dexbot 1.1.10

package/claw/skills/margin-trading/references/position-management.md

@@ -0,0 +1,203 @@
+# Position Management
+
+## Collateral Ratio Zones
+
+See `claw/docs/POSITION_HEALTH.md` for the full 3-zone CR model, dual-layer action system, and trend alignment logic.
+
+The skill-level summary:
+
+| Zone | CR Range | Status |
+|---|---|---|
+| Red (high) | Above `CR_ZONES.RED_HIGH` (3.0) | Over-collateralized |
+| Green | `CR_ZONES.RED_LOW` (1.7) – `CR_ZONES.RED_HIGH` (3.0) | Safe — operating target |
+| Red (low) | Below `CR_ZONES.RED_LOW` (1.7) | Not acceptable |
+
+MCR for HONEST.Assets is 1.4. See `honest-assets.md` for full ecosystem properties.
+
+Any position outside the green zone triggers two-layer adjustment (debt first, collateral second). See POSITION_HEALTH.md for full action tables.
+
+## Capital Efficiency Principle
+
+The goal is to maximize the BTS that is actively working — either as collateral backing a position or as funds on the order book generating volume — while keeping every open position in the green zone.
+
+BTS sitting idle in a wallet earns nothing. BTS locked as excess collateral beyond what the green zone requires is also underutilized. The balance:
+
+- Enough collateral to stay green (CR >= 1.7, with 2.0+ as a prudent operating target)
+- Remaining BTS deployed as order book liquidity or available for new positions
+
+## Maximizing Volume
+
+Volume comes from orders being filled. More fills means more volume. What drives fills:
+
+- **Orders on the book** — funds not on the book produce zero volume
+- **Maker placement** — 0% maker fee on HONEST.Assets means resting orders cost nothing in market fees
+- **Both sides** — having both buy and sell orders on the book means fills happen regardless of which direction the market moves
+- **Position recycling** — proceeds from a filled order can be redeployed into a new order
+
+Idle BTS or idle MPA holdings not placed as orders do not contribute to volume.
+
+## Being on the Right Side
+
+A position profits when the market moves in the direction the position is exposed to. The two directions:
+
+- **Short** profits when the MPA price (in BTS) decreases — the MPA bought back cheaper than it was sold
+- **Long** profits when the MPA price (in BTS) increases — the MPA sold higher than it was bought
+
+The shared trend-detection service provides signals based on AMA deviation from the feed price and the instantaneous premium/discount. These signals indicate which direction the smoothed market is moving relative to the feed anchor.
+
+Combining position direction with the trend signal:
+
+| Trend Signal | Position Aligned | Position Opposed |
+|---|---|---|
+| UP | Long | Short |
+| DOWN | Short | Long |
+| NEUTRAL | Either (no directional edge) | Either |
+
+Being on the right side means the position direction matches the detected trend. When no trend is confirmed, there is no directional edge — both sides are equivalent.
+
+## Shared Trend Service
+
+Trend detection, signal monitoring, and signal-driven bot-setting updates are owned by the shared trend-detection skill.
+
+- Use `../trend-detection/SKILL.md` for the trend model, tuning workflow, and bot-setting update boundary.
+- Keep this file focused on position health, CR discipline, and how margin positions consume the adaptive settings.
+
+## Control Split
+
+The bot should be treated as two layers:
+
+- a fixed structural grid layer
+- an adaptive trading layer on top of AMA
+
+### Fixed Structural Settings
+
+These should stay static unless you intentionally retune the bot for a different market regime:
+
+- `incrementPercent`
+- `targetSpreadPercent`
+- `activeOrders`
+
+These settings define the ladder geometry:
+
+- `incrementPercent` sets slot spacing
+- `targetSpreadPercent` sets the empty center buffer
+- spread is coupled to increment in the engine, so changing it changes slot count, order sizes, and gap structure
+
+Because of that, increment/spread are not good tactical knobs. They are setup choices.
+
+### Adaptive Trading Settings
+
+These are the practical knobs for building an advanced margin trader on top of the AMA price adapter:
+
+- debt adjustment
+- collateral adjustment
+- `weightDistribution`
+- `minPrice` / `maxPrice` ratio
+
+These settings change behavior without redefining the whole ladder:
+
+- CR is repaired through debt first, collateral second
