dexbot 1.6.15 → 1.6.16

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (102) hide show
  1. package/CHANGELOG.md +26 -0
  2. package/README.md +4 -4
  3. package/claw/docs/DEXBOT2_TUNING_CHEAT_SHEET.md +3 -2
  4. package/claw/modules/chain_queries.ts +3 -3
  5. package/claw/package.json +1 -1
  6. package/claw/runtimes/openclaw-plugin/openclaw.plugin.json +1 -1
  7. package/claw/runtimes/openclaw-plugin/package.json +1 -1
  8. package/dist/analysis/ama_fitting/fetch_lp_candles.js +8 -1
  9. package/dist/analysis/ama_fitting/optimizer_high_resolution.d.ts +50 -1
  10. package/dist/analysis/ama_fitting/optimizer_high_resolution.js +33 -7
  11. package/dist/analysis/bot_fitting/backtest_bot_fitting.d.ts +8 -6
  12. package/dist/analysis/bot_fitting/backtest_bot_fitting.js +8 -10
  13. package/dist/analysis/bot_key_utils.js +9 -7
  14. package/dist/analysis/chart_utils.d.ts +9 -2
  15. package/dist/analysis/chart_utils.js +51 -1
  16. package/dist/analysis/tradingview/tradingview_uplot_chart_generator.d.ts +2 -1
  17. package/dist/analysis/tradingview/tradingview_uplot_chart_generator.js +19 -5
  18. package/dist/analysis/trend_detection/dynamic_weight_chart_generator.js +21 -2
  19. package/dist/analysis/trend_detection/kalman_chart_generator.js +1 -0
  20. package/dist/analysis/trend_detection/regime_chart_generator.js +1 -1
  21. package/dist/analysis/trend_detection/volatility_chart_generator.js +1 -0
  22. package/dist/dexbot.js +89 -138
  23. package/dist/market_adapter/inputs/fetch_book_data.js +36 -12
  24. package/dist/market_adapter/inputs/fetch_lp_data.d.ts +6 -0
  25. package/dist/market_adapter/inputs/fetch_lp_data.js +65 -19
  26. package/dist/market_adapter/inputs/kibana_feed_source.js +45 -12
  27. package/dist/market_adapter/inputs/window_cache.d.ts +14 -1
  28. package/dist/market_adapter/inputs/window_cache.js +53 -17
  29. package/dist/market_adapter/lp_chart_core.js +1 -0
  30. package/dist/market_adapter/utils/chain.js +4 -3
  31. package/dist/market_adapter/utils/pair_orientation.d.ts +60 -0
  32. package/dist/market_adapter/utils/pair_orientation.js +142 -0
  33. package/dist/modules/account_bots.d.ts +37 -3
  34. package/dist/modules/account_bots.js +77 -37
  35. package/dist/modules/account_orders.d.ts +4 -1
  36. package/dist/modules/account_orders.js +4 -1
  37. package/dist/modules/bitshares-native/resolvers.js +3 -3
  38. package/dist/modules/bitshares_client.d.ts +12 -10
  39. package/dist/modules/bitshares_client.js +12 -10
  40. package/dist/modules/bot_settings.js +3 -3
  41. package/dist/modules/chain_keys.d.ts +5 -1
  42. package/dist/modules/chain_keys.js +5 -1
  43. package/dist/modules/chain_orders.d.ts +6 -2
  44. package/dist/modules/chain_orders.js +6 -2
  45. package/dist/modules/constants.d.ts +5 -0
  46. package/dist/modules/constants.js +7 -1
  47. package/dist/modules/dexbot_class.d.ts +3 -1
  48. package/dist/modules/dexbot_class.js +3 -1
  49. package/dist/modules/graceful_shutdown.d.ts +6 -2
  50. package/dist/modules/graceful_shutdown.js +6 -2
  51. package/dist/modules/launcher/bot_supervisor.js +4 -0
  52. package/dist/modules/launcher/launch_modes.js +6 -3
