L0 Data Capture / Quantpylib Binary eNcoding (QBN)

Quantpylib Binary eNcoding (QBN) is quantpylib's binary capture format for native market-data events. L0 data capture is the durable, lowest-level event log beneath higher-level functions such as market data replay and analytics. The capture path records the normalized event objects emitted by wrappers and gateway subscriptions. These artefacts can later be used to inspect trading system market data latency, backtesting, or combined with event traces to get markouts, et cetera.

The supported L0 event schemas are:

Use quantpylib.hft.feed.Feed to subscribe to market data and pass quantpylib.hft.l0.L0ArchiveSink into the feed as a global sink or per-feed sink. Feed owns event fanout; L0ArchiveSink owns queued durable capture; Writer and Reader classes are responsible for encoding and decoding QBN files.

Capture Path

wrapper websocket/http message
        |
        v
gateway executor subscription
        |
        v
schema-2 native event
BookUpdate / BBAUpdate / TradeUpdate
        |
        v
Feed fanout
        |
        +--> on_msg callback / materializer sinks
        |
        +--> L0ArchiveSink
                 |
                 v
              writer queue
                 |
                 v
              L0DirectArchiveSink
                 |
                 v
              L0Writer -> QBN file

Typical usage captures every event from a feed while keeping file IO off the feed callback path:

from quantpylib.hft.feed import Feed
from quantpylib.hft.l0 import L0ArchiveSink

sink = L0ArchiveSink(root="./archives")
feed = Feed(gateway=gateway, sinks=[sink])

await feed.add_l2_book_feed(
    exc="binance",
    ticker="BTCUSDT",
    depth=20,
)

Per-feed sinks can also be used when only selected subscriptions should be captured:

book_feed = await feed.add_l2_book_feed(
    exc="binance",
    ticker="BTCUSDT",
    sinks=[L0ArchiveSink(root="./archives")],
)

That's all. A .qbn file will be written. To understand the file layout and custom policies, read on.

Why QBN

Native event fidelity: QBN records quantpylib.standards.models.BookUpdate, quantpylib.standards.models.BBAUpdate, and quantpylib.standards.models.TradeUpdate events as they happened with lossless capture for critical fields used for event reconstruction.
Append-friendly writes: writers append sealed blocks and can reopen existing files.
Partial-tail recovery: readers ignore incomplete trailing blocks; writers can truncate an incomplete tail before appending.
Block-level validation: each block carries a CRC over the block header, instrument dictionary, and record payload.
Self-describing: the file header stores the schema id, record-domain string, venue, market class, and feed type with the captured data.
Compact symbology: records store a local_id; the per-block dictionary stores local_id -> symbol once for that block.
Index-friendly time-range scanning: block headers store exchange-time and receive-time bounds, so readers can skip out-of-range blocks before decoding payloads. Applications can design custom indexing based on file format and application semantics.

File Layout

A QBN file starts with one file header followed by zero or more sealed blocks. All fixed-width numeric fields are little-endian.

QBN file
|
+-- FileHeader
|
+-- Block 0
|   |
|   +-- BlockHeader
|   +-- InstrumentDictionary
|   +-- RecordPayload
|       |
|       +-- Record 0
|       +-- Record 1
|       +-- ...
|
+-- Block 1
|   |
|   +-- BlockHeader
|   +-- InstrumentDictionary
|   +-- RecordPayload
|
+-- ...

File Header

Field	Type	Description
`FILE_MAGIC`	`char[4]`	File magic. Current value: `QBN1`.
`schema_id`	`uint32`	Event schema id.
`record_domain_length`	`uint32`	Byte length of `record_domain`.
`record_domain`	`bytes`	UTF-8 record-domain string describing the record layout.
`venue_length`, `venue`	`uint16`, `bytes`	Venue name.
`market_class_length`, `market_class`	`uint16`, `bytes`	Market class.
`feed_type_length`, `feed_type`	`uint16`, `bytes`	Feed type.

