Crate backbeat_cli 

Library half of the backbeat CLI: the schema-driven dump tooling.

Everything is driven by each dump’s embedded schema registry, with no compiled-in knowledge of the producer’s event types. Living in a library (not just the binary) lets integration tests — and other tools — call these directly.

• model — load + decode one or more .bb dumps into a common in-memory form.
• convert — write the decoded records to sparse-wide Parquet.
• trace — write them to Chrome / Perfetto trace JSON, pairing spans into duration slices.
• inspect — summarize a dump (envelope, registry, per-shard counts).
• merge — splice several .bb dumps into one multi-instance .bb.
• views — generate DuckDB query DDL from a dump’s registry + its registered view sets.

Modules

convert

backbeat convert: decode a dump to sparse-wide Parquet, driven entirely by its embedded schema.

inspect

backbeat inspect: print a dump’s envelope, schema registry, and per-shard record counts.

merge

backbeat merge: combine several .bb dumps into one multi-instance .bb.

model

Shared dump-loading model behind both output formats.

trace

backbeat → Chrome / Perfetto trace JSON.

views

Generates DuckDB query DDL from a dump’s schema registry, and assembles it with the consumer-registered view sets carried in the dump.