dataset card: hyperliquid_mainnet_archive part 2 (dedicated collector), bbo_stream config, audited corrections
Browse files
README.md
CHANGED
|
@@ -33,6 +33,8 @@ configs:
|
|
| 33 |
data_files: "hyperliquid_mainnet_archive/funding/*/*.parquet"
|
| 34 |
- config_name: hl_archive_candles_5m
|
| 35 |
data_files: "hyperliquid_mainnet_archive/candles/*/*.parquet"
|
|
|
|
|
|
|
| 36 |
---
|
| 37 |
|
| 38 |
# variouscryptodata
|
|
@@ -45,7 +47,7 @@ nightly where collection is still running.
|
|
| 45 |
|---|---|---|---|
|
| 46 |
| `polymarket_updown_orderbook/` | Polymarket Up/Down (5m/15m) order books, 10 levels, BTC/ETH/SOL/XRP/DOGE/HYPE/BNB, with Binance spot reference | 2026-05-24 → present | appended nightly (previous UTC day) |
|
| 47 |
| `hyperliquid_trades/` | Hyperliquid perp trade prints (every fill) for BTC/ETH/SOL/HYPE/AVAX + 25 HIP-3 tradfi markets (`xyz:*` equities/commodities/FX, `cash:*`, `km:*`) | 2026-07-23 → present | appended nightly (previous UTC day) |
|
| 48 |
-
| `hyperliquid_mainnet_archive/` | Hyperliquid WebSocket capture for 16 perps (15 until 2026-07-20): **L2 order book snapshots (20 levels/side)**, trades, asset context (funding/OI/premium/oracle/mark/mid/impact), top-of-book+depth summary, mark/oracle, funding, 5m candles | 2026-07-18 (
|
| 49 |
|
| 50 |
---
|
| 51 |
|
|
@@ -157,7 +159,7 @@ print(t.group_by("coin").agg(pl.len(), (pl.col("px")*pl.col("sz")).sum().alias("
|
|
| 157 |
|
| 158 |
---
|
| 159 |
|
| 160 |
-
## hyperliquid_mainnet_archive (static
|
| 161 |
|
| 162 |
An 18-day capture (2026-07-18 ~14:03 UTC → 2026-08-04 ~13:25 UTC; first and last day partial) of Hyperliquid's public WebSocket feed by a shadow-trading
|
| 163 |
research bot (no orders were sent from this data). Coins were the bot's
|
|
@@ -168,13 +170,40 @@ per UTC day: `hyperliquid_mainnet_archive/<table>/date=YYYY-MM-DD/<table>.parque
|
|
| 168 |
|
| 169 |
| table | rows/day (≈) | columns |
|
| 170 |
|---|---:|---|
|
| 171 |
-
| `l2book` | 100–300 k | `coin`, `time_ms` (exchange), `recv_ms` (local receive), `bids`, `asks` — JSON `[[px, sz, n_orders], …]`,
|
| 172 |
| `trades` | 0.3–1.3 M | `coin`, `side` (`B` buyer-aggressor / `A` seller-aggressor), `px`, `sz`, `time_ms`, `tid` — de-duplicated on (`coin`,`tid`); wallet addresses and tx hashes removed |
|
| 173 |
| `asset_ctx` | 0.5–1.5 M | `coin`, `recv_ms`, `funding` (hourly rate), `open_interest`, `prev_day_px`, `day_ntl_vlm`, `day_base_vlm`, `premium`, `oracle_px`, `mark_px`, `mid_px`, `impact_bid`, `impact_ask` — streamed `activeAssetCtx` updates |
|
| 174 |
| `bbo` | 100–300 k | `coin`, `recv_ms`, `bid_px`, `bid_sz`, `ask_px`, `ask_sz`, `spread`, `spread_bps`, `bid_depth_sz`, `bid_depth_usd`, `ask_depth_sz`, `ask_depth_usd`, `bid_levels`, `ask_levels` — top of book plus summed depth over the 20 captured levels, computed by the collector for each captured `l2book` snapshot (same row count and ~5 s cadence as `l2book`) |
|
| 175 |
| `mark` | ~250 k | `coin`, `recv_ms`, `mark_px`, `oracle_px` |
|
| 176 |
| `funding` | ~22 k | `coin`, `recv_ms`, `funding_rate` |
|
| 177 |