Current schema ids:

Schema id	Event object	Reader	Writer
`20`	BookUpdate	L0BookUpdateReader	L0BookUpdateWriter
`21`	BBAUpdate	L0BBAUpdateReader	L0BBAUpdateWriter
`22`	TradeUpdate	L0TradeUpdateReader	L0TradeUpdateWriter

Block Layout

Blocks are the unit of validation and partial recovery.

Block
|
+-- BlockHeader
|   magic = QBLK1
|   block_length
|   record_count
|   min_ts_exch_ns / max_ts_exch_ns
|   min_ts_recv_ns / max_ts_recv_ns
|   dictionary_length
|   payload_length
|   crc
|
+-- InstrumentDictionary
|   symbol_count
|   repeated local_id, symbol_length, symbol
|
+-- RecordPayload
    repeated schema-specific QBN records

Field	Type	Description
`BLOCK_MAGIC`	`char[5]`	Block magic. Current value: `QBLK1`.
`block_length`	`uint32`	Total block bytes, including header, dictionary, and payload.
`record_count`	`uint32`	Number of records in the payload.
`min_ts_exch_ns`	`uint64`	Minimum exchange-event timestamp in the block.
`max_ts_exch_ns`	`uint64`	Maximum exchange-event timestamp in the block.
`min_ts_recv_ns`	`uint64`	Minimum local receive timestamp in the block.
`max_ts_recv_ns`	`uint64`	Maximum local receive timestamp in the block.
`dictionary_length`	`uint32`	Instrument dictionary byte length.
`payload_length`	`uint32`	Record payload byte length.
`crc`	`uint32`	CRC32 over the zero-CRC header plus dictionary and payload.

Blocks seal when either max_payload_bytes would be exceeded or max_block_seconds has elapsed since the first record in the pending block. Defaults are 32 MiB and 60 seconds.

Instrument Dictionary

Each block has its own compact symbol dictionary.

Field	Type	Description
`symbol_count`	`uint32`	Number of symbols in the block.
`local_id`	`uint32`	Block-local symbol id referenced by records.
`symbol_length`	`uint16`	Symbol byte length.
`symbol`	`bytes`	UTF-8 symbol.

Runtime event bytes are self-contained and include the ticker string. QBN record bytes are archive-local and replace the ticker string with local_id. The dictionary stores the symbol once per block.

Record Layouts

The payload is a sequence of records for the file schema. A file contains one schema id and one record domain.

BookUpdate

BookUpdate records use schema id 20.

Field	Type	Description
`record_length`	`uint32`	Total record bytes.
`local_id`	`uint32`	Instrument dictionary id.
`data_length`	`uint32`	Book-level buffer byte length.
`flags`	`uint32`	Snapshot, timestamp, and ordering flags.
`nbids`	`uint32`	Number of bid levels.
`nasks`	`uint32`	Number of ask levels.
`seq_type`	`uint32`	Sequence mode.
`reserved0`	`uint32`	Reserved, must be zero.
`ts_exch_ns`	`uint64`	Exchange-event timestamp.
`ts_disp_ns`	`uint64`	Provider dispatch timestamp.
`ts_recv_ns`	`uint64`	Local receive timestamp.
`seq0`, `seq1`, `seq2`	`int64`	Sequence fields.
`data`	`int64[]`	Bid price/size pairs followed by ask price/size pairs.

BBAUpdate

BBAUpdate records use schema id 21.

Field	Type	Description
`record_length`	`uint32`	Total record bytes.
`local_id`	`uint32`	Instrument dictionary id.
`flags`	`uint32`	Timestamp and side-availability flags.
`seq_type`	`uint32`	Sequence mode.
`reserved0`, `reserved1`	`uint32`	Reserved, must be zero.
`ts_exch_ns`	`uint64`	Exchange-event timestamp.
`ts_disp_ns`	`uint64`	Provider dispatch timestamp.
`ts_recv_ns`	`uint64`	Local receive timestamp.
`seq0`, `seq1`, `seq2`	`int64`	Sequence fields.
`bid_price`, `bid_size`	`int64`	Best bid price and size.
`ask_price`, `ask_size`	`int64`	Best ask price and size.