+- AMA remains the base anchor
+- `weightDistribution` changes where size is concentrated within the existing ladder
+- the min/max ratio changes the outer operating envelope slowly based on former price action
+
+Practical range-ratio bands:
+
+- below `2x` = very competitive
+- around `2x` = competitive
+- around `3x` = conservative
+- above `3x` = very conservative
+
+### Unified Plan
+
+The functions in `claw/modules/position_health.ts` (`assessPosition()`, `computePriceRangeRatioPlan()`, `computeOrderWeightBias()`) should therefore produce a unified plan with:
+
+- position assessment
+- target CR resolution
+- debt/collateral action plan
+- final `weightDistribution` bias
+- final min/max price ratio recommendation
+- a structural reset requirement when the CR plan changes debt or collateral
+
+This keeps the bot architecture clean:
+
+- AMA = anchor
+- weights = deployment bias within the grid
+- range ratio = slow structural width
+- debt/collateral actions = margin-risk control
+- CR execution is not the market-adapter's responsibility; the debt runtime applies the change and then rebuilds the grid from the new capital base
+
+Example: weak short, strong DOWN trend
+
+- repair the weak short by reducing debt first
+- front-load buys and flatten sells
+
+Example: sideways market
+
+- `weightDistribution` stays balanced or double-mountain
+- the bot keeps full deployment, centered by AMA, with no strong directional skew beyond the configured price bounds
+
+Interpretation:
+
+- CR below target → debt reduction first
+- if a CR leg executes, the bot must rebuild the grid before reusing fund assumptions
+- buys are front-loaded when they are the with-trend side
+- sells are flattened when price is moving away from them
+
+## Loss Minimization
+
+Losses come from:
+
+1. **Adverse price movement** — position is on the wrong side
+2. **Margin call** — CR drops below MCR, protocol force-covers at unfavorable price
+3. **Excess fees** — taker fills when maker placement was possible
+4. **Idle capital** — opportunity cost of funds not deployed
+
+Controls:
+
+- **CR zones** enforce collateral discipline — green zone positions have a buffer that absorbs adverse moves without reaching liquidation
+- **Maker-first placement** eliminates the largest recurring fee (0% vs 0.1%/0.2%)
+- **Trend alignment** reduces the probability of being on the wrong side
+- **Position sizing** relative to available capital — no single position should consume all available BTS as collateral
+
+## Profit Maximization
+
+Profits come from:
+
+1. **Spread capture** — buying at bid, selling at ask
+2. **Directional gain** — position on the right side of a trend
+3. **Fee asymmetry** — earning the spread while paying 0% maker fee
+4. **Capital velocity** — the faster proceeds are redeployed, the more cycles of profit per unit of time
+
+The lever is utilization: BTS that is working — on the book or backing a position — has the opportunity to earn. BTS that is idle does not. Maximizing available funds on the book and maintaining position alignment maximizes the rate at which profit accumulates.
+
+## Fund Flow
+
+```
+Total BTS
+├── Collateral (locked in debt positions)
+│   └── Target: CR in green zone (2.0 – 2.5)
+├── On book (resting limit orders)
+│   └── Target: as much as possible
+├── Pending (proceeds from fills, not yet redeployed)
+│   └── Target: minimize time in this state
+└── Reserve (fee budget for operations)
+    └── ~0.1 BTS per order operation
+```
+
+MPA holdings follow the same logic — MPA sitting in wallet balance is idle. MPA placed as a sell order or used to repay debt is working.
+
+## Decision Loop
+
+The evaluation pipeline runs as a cycle:
+
+```
+discover → trend → assess → recommend
+```
+
+1. **Discover** (`position_discovery.ts`) — scan account's on-chain call orders to find all open debt positions
+2. **Trend** (`feed_price_source.ts`) — fetch feed price and market mid-price per market, update the trend detector (KAMA-based, internal to the price source module)
+3. **Assess** (`position_health.ts`) — classify each position into a CR zone, check trend alignment, generate prioritized actions
+4. **Recommend** (`decision_loop.ts`) — sort actions by priority (immediate → soon → evaluate → fallback), return structured assessment
+
+The loop evaluates and recommends. Execution of recommended actions is a separate concern.