| `candles` | 2–5 k | `coin`, `interval` (`5m`), `open_ms`, `close_ms`, `open`, `high`, `low`, `close`, `volume`, `trade_count` — final state of each 5m candle |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 178 |
|
| 179 |
Notes: `time_ms`/`open_ms` are exchange timestamps; `recv_ms` is the
|
| 180 |
collector's receive time (single host, best-effort; latency typically
|
|
|
|
| 33 |
data_files: "hyperliquid_mainnet_archive/funding/*/*.parquet"
|
| 34 |
- config_name: hl_archive_candles_5m
|
| 35 |
data_files: "hyperliquid_mainnet_archive/candles/*/*.parquet"
|
| 36 |
+
- config_name: hl_archive_bbo_stream
|
| 37 |
+
data_files: "hyperliquid_mainnet_archive/bbo_stream/*/*.parquet"
|
| 38 |
---
|
| 39 |
|
| 40 |
# variouscryptodata
|
|
|
|
| 47 |
|---|---|---|---|
|
| 48 |
| `polymarket_updown_orderbook/` | Polymarket Up/Down (5m/15m) order books, 10 levels, BTC/ETH/SOL/XRP/DOGE/HYPE/BNB, with Binance spot reference | 2026-05-24 → present | appended nightly (previous UTC day) |
|
| 49 |
| `hyperliquid_trades/` | Hyperliquid perp trade prints (every fill) for BTC/ETH/SOL/HYPE/AVAX + 25 HIP-3 tradfi markets (`xyz:*` equities/commodities/FX, `cash:*`, `km:*`) | 2026-07-23 → present | appended nightly (previous UTC day) |
|
| 50 |
+
| `hyperliquid_mainnet_archive/` | Hyperliquid WebSocket capture for 16 perps (15 until 2026-07-20): **L2 order book snapshots (20 levels/side)**, trades, asset context (funding/OI/premium/oracle/mark/mid/impact), top-of-book+depth summary, mark/oracle, funding, 5m candles | part 1: 2026-07-18 (~14:03 UTC) → 2026-08-04 (~13:25 UTC), 16 perps (legacy bot); part 2: from 2026-08-22 ~19:49 UTC (first day partial), **all ~320 perp + HIP-3 markets** (dedicated collector) | part 2 appended nightly (previous UTC day) |
|
| 51 |
|
| 52 |
---
|
| 53 |
|
|
|
|
| 159 |
|
| 160 |
---
|
| 161 |
|
| 162 |
+
## hyperliquid_mainnet_archive (part 1 static 2026-07-18 → 2026-08-04; part 2 from 2026-08-22, appended nightly)
|
| 163 |
|
| 164 |
An 18-day capture (2026-07-18 ~14:03 UTC → 2026-08-04 ~13:25 UTC; first and last day partial) of Hyperliquid's public WebSocket feed by a shadow-trading
|
| 165 |
research bot (no orders were sent from this data). Coins were the bot's
|
|
|
|
| 170 |
|
| 171 |
| table | rows/day (≈) | columns |
|
| 172 |
|---|---:|---|
|
| 173 |
+
| `l2book` | 100–300 k | `coin`, `time_ms` (exchange), `recv_ms` (local receive), `bids`, `asks` — JSON `[[px, sz, n_orders], …]`, 20 levels per side in part 1 (up to 20 in part 2), best first, full precision (`nSigFigs=null`); one snapshot per coin every ~5.4 s (the exchange's push cadence, not every book update) |
|
| 174 |
| `trades` | 0.3–1.3 M | `coin`, `side` (`B` buyer-aggressor / `A` seller-aggressor), `px`, `sz`, `time_ms`, `tid` — de-duplicated on (`coin`,`tid`); wallet addresses and tx hashes removed |
|
| 175 |
| `asset_ctx` | 0.5–1.5 M | `coin`, `recv_ms`, `funding` (hourly rate), `open_interest`, `prev_day_px`, `day_ntl_vlm`, `day_base_vlm`, `premium`, `oracle_px`, `mark_px`, `mid_px`, `impact_bid`, `impact_ask` — streamed `activeAssetCtx` updates |