  53. package/dist/modules/launcher/monolithic_runtime.d.ts +39 -1
  54. package/dist/modules/launcher/monolithic_runtime.js +60 -1
  55. package/dist/modules/launcher/runtime_mode.d.ts +79 -0
  56. package/dist/modules/launcher/runtime_mode.js +223 -0
  57. package/dist/modules/market_adapter_whitelist.d.ts +8 -0
  58. package/dist/modules/market_adapter_whitelist.js +8 -0
  59. package/dist/modules/order/accounting.d.ts +15 -1
  60. package/dist/modules/order/accounting.js +15 -1
  61. package/dist/modules/order/async_lock.d.ts +9 -3
  62. package/dist/modules/order/async_lock.js +9 -3
  63. package/dist/modules/order/export.d.ts +0 -5
  64. package/dist/modules/order/export.js +0 -5
  65. package/dist/modules/order/format.d.ts +6 -4
  66. package/dist/modules/order/format.js +6 -4
  67. package/dist/modules/order/grid.d.ts +29 -18
  68. package/dist/modules/order/grid.js +29 -18
  69. package/dist/modules/order/index.d.ts +1 -1
  70. package/dist/modules/order/index.js +1 -1
  71. package/dist/modules/order/logger_state.d.ts +1 -1
  72. package/dist/modules/order/logger_state.js +1 -1
  73. package/dist/modules/order/sync_engine.d.ts +26 -11
  74. package/dist/modules/order/sync_engine.js +26 -11
  75. package/dist/modules/order/utils/math.d.ts +59 -21
  76. package/dist/modules/order/utils/math.js +59 -21
  77. package/dist/modules/order/utils/order.d.ts +64 -20
  78. package/dist/modules/order/utils/order.js +64 -20
  79. package/dist/modules/order/utils/system.d.ts +41 -11
  80. package/dist/modules/order/utils/system.js +41 -11
  81. package/dist/modules/order/utils/validate.d.ts +5 -2
  82. package/dist/modules/order/utils/validate.js +5 -2
  83. package/dist/modules/utils/asset_symbols.d.ts +5 -5
  84. package/dist/modules/utils/asset_symbols.js +5 -5
  85. package/dist/modules/utils/errors.d.ts +2 -1
  86. package/dist/modules/utils/errors.js +2 -1
  87. package/dist/scripts/rich_list.d.ts +32 -0
  88. package/dist/scripts/rich_list.js +16 -8
  89. package/dist/unlock.js +49 -18
  90. package/docs/BITSHARES_ONBOARDING.md +7 -0
  91. package/docs/COW_INVARIANTS.md +1 -1
  92. package/docs/DEXBOT_COMPARISON.md +5 -5
  93. package/docs/EVOLUTION.md +8 -7
  94. package/docs/FUND_MOVEMENT_AND_ACCOUNTING.md +3 -3
  95. package/docs/GRID_RECALCULATION.md +18 -6
  96. package/docs/LIFECYCLE.md +8 -8
  97. package/docs/ORDER_ENGINE_POST_1.0_RETROSPECTIVE.md +4 -4
  98. package/docs/README.md +1 -1
  99. package/docs/WORKFLOW.md +5 -5
  100. package/market_adapter/README.md +10 -3
  101. package/package.json +3 -1
  102. package/tests/README.md +29 -0
@@ -158,13 +158,19 @@ node dist/market_adapter/market_adapter.js --deltaPercent 2
158
158
  ```json
159
159
  {
160
160
  "ama": {
161
- "erPeriod": 10, // Efficiency Ratio lookback period (candles)
162
- "fastPeriod": 2, // Fast smoothing period for trending markets
163
- "slowPeriod": 30 // Slow smoothing period for choppy markets
161
+ "erPeriod": 781, // Efficiency Ratio lookback period (candles)
162
+ "fastPeriod": 5.2, // Fast smoothing period for trending markets
163
+ "slowPeriod": 83.6 // Slow smoothing period for choppy markets
164
164
  }
165
165
  }
166
166
  ```
167
167
 