package/claw/skills/margin-trading/references/trading-concepts.md

@@ -0,0 +1,80 @@
+# Trading Concepts
+
+## Positions
+
+### Short Position
+
+Betting that the MPA price (in BTS terms) will decrease.
+
+**Lifecycle:**
+
+1. **Borrow** — Issue MPA debt against BTS collateral (`call_order_update` with positive debt and collateral deltas). Collateral ratio must be in the green zone (CR >= 1.7).
+2. **Sell** — Place a limit order selling the borrowed MPA for BTS. As a maker order this incurs 0% fee on HONEST.Assets.
+3. **Cover** — Later, buy back the MPA at a lower price to repay the debt.
+4. **Close** — Repay the debt (`call_order_update` with negative debt delta) and release the collateral.
+
+Profit comes from the difference between the sell price and the buyback price, denominated in BTS.
+
+**Available operations (claw bridge):**
+
+| Step | Command | Broadcast |
+|---|---|---|
+| Plan open | `build-open-short-plan` | No |
+| Execute open | `open-short-bts` | Yes — borrows + places sell order |
+| Plan take-profit | `build-take-profit-plan` | No |
+| Execute take-profit | `take-profit-bts` | Yes — places buy order to cover |
+| Plan close | `build-close-short-plan` | No |
+| Execute close | `close-short-bts` | Yes — repays debt, releases collateral |
+
+### Long Position
+
+Betting that the MPA price (in BTS terms) will increase.
+
+**Lifecycle:**
+
+1. **Buy** — Place a limit order buying the MPA with BTS. As a maker order this incurs 0% fee on HONEST.Assets.
+2. **Hold** — Wait for price to move in your favor.
+3. **Sell** — Sell the MPA back for BTS at a higher price.
+
+No borrowing or collateral is required for a long position — it is a spot buy.
+
+**Available operations (claw bridge):**
+
+| Step | Command | Broadcast |
+|---|---|---|
+| Buy | `create-limit-order` | Yes |
+| Sell | `create-limit-order` | Yes |
+| Modify | `update-limit-order` | Yes |
+| Cancel | `cancel-limit-order` | Yes |
+
+## Collateral Management
+
+An open debt position can be adjusted without closing it:
+
+| Action | Command | Effect |
+|---|---|---|
+| Add collateral | `adjust-mpa-collateral` (positive delta) | Increases CR, reduces liquidation risk |
+| Remove collateral | `adjust-mpa-collateral` (negative delta) | Decreases CR, frees BTS |
+| Increase debt | `borrow-mpa` | Issues more MPA, decreases CR |
+| Repay partial | `repay-mpa` | Reduces debt, increases CR |
+
+Adjustments change the collateral ratio — the position must remain in the green zone (CR >= 1.7). See `references/position-management.md` for CR zones.
+
+## Position Queries
+
+| Query | Command | Returns |
+|---|---|---|
+| On-chain debt position | `mpa-position` | Debt amount, collateral, CR, feed price |
+| Account balances | `account-snapshot` | All balances and open orders |
+| Open limit orders | `open-orders` | Active orders with amounts and prices |
+| Market state | `market-snapshot` | Order book, ticker, global properties |
+
+## Settlement and Margin Calls
+
+If the collateral ratio drops below MCR (1.4 for HONEST.Assets), the position becomes subject to a margin call. The protocol will match the position against limit orders to cover the debt.
+
+Force settlement allows any holder to settle their MPA at the feed price, subject to settlement delay and offset parameters configured by the asset issuer.
+
+## Batch Operations
+
+Multiple operations can be combined into a single atomic transaction using `execute-batch`. This is useful for simultaneously opening a position and placing a take-profit order, or adjusting multiple positions at once.