TradeUpdate

TradeUpdate records use schema id 22.

Field	Type	Description
`record_length`	`uint32`	Total record bytes.
`local_id`	`uint32`	Instrument dictionary id.
`flags`	`uint32`	Timestamp flags.
`seq_type`	`uint32`	Sequence mode.
`ts_exch_ns`	`uint64`	Exchange-event timestamp.
`ts_disp_ns`	`uint64`	Provider dispatch timestamp.
`ts_recv_ns`	`uint64`	Local receive timestamp.
`seq0`, `seq1`, `seq2`	`int64`	Sequence fields.
`price`	`int64`	Trade price.
`size`	`int64`	Trade size.
`side`	`int32`	Aggressor side: buy, sell, or unknown.
`reserved0`	`uint32`	Reserved, must be zero.

The flags fields use the native model flag semantics documented in quantpylib.standards.models. Price and size fields use the scale documented in quantpylib.standards.models.

File Routing

Archive sinks route each event by calling:

context = policy.resolve(feed_id, event, metrics)

resolve() is the filesystem-policy contract. It must return an L0FileContext with:

Field	Meaning
`path`	Destination `.qbn` file path.
`venue`	Venue string written to the QBN file header.
`market_class`	Market class written to the QBN file header.
`feed_type`	Feed type written to the QBN file header.

Custom policies may route by feed id, event timestamp, current metrics, or another application-specific partitioning rule. The default DailyCaptureFileSystemPolicy uses event.ts_recv_ns and routes files by receive date:

root/venue/market_class/feed_type/YYYY-MM-DD.qbn

FeedTypeFileSystemPolicy is available for feed-type, ticker, or parameter-granular files. It accepts the same resolve(feed_id, event, metrics) call shape, but routes from feed_id and does not require the event.

Reading And Exporting

Use schema-specific readers when the schema is known, or reader_for_file when the file header should select the reader.

from quantpylib.hft.l0 import L0BookUpdateReader, reader_for_file

reader = L0BookUpdateReader("./archives/binance/perp/l2book/2026-07-08.qbn")
records = list(reader.records(symbol="BTCUSDT"))

reader = reader_for_file("./archives/binance/perp/l2book/2026-07-08.qbn")
df = reader.to_pandas(include_offsets=True)
parquet_path = reader.to_parquet()

Reader filters use exchange-event time:

start: inclusive ts_exch_ns.
end: exclusive ts_exch_ns.
symbol: event ticker after dictionary resolution.

Native Objects

Native event objects are documented in quantpylib.standards.models. The L0 reader decodes records back into those same object types; downstream materializers and replay tools should consume the native event objects rather than parsing QBN bytes directly unless they are implementing a new low-level reader.

Capture Sinks

`L0ArchiveSink`

Bases: FeedSink

Queued feed sink for live L0/QBN capture.

L0ArchiveSink accepts feed events without doing file IO on the feed callback path. Events are queued and written by a background L0DirectArchiveSink. Queue overflow behavior is controlled by backpressure.

Parameters:

Name	Type	Description	Default
`root`	`str \| Path`	Archive root used with the default daily capture policy.	`None`
`policy`	`L0FileSystemPolicy`	File routing policy. When not supplied, `root` is required and `DailyCaptureFileSystemPolicy` is used.	`None`
`max_payload_bytes`	`int`	Maximum unsealed block payload bytes.	`DEFAULT_L0_MAX_PAYLOAD_BYTES`
`max_block_seconds`	`float`	Maximum block age before sealing.	`DEFAULT_L0_MAX_BLOCK_SECONDS`
`writer_idle_seconds`	`float`	Idle age after which non-active writers are closed. Defaults to three block-age windows.	`None`
`writer_retire_interval_records`	`int`	Record interval for checking idle writers.	`1000`
`fsync`	`bool`	Whether to fsync after sealed block writes.	`False`
`validate_existing`	`bool`	Whether to validate existing blocks before appending to an existing QBN file.	`True`
`maxsize`	`int`	Queue capacity.	`QUEUE_MAXSIZE`
`backpressure`	`FeedSinkBackpressure \| str`	Queue overflow behavior. `FAIL` raises `FeedSinkBackpressureError`; `DROP_NEW` drops the new event and increments drop metrics.	`FAIL`
`logger`	`Logger`	Logger for queue, metrics, writer lifecycle, and write failures.	`None`
`metrics_log_interval_sec`	`float \| None`	Background metrics logging interval. Use `None` to disable background metrics logging.	`DEFAULT_L0_METRICS_LOG_INTERVAL_SECONDS`

`metrics` `property`

Sink and writer metrics.

`writers` `property`

Active QBN writers keyed by path.

`close()`

Stop the worker, flush, and close active writers. Idempotent.

`flush()`

Wait for queued events to be written and flush active writers.

`write(feed_id, event)`

Enqueue one native event for L0 capture.

Parameters:

Name	Type	Description	Default
`feed_id`	`FeedId`	Feed identity.	required
`event`	`BookUpdate \| BBAUpdate \| TradeUpdate`	Native event to capture.	required

Returns:

Name	Type	Description
`bool`		`True` when the event was enqueued, `False` when the sink is already closed or the event was dropped by `DROP_NEW` backpressure.

Raises:

Type	Description
`L0WriteError`	If the worker has failed.
`FeedSinkBackpressureError`	If the queue is full and backpressure mode is `FAIL`.

`L0DirectArchiveSink`

Bases: FeedSink

Synchronous feed sink that writes native events to L0/QBN files.

L0DirectArchiveSink owns one writer per resolved QBN path. It is useful when the caller wants file IO on the feed callback path or when another queue already owns backpressure. Most live feed capture should use L0ArchiveSink, which moves file IO to a worker thread.

Parameters:

Name	Type	Description	Default
`root`	`str \| Path`	Archive root used with the default daily capture policy.	`None`
`policy`	`L0FileSystemPolicy`	File routing policy. When not supplied, `root` is required and `DailyCaptureFileSystemPolicy` is used.	`None`
`max_payload_bytes`	`int`	Maximum unsealed block payload bytes.	`DEFAULT_L0_MAX_PAYLOAD_BYTES`
`max_block_seconds`	`float`	Maximum block age before sealing.	`DEFAULT_L0_MAX_BLOCK_SECONDS`
`writer_idle_seconds`	`float`	Idle age after which non-active writers are closed. Defaults to three block-age windows.	`None`
`writer_retire_interval_records`	`int`	Record interval for checking idle writers.	`1000`
`fsync`	`bool`	Whether to fsync after sealed block writes.	`False`
`validate_existing`	`bool`	Whether to validate existing blocks before appending to an existing QBN file.	`True`
`metrics`	`L0DirectArchiveMetrics`	Metrics object to update.	`None`
`logger`	`Logger`	Logger for metrics, writer lifecycle, and write failures.	`None`
`metrics_log_interval_sec`	`float \| None`	Background metrics logging interval. Use `None` to disable background metrics logging.	`DEFAULT_L0_METRICS_LOG_INTERVAL_SECONDS`

`close()`

Flush and close every active writer. Idempotent.

`flush()`

Seal pending blocks and flush all active writers.

`log_metrics()`

Emit a metrics snapshot through the configured logger.

`write(feed_id, event)`

Write one native event to the resolved QBN file.

Parameters:

Name	Type	Description	Default
`feed_id`	`FeedId`	Feed identity.	required
`event`	`BookUpdate \| BBAUpdate \| TradeUpdate`	Native event to capture.	required

Returns:

Name	Type	Description
`bool`		`True` when the event was written, `False` when the sink is already closed.