168
+ > Optional. A bot-level `ama` block overrides the preset (or matched pair profile)
169
+ > field by field, and supplying any period marks the bot's AMA as `custom` in the
170
+ > adapter state. The values above are the fitted `AMA3` built-in defaults
171
+ > (`MARKET_ADAPTER.AMAS`): `erPeriod: 781` and `fastPeriod: 5.2` are shared, while
172
+ > `slowPeriod` runs `62.1` (AMA1) to `96.9` (AMA4).
173
+
168
174
  > **`gridPrice` is the only switch.** A bot is AMA-driven when `gridPrice` says
169
175
  > so — `usesAmaGridPrice()`, the source of truth shared by the bot, maintenance
170
176
  > and launcher runtimes. There is deliberately no `ama.enabled`: a second switch
@@ -199,9 +205,15 @@ The pool is fetched directly by ID, bypassing pool discovery. Useful when the tr
199
205
  - Higher = more lag, filters noise
200
206
  - Built-in presets range from `AMA1` (fastest) to `AMA4` (slowest) — see `MARKET_ADAPTER.AMAS` in `modules/constants.ts` for the exact `slowPeriod` values
201
207
 
202
- AMA parameters are resolved exclusively from **presets** (`AMA1`–`AMA4`): a matched
203
- pair profile in `profiles/market_profiles.json` wins, otherwise the built-in defaults
204
- (`MARKET_ADAPTER.AMAS` in `modules/constants.ts`, default key `AMA3`) are used.
208
+ AMA parameters resolve in this order: a matched pair profile in
209
+ `profiles/market_profiles.json` wins for that pair; otherwise the bot's own `ama`
210
+ block overrides the preset field by field; any period the bot does not supply is
211
+ taken from the preset named by `gridPrice` (`"ama"` → the pair's `defaultAma`,
212
+ which falls back to `AMA3`; `"ama1"`–`"ama4"` → that exact preset); and any
213
+ remaining missing field falls back to the built-in `MARKET_ADAPTER.AMAS` defaults
214
+ (`modules/constants.ts`, default key `AMA3`). A bot that hand-writes any of the
215
+ three periods is reported as `custom` in the adapter state rather than by a preset
216
+ name.
205
217
 
206
218
  ### How It Works
207
219
 
package/docs/LIFECYCLE.md CHANGED
@@ -90,7 +90,7 @@ sequenceDiagram
90
90
  participant CO as chain_orders (broadcast)
91
91
  participant STORE as Storage
92
92
 
93
- CH->>Q: enqueue fill(s) (dexbot_fill_runtime.ts:466)
93
+ CH->>Q: enqueue fill(s) (dexbot_fill_runtime.ts:463)
94
94
  FR->>FR: drain queue -> _processFillsWithBatching
95
95
  FR->>OM: processFilledOrders() (manager.ts)
96
96
  OM->>AC: processFillAccounting() (single call, batch)
@@ -134,9 +134,9 @@ sequenceDiagram
134
134
  participant STORE as Storage
135
135
 
136
136
  MA->>CL: AMA center snapshot updated
137
- CL->>CL: _performPeriodicGridChecks() (dexbot_class.ts:1720)
137
+ CL->>CL: _performPeriodicGridChecks() (dexbot_class.ts:1747)
138
138
  CL->>RT: performPeriodicGridChecks() (dexbot_maintenance_runtime.ts:1677)
139
- RT->>RT: runGridMaintenance(bot,'periodic') (dexbot_maintenance_runtime.ts:2953)
139
+ RT->>RT: runGridMaintenance(bot,'periodic') (dexbot_maintenance_runtime.ts:2981)
140
140
  RT->>RT: executeMaintenanceLogic() (dexbot_maintenance_runtime.ts:2462)
141
141
  RT->>AC: recalculate funds from balances
142
142
  RT->>GD: promote AMA center -> grid center
@@ -157,13 +157,13 @@ The AMA signal stack (AMA/Kalman/Hurst/PE) is *research-tuned* in `analysis/` an
157
157
  the parameters.
158
158
 
159
159
  > Note: `runMaintenance()` is a **different** subsystem — the credit/MPA debt
160
- > runtime (`modules/credit_runtime.ts:3441`, reached via
161
- > `_runCreditRuntimeMaintenance` at `dexbot_class.ts:1837`). The grid maintenance
160
+ > runtime (`modules/credit_runtime.ts:3623`, reached via
161
+ > `_runCreditRuntimeMaintenance` at `dexbot_class.ts:1864`). The grid maintenance
162
162
  > chain above is the one that matters for order/price upkeep.
163
163
 
164
- References: `modules/dexbot_class.ts:1720` (`_performPeriodicGridChecks`) →
164
+ References: `modules/dexbot_class.ts:1747` (`_performPeriodicGridChecks`) →
165
165
  `modules/dexbot_maintenance_runtime.ts:1677` (`performPeriodicGridChecks`) →
166
- `:2953` (`runGridMaintenance`) → `:2462` (`executeMaintenanceLogic`),
166
+ `:2981` (`runGridMaintenance`) → `:2462` (`executeMaintenanceLogic`),
167
167
  `docs/GRID_RECALCULATION.md`, `docs/GRID_RECONCILE.md`.
168
168
 
