Deep-clone source metadata during chunk materialization

This commit is contained in:
2026-07-18 02:10:05 +00:00
parent 205e2a9908
commit 561d65a505
6 changed files with 279 additions and 77 deletions

View File

@@ -162,7 +162,10 @@ func MaterializeChunkPlan(doc *SourceDocument, plan ChunkPlan) ([]Chunk, error)
for index, chunkRange := range plan.Ranges {
start, _ := UnitIndex(doc, chunkRange.StartUnitID)
end, _ := UnitIndex(doc, chunkRange.EndUnitID)
units := cloneSourceUnits(doc.Units[start : end+1])
units, err := cloneSourceUnits(doc.Units[start : end+1])
if err != nil {
return nil, fmt.Errorf("clone chunk plan range[%d] units: %w", index, err)
}
content, err := json.Marshal(struct {
Units []SourceUnit `json:"units"`
}{Units: units})
@@ -189,52 +192,18 @@ func MaterializeChunkPlan(doc *SourceDocument, plan ChunkPlan) ([]Chunk, error)
return chunks, nil
}
func cloneSourceUnits(units []SourceUnit) []SourceUnit {
func cloneSourceUnits(units []SourceUnit) ([]SourceUnit, error) {
if len(units) == 0 {
return nil
return nil, nil
}
cloned := make([]SourceUnit, len(units))
for i, unit := range units {
cloned[i] = unit
cloned[i].Metadata = cloneJSONMap(unit.Metadata)
}
return cloned
}
func cloneJSONMap(values map[string]any) map[string]any {
if len(values) == 0 {
return nil
}
cloned := make(map[string]any, len(values))
for key, value := range values {
cloned[key] = cloneJSONValue(value)
}
return cloned
}
func cloneJSONValue(value any) any {
switch typed := value.(type) {
case map[string]any:
return cloneJSONMap(typed)
case []any:
cloned := make([]any, len(typed))
for i := range typed {
cloned[i] = cloneJSONValue(typed[i])
metadata, err := CloneMetadata(unit.Metadata)
if err != nil {
return nil, fmt.Errorf("source unit[%d] metadata: %w", i, err)
}
return cloned
case json.RawMessage:
return append(json.RawMessage(nil), typed...)
case []byte:
return append([]byte(nil), typed...)
case []string:
return append([]string(nil), typed...)
case map[string]string:
cloned := make(map[string]string, len(typed))
for key, item := range typed {
cloned[key] = item
}
return cloned
default:
return value
cloned[i].Metadata = metadata
}
return cloned, nil
}