package/claw/skills/memu-memory/SKILL.md

@@ -0,0 +1,218 @@
+# memU Proactive Memory for DEXBot2
+
+memU provides 24/7 always-on proactive memory for AI agents integrated with DEXBot2. It captures user intent, reduces LLM token costs, and enables context-aware trading assistance.
+
+## Overview
+
+memU treats memory like a file system — structured, hierarchical, and instantly accessible:
+
+| File System | memU Memory |
+|-------------|-------------|
+| Folders | Categories (auto-organized topics) |
+| Files | Memory Items (extracted facts, preferences, skills) |
+| Symlinks | Cross-references (related memories linked) |
+| Mount points | Resources (conversations, documents, images) |
+
+## Memory Hierarchy
+
+```
+memory/
+├── preferences/
+│   ├── communication_style.md
+│   └── trading_preferences.md
+├── relationships/
+│   ├── contacts/
+│   └── interaction_history/
+├── knowledge/
+│   ├── domain_expertise/
+│   └── trading_strategies/
+└── context/
+    ├── recent_conversations/
+    └── pending_tasks/
+```
+
+## Prerequisites
+
+- Python 3.13+
+- memU package: `pip install memu-py`
+- LLM API key (OpenAI, OpenRouter, etc.)
+
+## Available Tools
+
+### Core Memory Operations
+
+| Tool | Description | Required Args |
+|------|-------------|---------------|
+| `memu_memorize` | Store a resource as memory | `resourceUrl`, `modality` |
+| `memu_retrieve` | Query stored memories | `queries` |
+| `memu_list_categories` | List memory categories | none |
+| `memu_list_items` | List memory items | none |
+| `memu_create_item` | Create a memory item directly | `categoryId` or `categoryName`, `summary` |
+| `memu_status` | Get memU service status | none |
+
+### Trading-Specific Operations
+
+| Tool | Description | Required Args |
+|------|-------------|---------------|
+| `memu_memorize_conversation` | Memorize a conversation | `messages` |
+| `memu_memorize_trading_context` | Memorize trading context | `context` |
+| `memu_retrieve_trading_context` | Retrieve trading memories | `query` |
+
+## Modalities
+
+| Modality | Use Case |
+|----------|----------|
+| `conversation` | Chat logs, user-bot interactions |
+| `document` | Trading reports, market analysis, bot configs |
+| `image` | Chart screenshots, price graphs |
+| `video` | Trading tutorials, market commentary |
+| `audio` | Voice notes, trading calls |
+
+## Usage Examples
+
+### Memorize a Conversation
+
+```json
+{
+  "tool": "memu_memorize_conversation",
+  "arguments": {
+    "messages": [
+      {"role": "user", "content": "I prefer BTS/USD grid bots with 2% increment"},
+      {"role": "assistant", "content": "I'll configure a grid bot with those settings"}
+    ],
+    "user": {"user_id": "trader-123"}
+  }
+}
+```
+
+### Memorize Trading Context
+
+```json
+{
+  "tool": "memu_memorize_trading_context",
+  "arguments": {
+    "context": {
+      "bot": "BTS/USD-grid",
+      "event": "price_dropped_5_percent",
+      "action_taken": "rebalanced_grid",
+      "timestamp": "2026-05-18T10:30:00Z"
+    },
+    "user": {"user_id": "trader-123"}
+  }
+}
+```
+
+### Retrieve Trading Context
+
+```json
+{
+  "tool": "memu_retrieve_trading_context",
+  "arguments": {
+    "query": "What are my preferences for BTS/USD grid bots?",
+    "user": {"user_id": "trader-123"}
+  }
+}
+```
+
+### Retrieve with LLM Deep Reasoning
+
+```json
+{
+  "tool": "memu_retrieve",
+  "arguments": {
+    "queries": [
+      {"role": "user", "content": {"text": "How should I adjust my grid based on recent market behavior?"}}
+    ],
+    "method": "llm",
+    "where": {"user_id": "trader-123"}
+  }
+}
+```
+
+## Integration with DEXBot2 Claw
+
+The memU bridge integrates with the DEXBot2 claw subsystem:
+
+### CLI Usage
+
+```bash
+# From claw/ directory
+npm run memu:status
+npm run memu:mcp  # Start MCP server
+```
+
+### MCP Server Configuration
+
+For Hermes:
+
+```yaml
+mcp_servers:
+  memu:
+    command: "tsx"
+    args: ["/path/to/claw/scripts/memu_mcp_server.ts", "--memu-dir", "/path/to/claw/data/memu"]
+```
+
+For NanoBot/PicoClaw:
+
+```bash
+tsx scripts/memu_mcp_server.ts --memu-dir /path/to/claw/data/memu
+```
+
+## Proactive Memory Patterns
+
+### Pattern 1: Learning User Preferences
+
+When the user mentions trading preferences:
+
+1. Extract the preference from the conversation
+2. Call `memu_memorize_conversation` to store it
+3. On future interactions, call `memu_retrieve_trading_context` to recall preferences
+
+### Pattern 2: Trading Event Memory
+
+When significant trading events occur:
+
+1. Call `memu_memorize_trading_context` with event details
+2. Store bot actions, market conditions, and outcomes
+3. Later, retrieve to inform similar decisions
+
+### Pattern 3: Context-Aware Assistance
+
+Before responding to trading queries:
+
+1. Call `memu_retrieve_trading_context` with the query
+2. Use retrieved context to provide personalized responses
+3. Reference past decisions and preferences
+
+## Memory Types
+
+| Type | Description |
+|------|-------------|
+| `profile` | User preferences and settings |
+| `knowledge` | Factual information about markets, assets |
+| `skill` | Learned trading strategies and techniques |
+| `behavior` | Interaction patterns and habits |
+| `event` | Specific trading events and outcomes |
+| `tool` | Tool call memory and outcomes |
+
+## Best Practices
+
+1. **Scope memories to users**: Always pass `user` with `user_id` when multiple users share the system
+2. **Use appropriate modalities**: Choose the right modality for the content type
+3. **Retrieve before acting**: Check existing memories before making recommendations
+4. **Use RAG for speed, LLM for depth**: `method: "rag"` is fast, `method: "llm"` provides deeper reasoning
+5. **Clean up old memories carefully**: Prefer scoped `memu_clear` calls such as `{"where":{"user_id":"trader-123"}}` on shared systems
+
+## Environment Variables
+
+| Variable | Description |
+|----------|-------------|
+| `OPENAI_API_KEY` | OpenAI API key for LLM operations |
+| `MEMU_PYTHON` | Path to Python interpreter (default: `python3`) |
+
+## Notes
+
+- memU runs as a Python subprocess bridge from Node.js
+- Memory is persisted to SQLite under `claw/data/memu` by default
+- For production, configure PostgreSQL with pgvector for persistent storage
+- The MCP server uses stdio transport with newline-delimited JSON-RPC