169
169
  ---
@@ -192,7 +192,7 @@ enforced*, not compiler-enforced — learn them or you will introduce fund bugs.
192
192
  | How a fill becomes orders | `modules/dexbot_fill_runtime.ts` → `modules/order/manager.ts` |
193
193
  | Grid math / recalculation | `modules/order/grid.ts`, `docs/GRID_RECALCULATION.md` |
194
194
  | Funds & accounting | `modules/order/accounting.ts`, `docs/FUND_MOVEMENT_AND_ACCOUNTING.md` |
195
- | Periodic loop & AMA hook | `modules/dexbot_maintenance_runtime.ts`, `modules/dexbot_class.ts:1720` |
195
+ | Periodic loop & AMA hook | `modules/dexbot_maintenance_runtime.ts`, `modules/dexbot_class.ts:1747` |
196
196
  | Market signal source | `market_adapter/market_adapter.ts`, `analysis/README.md` |
197
197
  | Startup & orchestration | `modules/dexbot_class.ts`, `docs/developer_guide.md` §"Startup Sequence" |
198
198
  | Why COW exists | `docs/architecture.md` §"Copy-on-Write (COW) Grid Pattern", `docs/COW_INVARIANTS.md` |
@@ -157,7 +157,7 @@ plans, `LANDED`/`REVERTED`/`SUPERSEDED` statuses, commit hashes, and verificatio
157
157
 
158
158
  ### 1. Gap-Band Orphan — Prevention Plan (P1–P6)
159
159
 
160
- > **Source:** `GAP_BAND_ORPHAN_PREVENTION_PLAN.md` · **Status:** implemented `f94d6ec4` (2026-08-29), with subsequent amendments — landing map below.
160
+ > **Source:** former `GAP_BAND_ORPHAN_PREVENTION_PLAN.md` (deleted 2026-08-31; merged here) · **Status:** implemented `f94d6ec4` (2026-08-29), with subsequent amendments — landing map below.
161
161
 
162
162
  **Landing map (corrected 2026-08-31):**
163
163
  - **P1** — landed `f94d6ec4` (same-batch CANCEL injection in `COWRebalanceEngine.execute` + `_assertGapBandIntactPostCommit` cancelOnly). **REVERTED `e2898e51`** ("remove placement guards that interfered with normal operation") — both sites removed; stranding is now handled by `calculateIdealBoundary` + `LAST-FILL-GUARD` + adoption path (P3 gate removed `e7231534`).
@@ -226,7 +226,7 @@ Sequence `Restored boundary index: 131` *then* `[SNAPSHOT-REJECT] Deleting corru
226
226
 
227
227
  ### 2. Ladder Recenter Orphan — Root Cause (Aug 26–30 2026)
228
228
 
229
- > **Source:** `LADDER_RECENTER_ORPHAN_ROOT_CAUSE.md` · **Status:** fixes #1–#7 landed 2026-08-30.
229
+ > **Source:** former `LADDER_RECENTER_ORPHAN_ROOT_CAUSE.md` (deleted 2026-08-31; merged here) · **Status:** fixes #1–#7 landed 2026-08-30.
230
230
 
231
231
  #### 2.1 Symptom
232
232
 
@@ -317,7 +317,7 @@ Throttle `Structural resync defer` to first/last per cap window (was 120/30s); l
317
317
 
318
318
  ### 3. Orphan-Fill & Fund-Invariant — Root Cause & Fix Plan
319
319
 
320
- > **Source:** `ORPHAN_FILL_INVARIANT_ROOT_CAUSE_AND_FIX.md` · **Status:** P0s + rotation landed `1b27f6eb`/`4838bcb0`/`d808c052`; P1 atomicity partial; duplicate-cancel hard-error and P2 counters still open.
320
+ > **Source:** former `ORPHAN_FILL_INVARIANT_ROOT_CAUSE_AND_FIX.md` (deleted 2026-08-31; merged here) · **Status:** P0s + rotation landed `1b27f6eb`/`4838bcb0`/`d808c052`; P1 atomicity partial; duplicate-cancel hard-error and P2 counters still open.
321
321
  > **Scope:** `dexbot_cow_runtime.ts`, `sync_engine.ts`, `chain_orders.ts`, `dexbot_state_recovery.ts`, `dexbot_maintenance_runtime.ts`, `strategy.ts`, `constants.ts` · **Evidence:** `profiles/logs/<bot>.log` Aug 28 17:40–17:57, Aug 29 15:13–15:19.
322
322
 