Raises:

Type	Description
`L0WriteError`	If the event cannot be written.

Filesystem Policies

`L0FileSystemPolicy`

Base contract for L0/QBN filesystem routing policies.

Archive sinks call resolve(feed_id, event, metrics) for each captured event to choose the destination QBN file and file-header context. Custom policies should return an L0FileContext and may use feed_id, event timestamps, or current sink metrics to choose a path.

`resolve(feed_id, event, metrics)`

Resolve an event destination.

Parameters:

Name	Type	Description	Default
`feed_id`	`FeedId`	Feed identity.	required
`event`	`Any`	Native schema-2 event being captured.	required
`metrics`	`L0DirectArchiveMetrics`	Current sink metrics.	required

Returns:

Name	Type	Description
`L0FileContext`	`L0FileContext`	Destination file and header context.

`DailyCaptureFileSystemPolicy`

Bases: L0FileSystemPolicy

Resolve QBN files by receive-time calendar date.

The default layout is root/venue/market_class/feed_type/YYYY-MM-DD.qbn.

Parameters:

Name	Type	Description	Default
`root`	`str \| Path`	Archive root.	`'.'`
`suffix`	`str`	File suffix. Defaults to `.qbn`.	`'.qbn'`
`tz`	`tzinfo`	Time zone used to derive the receive-date partition from `event.ts_recv_ns`.	`utc`

`resolve(feed_id, event, metrics=None)`

Resolve a destination from feed identity and receive date.

The receive-date partition is derived from event.ts_recv_ns in this policy's timezone. The returned context path is root/venue/market_class/feed_type/YYYY-MM-DD.qbn by default.

Parameters:

Name	Type	Description	Default
`feed_id`	`FeedId`	Feed identity.	required
`event`	`Any`	Native schema-2 event with `ts_recv_ns`.	required
`metrics`	`L0DirectArchiveMetrics`	Current sink metrics.	`None`

Returns:

Name	Type	Description
`L0FileContext`	`L0FileContext`	Destination file and header context.

`FeedTypeFileSystemPolicy`

Bases: L0FileSystemPolicy

Resolve QBN files by feed context instead of receive date.

Parameters:

Name	Type	Description	Default
`root`	`str \| Path`	Archive root.	`'.'`
`suffix`	`str`	File suffix. Defaults to `.qbn`.	`'.qbn'`
`granularity`	`str`	File grouping mode. Supported values are `feed_type`, `ticker`, and `params`.	`'feed_type'`

`resolve(feed_id, event=None, metrics=None)`

Resolve a destination from feed identity.

event and metrics are accepted for the policy interface but are not used by this policy. The returned context path is rooted at root/venue/market_class and then grouped by granularity.

Parameters:

Name	Type	Description	Default
`feed_id`	`FeedId`	Feed identity.	required
`event`	`Any`	Ignored.	`None`
`metrics`	`L0DirectArchiveMetrics`	Ignored.	`None`

Returns:

Name	Type	Description
`L0FileContext`	`L0FileContext`	Destination file and header context.

`L0FileContext` `dataclass`

File destination and semantic context for an L0 writer.

Parameters:

Name	Type	Description	Default
`path`	`Path`	QBN file path.	required
`venue`	`str`	Venue name stored in the file header.	required
`market_class`	`str`	Market class stored in the file header.	required
`feed_type`	`str`	Feed type stored in the file header.	required

Writers

`L0Writer`

Base writer for one L0/QBN event schema.

A writer owns one QBN file and accepts one native event type. The file header is written lazily on the first event. Existing files are reopened by validating complete blocks and truncating any incomplete trailing block before appending.