package/claw/skills/shared/references/js-automation-overview.md

@@ -0,0 +1,24 @@
+# BitShares JS Automation Overview
+
+Use this reference when the goal is long-running JavaScript automation against BitShares.
+
+Focus on the practical module split:
+
+- one shared read/subscription client
+- one per-account signing client
+- separate query, action, and state layers
+- local-only signing through the credential path the runtime provides
+
+Keep the implementation reference explicit:
+
+- DEXBot2 shows a real Node.js/CommonJS bot runtime with shared reads, per-account signing, persistence, and reconciliation
+- AI-Bot should reuse that pattern rather than duplicating it
+
+Preferred baseline:
+
+- `Node.js` / CommonJS first
+- DEXBot2 native BitShares client for chain connectivity
+- shared read connection for market and account state
+- dedicated account client for signing and broadcast
+
+Use this as a practical implementation guide, not as protocol authority.

package/claw/skills/shared/references/safety-and-staleness.md

@@ -0,0 +1,11 @@
+# Safety and Staleness Labels
+
+- `current`: primary source surfaces for the current release, when present in the repo.
+- `reference-only`: presentation reference material and any presentation examples derived from it.
+- `historical-only`: legacy archive material or older reference material that should not be treated as current guidance.
+
+Rules:
+- Treat `current` files as the only place to make operational claims.
+- Treat `reference-only` material as layout, tone, and framing guidance only.
+- Treat `historical-only` material as stale until explicitly relabeled by the maintainer.
+- If a file mixes current and historical material, split or relabel it before reuse.

package/claw/skills/shared/references/skill-boundaries.md

@@ -0,0 +1,17 @@
+# Skill Boundaries
+
+| Skill | Boundary | Allowed work | Not allowed |
+| --- | --- | --- | --- |
+| `bitshares-guide` | presentation-only | Pack presentation, onboarding copy, and guide structure | Operational instructions or source dumps |
+| `trend-detection` | service-control-plane | Candle-driven signal monitoring, parameter tuning, and bot-setting updates | Order placement or margin-trading policy |
+| `margin-trading` | concept-reference | Trading concepts, HONEST.Asset properties, position lifecycle, available operations | Prescribing specific parameter values as optimal, strategy recommendations, profit/loss predictions |
+| `launcher-ops` | operational-orchestration | PM2 startup, unlock, claw-only mode, Docker-friendly launcher flow, launcher validation | Trading advice, credential disclosure, or bridge internals |
+
+## Non-Goals
+
+- No secrets or credentials.
+- No unsupported runtime claims.
+- No duplicated canonical source pointers inside individual `SKILL.md` files.
+- No mixing of current, reference-only, and historical-only material without labels.
+- Keep launcher orchestration guidance isolated to `launcher-ops`.
+- No reuse of presentation reference material as an operational source.