323
323
  #### 3.1 TL;DR — one cascade
@@ -388,7 +388,7 @@ All eight surface through `getMetrics()`; the five baseline counters via the exi
388
388
 
389
389
  ### 4. Price-First Alignment Plan
390
390
 
391
- > **Source:** `PRICE_FIRST_ALIGNMENT_PLAN.md` · **Status:** Phase 1 shipped and live as shadow telemetry; Phase 2 projection was enabled (`1bbf1a23`) then **removed** (`3713c496`) after a live-bot incident — the anchor veto overrode chain evidence and drove destructive gap-band cancels. The anchor is now observability-only; the boundary is chain-evidence-derived. Phase 3 is moot (no second write path to retire); boundary geometry is pinned by `tests/test_boundary_anchor_recovery.ts` (ANCHOR-001..004) and `tests/test_grid_price_slot_invariant.ts`.
391
+ > **Source:** former `PRICE_FIRST_ALIGNMENT_PLAN.md` (deleted 2026-08-31; merged here) · **Status:** Phase 1 shipped and live as shadow telemetry; Phase 2 projection was enabled (`1bbf1a23`) then **removed** (`3713c496`) after a live-bot incident — the anchor veto overrode chain evidence and drove destructive gap-band cancels. The anchor is now observability-only; the boundary is chain-evidence-derived. Phase 3 is moot (no second write path to retire); boundary geometry is pinned by `tests/test_boundary_anchor_recovery.ts` (ANCHOR-001..004) and `tests/test_grid_price_slot_invariant.ts`.
392
392
  > **Scope:** grid boundary state `modules/order/` · **Related:** three-layer crawl/anchor/guard fix.
393
393
 
394
394
  #### 4.1 Goal
package/docs/README.md CHANGED
@@ -2,7 +2,7 @@
2
2
 
3
3
  This directory contains the comprehensive technical documentation for the DEXBot2 trading bot. It is designed to guide developers from high-level architecture down to the nuances of fund accounting and state management.
4
4
 
5
- **Version context:** v1.6.15 (released).
5
+ **Version context:** v1.6.16 (released).
6
6
 