Parameters:

Name	Type	Description	Default
`context`	`L0FileContext`	Destination file and header context.	required
`max_payload_bytes`	`int`	Maximum unsealed block payload bytes.	`DEFAULT_L0_MAX_PAYLOAD_BYTES`
`max_block_seconds`	`float`	Maximum block age before sealing.	`DEFAULT_L0_MAX_BLOCK_SECONDS`
`fsync`	`bool`	Whether to fsync after sealed block writes.	`False`
`validate_existing`	`bool`	Whether to validate existing blocks before appending to an existing QBN file.	`True`
`metrics`	`L0FileMetrics`	File metrics object to update.	`None`

`close()`

Flush and close the file handle. Idempotent.

`flush()`

Seal the pending block and flush the file handle.

`record_length(event)` `classmethod`

Return the QBN record length for an event.

`write(feed_id, event)`

Append one event to the current block.

Parameters:

Name	Type	Description	Default
`feed_id`	`FeedId`	Feed identity used when creating a new file.	required
`event`	`BookUpdate \| BBAUpdate \| TradeUpdate`	Native event matching the writer schema.	required

Raises:

Type	Description
`L0WriteError`	If the writer is closed or the event type does not match the writer schema.

`L0BookUpdateWriter`

Bases: L0Writer

L0/QBN writer for quantpylib.standards.models.BookUpdate events.

`L0BBAUpdateWriter`

Bases: L0Writer

L0/QBN writer for quantpylib.standards.models.BBAUpdate events.

`L0TradeUpdateWriter`

Bases: L0Writer

L0/QBN writer for quantpylib.standards.models.TradeUpdate events.

Readers

`reader_for_file(path)`

Open a QBN file with the reader matching its header schema id.

Parameters:

Name	Type	Description	Default
`path`	`str \| Path`	QBN file path.	required

Returns:

Type	Description
`L0Reader`	Reader instance for the file schema.

`L0Reader`

Base reader for one L0/QBN event schema.

Readers validate the file schema, skip incomplete trailing blocks, validate complete block CRCs, and decode QBN records back into native event objects.

Parameters:

Name	Type	Description	Default
`path`	`str \| Path`	QBN file path.	required

`blocks(*, symbol=None, start=None, end=None)`

Iterate decoded QBN blocks.

Parameters:

Name	Type	Description	Default
`symbol`	`str`	Symbol filter.	`None`
`start`	`int`	Inclusive exchange-event timestamp filter in nanoseconds.	`None`
`end`	`int`	Exclusive exchange-event timestamp filter in nanoseconds.	`None`

Yields:

Type	Description
`dict`	Block metadata with decoded `symbols` and `records`.

`header()`

Read the QBN file header.

Returns:

Type	Description
`dict`	Header fields: `schema_id`, `record_domain`, `venue`, `market_class`, and `feed_type`.

`records(*, symbol=None, start=None, end=None, with_context=False)`

Iterate decoded native events.

Parameters:

Name	Type	Description	Default
`symbol`	`str`	Symbol filter.	`None`
`start`	`int`	Inclusive exchange-event timestamp filter in nanoseconds.	`None`
`end`	`int`	Exclusive exchange-event timestamp filter in nanoseconds.	`None`
`with_context`	`bool`	When true, yield dictionaries containing file context, block offset, record index, symbol, and `event`.	`False`

Yields:

Type	Description
`BookUpdate \| BBAUpdate \| TradeUpdate \| dict`	Native event, or a context dictionary when `with_context` is true.

`scan_valid_blocks()`

Return all currently valid decoded blocks.

Returns:

Type	Description
`list`	Decoded block dictionaries.

`to_pandas(*, symbol=None, start=None, end=None, include_offsets=False)`

Materialize decoded records as a pandas DataFrame.

Parameters:

Name	Type	Description	Default
`symbol`	`str`	Symbol filter.	`None`
`start`	`int`	Inclusive exchange-event timestamp filter in nanoseconds.	`None`
`end`	`int`	Exclusive exchange-event timestamp filter in nanoseconds.	`None`
`include_offsets`	`bool`	Include `block_offset` and `record_index` columns.	`False`

Returns:

Type	Description
`DataFrame`	Decoded records in tabular form.

