plan":2}` {
+		t.Fatalf("debug annotations = %#v / %#v", envelope.Annotations, envelope.PlanAnnotations)
+	}
+	envelope.Annotations["same"][0] = '['
+	envelope.PlanAnnotations["same"][0] = '['
+	if string(chunk.Annotations["same"]) != `{"range":1}` || string(chunk.PlanAnnotations["same"]) != `{"plan":2}` {
+		t.Fatal("debugSourceChunkEnvelope() shares annotation bytes")
+	}
+}