Expand description
Rerun arrow metadata and record batch definitions.
Handles the structure of arrow record batches and their meta data for different use cases for Rerun.
An arrow record batch that follows a specific schema is called a SorbetBatch.
Some SorbetBatches has even more constrained requirements, such as ChunkBatch and DataframeBatch.
- Every
ChunkBatchis aSorbetBatch. - Every
DataframeBatchis aSorbetBatch.
NOTE: DataframeBatch has not yet been implemented.
Each batch type has a matching schema type:
SorbetBatchhas aSorbetSchemaChunkBatchhas aChunkSchemaDataframeBatchwill have aDataframeSchema
Modules§
Structs§
- The [
ArrowRecordBatch] representation of a Rerun chunk. - The parsed schema of a Rerun chunk, i.e. multiple columns of data for a single entity.
- Describes a data/component column, such as
Position3D, in a dataframe. - Describes a time column, such as
log_time. - Describes the schema of the primary [
RowId] column. - Any rerun-compatible [
ArrowRecordBatch]. - The parsed schema of a
SorbetBatch.
Enums§
- The type of
SorbetBatch. - The type of column in a sorbet batch.
Traits§
- Make it more ergonomic to work with arrow metadata.
Functions§
- Encode an arrow schema as IPC bytes.
- Migrate old renamed types to new types.
- Decode an arrow schema from IPC bytes.
Type Aliases§
- Arrow metadata for an arrow record batch.
- Arrow metadata for a column/field.