Deserialization Skill

Converts raw binary message data into structured JavaScript objects.

Flow

Raw bytes (from source)
    │
    ▼
DeserializingIterableSource (applies parseChannel-based decode)
    │
    ▼
Decoded message objects (ready for panels)

Schema Encodings

Core Entry Point: parseChannel()

packages/mcap-support/src/parseChannel.ts:

function parseChannel(channel: Channel): ParsedChannel {
  // Returns { deserialize: (data: ArrayBufferView) => unknown, datatypes: Map }
  // Dispatches based on channel.schema.encoding
}

This is the single function that resolves any channel's schema into a deserializer.

Per-encoding parsers

• packages/mcap-support/src/parseProtobufSchema.ts
• packages/mcap-support/src/parseFlatbufferSchema.ts
• packages/mcap-support/src/parseJsonSchema.ts

DeserializingIterableSource

packages/suite-base/src/players/IterablePlayer/DeserializingIterableSource.ts:

• Wraps a serialized source (ISerializedIterableSource)
• Applies parseChannel() to each channel's schema at initialization
• Deserializes messages on-the-fly as they are iterated
• Produces IDeserializedIterableSource (messages are JS objects, not raw bytes)
• Supports sampling (setSamplingWindowEnd() for latest-per-render-tick mode)

⚠️ This wrapper is only inserted for serialized sources. Sources that already implement IDeserializedIterableSource skip it.

WASM Decompression Handlers

packages/mcap-support/src/decompressHandlers.ts:

• zstd: @lichtblick/wasm-zstd — best compression ratio
• lz4: @lichtblick/wasm-lz4 — fastest decompression
• bz2: @lichtblick/wasm-bz2 — legacy support

Handlers are loaded once (singleton promise) and shared across all readers.

Performance Notes

• Deserialization is CPU-bound — runs in a Worker via WorkerIterableSource
• Protobuf: root.lookupType() is cached after first call
• JSON: postprocessValue handles base64 → Uint8Array for bytes fields
• Flatbuffers: zero-copy read from the shared buffer (fastest)
• ROS: MessageReader precompiles field offsets at schema-parse time

Skills Reference

• For MCAP chunk structure: load mcap-format skill
• For Worker patterns: load web-workers skill
• For the iterator layering: load player-internals skill