7
7
  ---
8
8
 
package/docs/WORKFLOW.md CHANGED
@@ -257,12 +257,12 @@ the canonical name is preferred in scripts and docs.
257
257
  | `dexbot dw <target>` | — | Dynamic-weight research chart: same targets/flags as `tv` |
258
258
  | `dexbot pnl <account>` | — | HTML PnL report for a bot/account over `--month N`, optional `--pair BASE/QUOTE` filter |
259
259
  | `dexbot git` | — | Analyze repository git history and write an added/deleted-lines HTML chart (`repo-stats.html`; source checkouts only) |
260
- | `dexbot status` | `stat`, `stats` | Unified runtime health — daemon, adapter, bots |
260
+ | `dexbot status` | `stat`, `stats` | Unified runtime health — daemon, adapter, bots. Whole-runtime: an optional bot name is validated, then ignored |
261
261
  | `dexbot start` | `unlock` | Run credential daemon + bot (equivalent to running the `unlock` runtime, `dist/unlock.js`) |
262
- | `dexbot stop` | `stp`, `stopall` | Stop the monolithic runtime (unlock mode) |
263
- | `dexbot reload` | `reloadall` | Reload the monolithic runtime without touching the credential daemon (unlock mode) |
264
- | `dexbot restart` | `restartall` | Restart the monolithic runtime (unlock mode, re-unlocks credential daemon) |
265
- | `dexbot delete` | — | Shut down and clean up the monolithic runtime (unlock mode) |
262
+ | `dexbot stop [<bot>]` | `stp`, `stopall` | Stop the runtime. Monolithic: whole runtime (a bot argument folds into it with a notice); `--isolated`: just that bot |
263
+ | `dexbot reload [<bot>]` | `reloadall` | Reload without touching the credential daemon. Same per-bot target rules as `stop` |
264
+ | `dexbot restart [<bot>]` | `restartall` | Restart and re-unlock the credential daemon. Same per-bot target rules as `stop` |
265
+ | `dexbot delete` | — | Shut down and clean up the monolithic runtime. Whole-runtime only; rejects a bot argument |
266
266
  | `dexbot clear` | — | Delete `*.log`, rotated `*.log.*` and `*.jsonl*` from the logs directory (`<profiles>/logs`) |
267
267
  | `dexbot clear-orders` | — | Delete all persisted order files (`<profiles>/orders`); grids regenerate on next start |
268
268
  | `dexbot clear-market-adapter` | — | Delete market adapter data, state (incl. `market_adapter.lock`), and adapter logs |
@@ -240,13 +240,18 @@ Pair-specific AMA profiles live in `profiles/market_profiles.json`:
240
240
  "intervalSeconds": 3600,
241
241
  "defaultAma": "AMA1",
242
242
  "amas": {
243
- "AMA1": { "erPeriod": 200, "fastPeriod": 3, "slowPeriod": 80 },
244
- "AMA2": { "erPeriod": 500, "fastPeriod": 5, "slowPeriod": 120 }
243
+ "AMA1": { "erPeriod": 781, "fastPeriod": 5.2, "slowPeriod": 62.1 },
244
+ "AMA2": { "erPeriod": 781, "fastPeriod": 5.2, "slowPeriod": 72.0 }
245
245
  }
246
246
  }]
247
247
  }
248
248
  ```
249
249
 
250
+ > The `amas` periods above are the fitted built-in presets (`MARKET_ADAPTER.AMAS`),
251
+ > shown here only to make the shape concrete. A real pair profile needs just the
252
+ > fields it changes — e.g. `"AMA3": { "slowPeriod": 90 }` keeps the built-in
253
+ > `erPeriod`/`fastPeriod` and overrides only `slowPeriod`.
254
+
250
255
  `profiles/market_adapter_settings.json` layers:
251
256
 
252
257
  ```json
@@ -580,6 +585,7 @@ Typical fields:
580
585
  market_adapter/
581
586
  |-- market_adapter.ts main adapter daemon
582
587
  |-- ama_signal_runner.ts one-cycle JSON signal CLI
588
+ |-- cli_args.ts shared CLI argument walker
583
589
  |-- candle_utils.ts candle transforms, gap detection, pruning
584
590
  |-- interval_utils.ts shared interval label + chain-time helpers
585
591
  |-- lp_chart_core.ts chart HTML renderer
@@ -623,6 +629,7 @@ market_adapter/
623
629
  | |-- file_lock.ts single-instance file lock
624
630
  | |-- file_json_cache.ts write-through JSON read cache
625
631
  | |-- month_shards.ts calendar-month shard naming + coverage-span primitives (shared by candle and fills caches)
632
+ | |-- pair_orientation.ts canonical pair orientation + reciprocal inversion helpers
626
633
  | |-- data_discovery.ts data directory auto-discovery + shard-family assembly
627
634
  | |-- atomic_write.ts atomic file write utility
628
635
  | |-- dynamic_grid_snapshot.ts dynamic grid snapshot helpers
@@ -863,7 +870,7 @@ closed 1h candles.
863
870
 