|
| 176 |
| `bbo` | 100–300 k | `coin`, `recv_ms`, `bid_px`, `bid_sz`, `ask_px`, `ask_sz`, `spread`, `spread_bps`, `bid_depth_sz`, `bid_depth_usd`, `ask_depth_sz`, `ask_depth_usd`, `bid_levels`, `ask_levels` — top of book plus summed depth over the 20 captured levels, computed by the collector for each captured `l2book` snapshot (same row count and ~5 s cadence as `l2book`) |
|
| 177 |
| `mark` | ~250 k | `coin`, `recv_ms`, `mark_px`, `oracle_px` |
|
| 178 |
| `funding` | ~22 k | `coin`, `recv_ms`, `funding_rate` |
|
| 179 |
| `candles` | 2–5 k | `coin`, `interval` (`5m`), `open_ms`, `close_ms`, `open`, `high`, `low`, `close`, `volume`, `trade_count` — final state of each 5m candle |
|
| 180 |
+
| `bbo_stream` (part 2 only) | ~5–10 M | `coin`, `time_ms`, `recv_ms`, `bid_px`, `bid_sz`, `bid_n`, `ask_px`, `ask_sz`, `ask_n` — Hyperliquid's high-frequency `bbo` WebSocket channel (every top-of-book change) for the ~40 highest-volume markets |
|
| 181 |
+
|
| 182 |
+
**Part 2 (from 2026-08-22 ~19:49 UTC, first day partial; appended nightly):**
|
| 183 |
+
a dedicated read-only collector subscribes `l2Book` + `trades` for **every**
|
| 184 |
+
perp on the main exchange and every HIP-3 market (≈320 markets, re-discovered
|
| 185 |
+
every 6 h, so newly listed markets appear automatically), and `bbo` for the
|
| 186 |
+
~40 highest-volume markets (ranked at collector start; new high-volume entrants
|
| 187 |
+
are added at discovery, none removed). Same table layout and column names as
|
| 188 |
+
part 1, with these differences:
|
| 189 |
+
|
| 190 |
+
- `asset_ctx`, `mark` and `funding` come from the REST `metaAndAssetCtxs`
|
| 191 |
+
endpoint once per minute (part 1: streamed, ~5 s).
|
| 192 |
+
- `bbo` is derived from each `l2book` snapshot (identical definition to part 1).
|
| 193 |
+
- `l2book`: up to 20 levels per side — thin HIP-3 markets can have fewer (see
|
| 194 |
+
`bid_levels`/`ask_levels`); the ~5.4 s cadence is the exchange's push rate.
|
| 195 |
+
- `candles` are 5-minute bars built from `trades` captured live (receive
|
| 196 |
+
latency ≤ 60 s), only buckets with ≥ 1 trade (part 1: exchange candle stream
|
| 197 |
+
incl. empty buckets); buckets around collector (re)starts may be partial.
|
| 198 |
+
- `trades`: on every (re)subscription Hyperliquid replays the last ~30 trades
|
| 199 |
+
of a market, so each daily file can contain a few older trades per market —
|
| 200 |
+
filter on `time_ms` if that matters.
|
| 201 |
+
- Part-2 volumes: `l2book`/`bbo` ≈ 5 M rows/day, `trades` ≈ 5–8 M,
|
| 202 |
+
`bbo_stream` ≈ 9–11 M, `asset_ctx`/`mark`/`funding` ≈ 0.46 M, `candles` ≤ 92 k.
|
| 203 |
+
- Receive latency (recv_ms − time_ms): l2book ≈ 0.6 s median / ≈ 1–2 s p99;
|
| 204 |
+
trades/bbo ≈ 0.35 s median; bursts up to ~10 s at (re)subscribe.
|
| 205 |
+
|
| 206 |
+
Gap between part 1 and part 2: 2026-08-04 13:25 → 2026-08-22 19:49 UTC.
|
| 207 |
|
| 208 |
Notes: `time_ms`/`open_ms` are exchange timestamps; `recv_ms` is the
|
| 209 |
collector's receive time (single host, best-effort; latency typically
|