`to_parquet(path=None, *, symbol=None, start=None, end=None, include_offsets=False, **kwargs)`

Export decoded records to a derived Parquet file.

Parameters:

Name	Type	Description	Default
`path`	`str \| Path`	Output path. Defaults to the QBN path with a `.parquet` suffix.	`None`
`symbol`	`str`	Symbol filter.	`None`
`start`	`int`	Inclusive exchange-event timestamp filter in nanoseconds.	`None`
`end`	`int`	Exclusive exchange-event timestamp filter in nanoseconds.	`None`
`include_offsets`	`bool`	Include `block_offset` and `record_index` columns.	`False`
`**kwargs`	`Any`	Forwarded to `pandas.DataFrame.to_parquet`.	`{}`

Returns:

Type	Description
`Path`	Parquet output path.

`L0BookUpdateReader`

Bases: L0Reader

L0/QBN reader for quantpylib.standards.models.BookUpdate files.

`L0BBAUpdateReader`

Bases: L0Reader

L0/QBN reader for quantpylib.standards.models.BBAUpdate files.

`L0TradeUpdateReader`

Bases: L0Reader

L0/QBN reader for quantpylib.standards.models.TradeUpdate files.

L0 Data Capture / Quantpylib Binary eNcoding (QBN)

Capture Path

Why QBN

File Layout

File Header

Block Layout

Instrument Dictionary

Record Layouts

BookUpdate

BBAUpdate

TradeUpdate

File Routing

Reading And Exporting

Native Objects

Capture Sinks

L0ArchiveSink

metrics property

writers property

close()

flush()

write(feed_id, event)

L0DirectArchiveSink

close()

flush()

log_metrics()

write(feed_id, event)

Filesystem Policies

L0FileSystemPolicy

resolve(feed_id, event, metrics)

DailyCaptureFileSystemPolicy

resolve(feed_id, event, metrics=None)

FeedTypeFileSystemPolicy

resolve(feed_id, event=None, metrics=None)

L0FileContext dataclass

Writers

L0Writer

close()

flush()

record_length(event) classmethod

write(feed_id, event)

L0BookUpdateWriter

L0BBAUpdateWriter

L0TradeUpdateWriter

Readers

reader_for_file(path)

L0Reader

blocks(*, symbol=None, start=None, end=None)

header()

records(*, symbol=None, start=None, end=None, with_context=False)

scan_valid_blocks()

to_pandas(*, symbol=None, start=None, end=None, include_offsets=False)

to_parquet(path=None, *, symbol=None, start=None, end=None, include_offsets=False, **kwargs)

L0BookUpdateReader

L0BBAUpdateReader

L0TradeUpdateReader

`L0ArchiveSink`

`metrics` `property`

`writers` `property`

`close()`

`flush()`

`write(feed_id, event)`

`L0DirectArchiveSink`

`close()`

`flush()`

`log_metrics()`

`write(feed_id, event)`

`L0FileSystemPolicy`

`resolve(feed_id, event, metrics)`

`DailyCaptureFileSystemPolicy`

`resolve(feed_id, event, metrics=None)`

`FeedTypeFileSystemPolicy`

`resolve(feed_id, event=None, metrics=None)`

`L0FileContext` `dataclass`

`L0Writer`

`close()`

`flush()`

`record_length(event)` `classmethod`

`write(feed_id, event)`

`L0BookUpdateWriter`

`L0BBAUpdateWriter`

`L0TradeUpdateWriter`

`reader_for_file(path)`

`L0Reader`

`blocks(*, symbol=None, start=None, end=None)`

`header()`

`records(*, symbol=None, start=None, end=None, with_context=False)`

`scan_valid_blocks()`

`to_pandas(*, symbol=None, start=None, end=None, include_offsets=False)`

`to_parquet(path=None, *, symbol=None, start=None, end=None, include_offsets=False, **kwargs)`

`L0BookUpdateReader`

`L0BBAUpdateReader`

`L0TradeUpdateReader`