864
871
  #### Shared Month-Shard Cache and Fetch Robustness
865
872
 
866
- Pool, book, and feed candle fetches share one cache entry point (`runCachedWindows` in `market_adapter/inputs/window_cache.ts`): candles live in fixed calendar-month shards (`<base>.shard_YYYY-MM.json`, UTC) whose names never shift, so a run loads only the shards overlapping its requested range, queries only genuinely missing buckets plus a bounded tail refresh, and rewrites only shards that gained buckets or query coverage — pure-reuse runs perform zero writes and zero deletes. Shard metas record the ranges actually queried (`meta.queriedRanges`, monotonically unioned) together with `at`, the time each query ran; spans merge only when they assert the same verification time, so per-bucket freshness stays exact. A missing range is pruned only when recorded query coverage genuinely covers it — the absence of local buckets alone never certifies history as empty — and only once the range ended more than `GAP_SETTLE_HORIZON_MS` ago, judged per gap rather than per window (a month-old gap at the leading edge of the newest window is as settled as one in a fully past window). The tail window re-queries only buckets that no query has seen `TAIL_SETTLE_LAG_MS` after they closed (`settleCoverage`), so a rerun costs ~lag+elapsed hours instead of a fixed 48h, and the refresh never reaches back past `TAIL_REFRESH_HOURS` (an unsettled bucket must be refreshed because index lag is still open, but a coverage hole older than 48h is past any plausible lag and must not drag the boundary back weeks); caches without `at` fall back to that fixed window. Clusters of gaps within `GAP_MERGE_TOLERANCE_BUCKETS` collapse into one query, and the sub-range path is bounded by merged span hours rather than gap count, so one extra small hole can no longer escalate a run to a full-window fetch. Because spans merge only on an identical `at`, a shard gains one span per re-verifying run; the persisted list is capped at `MAX_COVERAGE_SPANS` by dropping the oldest spans, which can never invent emptiness (it can only make those buckets look unsettled, bounded by the same 48h ceiling). A run that re-verifies the tail therefore rewrites that shard even when the candles come back identical — the moved verification time is the news, and skipping the write would throw it away; pure-reuse runs still write nothing. Partial windows merge into the run output but are never persisted. Stable month shards are the only supported cache format; obsolete run-relative cache files are ignored. Every range fetch runs through `fetchRangeWithRetry` (per-range attempts + linear backoff + abort-signal timeout; the LP path keeps a 4-attempt budget), one-shot Kibana queries retry transient errors (3 attempts), paged fetchers cap at `kibanaMaxPages` (500), and bidirectional fetches tolerate a one-direction failure.
873
+ Pool, book, and feed candle fetches share one cache entry point (`runCachedWindows` in `market_adapter/inputs/window_cache.ts`): candles live in fixed calendar-month shards (`<base>.shard_YYYY-MM.json`, UTC) whose names never shift, so a run loads only the shards overlapping its requested range, queries only genuinely missing buckets plus a bounded tail refresh, and rewrites only shards that gained buckets or query coverage — pure-reuse runs perform zero writes and zero deletes. Shard metas record the ranges actually queried (`meta.queriedRanges`, monotonically unioned) together with `at`, the time each query ran; spans merge only when they assert the same verification time, so per-bucket freshness stays exact. A missing range is pruned only when recorded query coverage genuinely covers it — the absence of local buckets alone never certifies history as empty — and only once the range ended more than `GAP_SETTLE_HORIZON_MS` ago, judged per gap rather than per window (a month-old gap at the leading edge of the newest window is as settled as one in a fully past window). The tail window re-queries only buckets that no query has seen `TAIL_SETTLE_LAG_MS` after they closed (`settleCoverage`), so a rerun costs ~lag+elapsed hours instead of a fixed 48h, and the refresh never reaches back past `TAIL_REFRESH_HOURS` (an unsettled bucket must be refreshed because index lag is still open, but a coverage hole older than 48h is past any plausible lag and must not drag the boundary back weeks); caches without `at` fall back to that fixed window. Clusters of gaps within `GAP_MERGE_TOLERANCE_BUCKETS` collapse into one query, and the sub-range path is bounded by merged span hours rather than gap count, so one extra small hole can no longer escalate a run to a full-window fetch. Because spans merge only on an identical `at`, a shard gains one span per re-verifying run; the persisted list is capped at `MAX_COVERAGE_SPANS` by dropping the oldest spans, which can never invent emptiness (it can only make those buckets look unsettled, bounded by the same 48h ceiling). A run that re-verifies the tail therefore rewrites that shard even when the candles come back identical — the moved verification time is the news, and skipping the write would throw it away; pure-reuse runs still write nothing. Partial windows merge into the run output but are never persisted. Stable month shards are the only supported cache format; obsolete run-relative cache files are ignored. Pair orientation is canonicalized (the lower asset id is always leg A on disk), so `bts/iob.xrp` and `iob.xrp/bts` share one shard family: the fetch queries the canonical side and the result is inverted back for the requested side (inverse volume uses the same base×close approximation as the TV chart's inverse view). A legacy inverse-orientation family is read once as a migration source and folded into the canonical shards instead of being re-fetched. Every range fetch runs through `fetchRangeWithRetry` (per-range attempts + linear backoff + abort-signal timeout; the LP path keeps a 4-attempt budget), one-shot Kibana queries retry transient errors (3 attempts), paged fetchers cap at `kibanaMaxPages` (500), and bidirectional fetches tolerate a one-direction failure.
867
874
 