package/claw/skills/trend-detection/SKILL.md

@@ -0,0 +1,33 @@
+---
+name: trend-detection
+description: Shared DEXBot/Claw trend-detection service for candle-driven signals, parameter tuning, and bot-setting updates.
+---
+
+# Trend Detection
+
+Use this skill when the task is about the shared trend service rather than margin-trading specifically.
+
+This skill covers:
+
+- candle-fed trend detection and signal monitoring
+- tuning trend-detection parameters and general settings
+- applying signal-driven bot-setting updates for `weightDistribution`
+
+## Workflow
+
+1. Inspect the current trend signal, confidence, and premium/discount state.
+2. Check the candle source and the active bot context.
+3. Adjust trend-detection parameters in the appropriate configuration surface.
+4. Preview the bot-settings patch before applying it.
+5. Apply the update only when the signal and validation agree.
+6. Monitor the next cycle through the adapter state and recalc trigger files.
+
+## Boundaries
+
+- Keep signal generation in `analysis/trend_detection/` and `market_adapter/`.
+- Keep setting updates in the Claw/Dexbot profile layer.
+- Do not tie the service to margin-trading-only policy or order placement.
+
+## Reference
+
+See [references/service.md](references/service.md) for the signal model, control surfaces, and tuning boundaries.

package/claw/skills/trend-detection/agents/openai.yaml

@@ -0,0 +1,3 @@
+interface:
+  display_name: "Trend Detection"
+  short_description: "Tune and monitor the shared DEXBot/Claw trend service"

package/claw/skills/trend-detection/references/service.md

@@ -0,0 +1,45 @@
+# Trend Detection Service
+
+## Scope
+
+This is the shared DEXBot/Claw trend service. It is independent of margin-trading and should be used for any bot that needs candle-driven signal generation, parameter tuning, and bot-setting updates.
+
+## Signal Flow
+
+1. `market_adapter/market_adapter.ts` keeps candles current and emits the per-bot center/state snapshot.
+2. `analysis/trend_detection/kalman_trend_analyzer.ts` turns candle input plus feed price into trend output.
+3. The service monitors the signal output, premium/discount, and price-ratio context.
+4. Claw applies supported setting changes through `bot-settings-preview` and `bot-settings-apply`.
+
+## Inputs
+
+- Candle history from the market adapter
+- Feed price / settlement reference
+- Trend analyzer parameters:
+  - `erPeriod`
+  - `fastPeriod`
+  - `slowPeriod`
+  - `thresholdPercent`
+  - `minBarsForConfirmation`
+  - `deadZonePercent`
+- Bot settings fields that the signal may update:
+  - `weightDistribution`
+## Outputs
+
+- `UP`, `DOWN`, or `NEUTRAL`
+- Confidence score
+- Premium / discount signal
+- Oscillation / range context
+- Proposed bot-setting patch for the next cycle
+
+## Control Surfaces
+
+- General settings: `profiles/general.settings.json`
+- Per-bot settings: `profiles/bots.json`
+- Recalc triggers: `profiles/recalculate.<botKey>.trigger`
+- Signal state: `market_adapter/state/market_adapter_state.json`
+- Center snapshot: `market_adapter/state/market_adapter_centers.json`
+
+## Operating Rule
+
+Treat the trend signal as configuration input, not as trading logic. The shared service can tune bot settings and trigger recalculation, but it should not own order placement or margin-position policy.

package/dist/analysis/ama_fitting/analyze_ama_price_changes.d.ts

@@ -0,0 +1,3 @@
+#!/usr/bin/env node
+export {};
+//# sourceMappingURL=analyze_ama_price_changes.d.ts.map

package/dist/analysis/ama_fitting/analyze_ama_price_changes.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"analyze_ama_price_changes.d.ts","sourceRoot":"","sources":["../../../analysis/ama_fitting/analyze_ama_price_changes.ts"],"names":[],"mappings":""}