Market Data¶

The SDK provides two market data streams:

Stream	Method	Use case
Live exchange (all assets)	`p.market()`	Regular trading
Arena stream	`p.arena_market()`	Arena bots / training

Both arrive over the TCP market relay the server advertises at login.

Getting a subscription¶

market       = p.market()        # live exchange assets
arena_market = p.arena_market()  # arena stream (use in 07a/07b)

Both return a Market object that receives data in a background thread.

Reading the latest snapshot¶

md = market.get_latest(asset_id=1)

if md is None:
    # No data received yet — wait a moment and retry
    pass
else:
    bid = md["bid"]      # best bid price (raw ticks)
    ask = md["ask"]      # best ask price (raw ticks)
    mid = (bid + ask) // 2

    bid_qty = md["bid_qty"]
    ask_qty = md["ask_qty"]

    # Full book depth — list of (price, qty) tuples
    bids = md["bids"]    # [( price, qty ), ...]  best first
    asks = md["asks"]

No mid key

There is no "mid" field. Always compute: mid = (md["bid"] + md["ask"]) // 2. md.get("mid") always returns None.

All available keys¶

Key	Type	Description
`"bid"`	`int`	Best bid price (raw ticks)
`"ask"`	`int`	Best ask price (raw ticks)
`"bid_qty"`	`int`	Volume at best bid
`"ask_qty"`	`int`	Volume at best ask
`"bids"`	`list[(int,int)]`	All bid levels `(price, qty)`
`"asks"`	`list[(int,int)]`	All ask levels `(price, qty)`
`"timestamp"`	`int`	Nanosecond publish timestamp
`"seq_id"`	`int`	Monotone update counter
`"lifecycle_state"`	`int`	Engine state
`"volume_buy"`	`int`	Session cumulative buy volume
`"volume_sell"`	`int`	Session cumulative sell volume
`"ntb"`	`int`	Number of buy trades (session)
`"nts"`	`int`	Number of sell trades (session)
`"nlb"`	`int`	Live bid order count
`"nls"`	`int`	Live ask order count

In an arena trading loop¶

Use p.arena_market() — not conn.lob():

arena_market = p.arena_market()

while RUNNING:
    md  = arena_market.get_latest(asset_id)
    if md is None:
        time.sleep(0.1)
        continue

    bid = md["bid"]
    ask = md["ask"]
    mid = (bid + ask) // 2

    # your signal logic here...
    time.sleep(0.3)

conn.lob() in arena context

conn.lob(asset_id) is always empty in arena trader connections (best_bid=None, best_ask=None). Book data comes from p.arena_market(), not through the trading connection.

Polars export¶

# All snapshots as DataFrame
df = market.to_polars(asset_id=1)

# Public trade tape
trades_df = market.trades_to_polars(asset_id=1)

# Tick event history
ticks_df  = market.ticks_to_polars(asset_id=1)

Status check¶

market.is_connected()    # True if data received in the last 2s
market.history_len()     # number of stored snapshots

Price conventions reminder¶

All prices in the market data dict are raw ticks (u64 integers):

from fastboxex.price import to_display, from_float

raw     = from_float(100.00)   # 1_000_000
display = to_display(raw)      # 100.0

mid_display = to_display((md["bid"] + md["ask"]) // 2)