868
875
  #### AMA Warmup Window — Why Candle Length Matters
869
876
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "dexbot",
3
- "version": "1.6.15",
3
+ "version": "1.6.16",
4
4
  "description": "The first open source trading bot with zero runtime dependencies and a fully adaptive market making strategy.",
5
5
  "main": "dist/modules/dexbot_class.js",
6
6
  "exports": {
@@ -109,6 +109,7 @@
109
109
  "verify:browser-bundle": "npm run build && node dist/scripts/verify-browser-bundle.js",
110
110
  "test": "npm run build && npm run build:tests && npm run build --prefix claw && node dist/scripts/run-tests.js",
111
111
  "test:live": "RUN_LIVE_BITSHARES_TESTS=1 npm run build && npm run build:tests && npm run build --prefix claw && node dist/scripts/run-tests.js",
112
+ "test:playwright": "npm run build && NODE_PATH=\"$(npm root -g)\" playwright test --config tests/playwright/playwright.config.mjs",
112
113
  "clean": "node scripts/clean-dist.js",
113
114
  "build": "echo \"tsc is compiling ...\" && tsc && chmod +x dist/dexbot.js dist/bot.js dist/unlock.js dist/pm2.js dist/credential-daemon.js",
114
115
  "build:clean": "npm run clean && npm run build",
@@ -157,6 +158,7 @@
157
158
  "./dist/modules/launcher/status_reporting.js": false,
158
159
  "./dist/modules/launcher/supervisor_control.js": false,
159
160
  "./dist/modules/launcher/monolithic_runtime.js": false,
161
+ "./dist/modules/launcher/runtime_mode.js": false,
160
162
  "./dist/modules/launcher/market_adapter_runtime.js": false,
161
163
  "./dist/modules/launcher/market_adapter_watchdog.js": false,
162
164
  "./dist/modules/launcher/foreign_cred_daemon.js": false,
package/tests/README.md CHANGED
@@ -108,6 +108,35 @@ PM2 lifecycle, startup ordering, bot supervision, launcher version reporting.
108
108
  Interactive tools and performance benchmarks (not part of CI).
109
109
  *Examples:* `connection_test.ts`, `diag_adapter_client.ts`, `diag_ws_nodes.ts`, `diag_ws_lifecycle.ts`, `benchmark_cow.ts`, `sim_batching.ts`, `repro_phantom_orders.ts`
110
110
 
111
+ ### Browser Charts (Playwright)
112
+ End-to-end tests for the generated HTML charts (currently the `dexbot tv`
113
+ TradingView exporter). These load a real chart page in Firefox and assert on
114
+ canvas draw work and pixels, so they live under `tests/playwright/` as
115
+ `.spec.mjs` files and are **not** picked up by the plain `node` runner.
116
+
117
+ ```bash
118
+ npm run test:playwright # builds, then runs tests/playwright
119
+ ```
120
+
121
+ The runner and browsers come from a global Playwright install (the repo keeps
122
+ zero runtime deps); `test:playwright` points `NODE_PATH` at `npm root -g`, and
123
+ the spec resolves `playwright/test` dynamically. That inline `NODE_PATH=` is
124
+ POSIX-shell syntax (Linux/macOS); on Windows set it in your shell
125
+ (`$env:NODE_PATH=(npm root -g)`) or add a local `playwright` devDependency.
126
+ **Firefox is the default target** (set in
127
+ `tests/playwright/playwright.config.mjs`) because that is the browser the
128
+ generated charts are hand-checked in. If Playwright or the Firefox build is
129
+ missing: `npm i -g playwright && npx playwright install firefox`.
130
+
131
+ `tv_chart_perf.spec.mjs` pins the range-band draw hook to the visible bar
132
+ window: it counts `CanvasRenderingContext2D.lineTo` calls to prove the band's
133
+ per-redraw cost scales with the viewport (not the full series), and diffs
134
+ canvas pixels to prove the band still spans the plot when zoomed in. By default
135
+ it generates a synthetic TV page; set `TV_CHART_HTML=<path>` to run the load
136
+ check against any already-generated chart HTML (the uPlot interception is
137
+ generic; the Range-band assertions are skipped when the page has no Range
138
+ toggle).
139
+
111
140
  ### Edge Cases & Regression
112
141
  Tests targeting specific bugs, race conditions, and failure modes.
113
142
  *Examples:* `test_critical_bug_fixes.ts`, `test_race_condition_fixes_batch1.ts`, `test_patch17_invariants.ts`, `test_shutdown_reentrancy.ts`, `test_multifill_opposite_partial.ts`, `test_correction_queue_staleness.ts`, `test_spread_pure_fund_driven.ts`, `test_spread_correction_resilience.ts`, `test_sync_lock_id_verification.ts`