DSL Primitive Browser & Strategy Builder

Trading Strategy Domain Specific Language - Agencio Predict

417

Primitives

1. System Overview

The DSL system enables users to create trading strategies without writing code. Primitives are building blocks that return values (numbers, booleans, strings) which are combined using operators and logic to form trading rules.

DSL System Architecture

flowchart TB subgraph UserInput["User Input Methods"] NL["Natural Language\nFrom English"] VB["Visual Builder\nGuided Builder"] DSL["Manual DSL\nWrite DSL"] TPL["Templates\nPre-built strategies"] end subgraph DSLCore["DSL Core System"] PARSE["DSL Parser\nYAML to AST"] VALID["Validator\nPrimitive whitelist"] EVAL["Evaluator\nBar-by-bar execution"] end subgraph Primitives["417 Primitives"] P1["Price (32)\nprice, volume, vwap"] P2["Technical (85)\nrsi, macd, sma, bb"] P3["Sentiment (28)\nnews, social, analyst"] P4["Macro (24)\nvix, yield_curve, fed"] P5["Pattern (35)\nfvg, candlestick, sr"] P6["Options (18)\niv, put_call, gamma"] P7["PINN (60+)\nneural network signals"] P8["Others (135)\nfundamental, sizing, etc"] end subgraph Execution["Execution Layer"] BACKTEST["Backtest Engine\n90-day historical"] PAPER["Paper Executor\nSimulated trading"] LIVE["Live Executor\nReal broker orders"] end NL --> |"LLM Translation"| PARSE VB --> |"Generate YAML"| PARSE DSL --> PARSE TPL --> PARSE PARSE --> VALID VALID --> EVAL P1 & P2 & P3 & P4 & P5 & P6 & P7 & P8 --> EVAL EVAL --> BACKTEST & PAPER & LIVE

2. Primitive Categories (14)

Category Overview

flowchart TB subgraph MarketData["Market Data"] PRICE["Price\n32 primitives"] TECH["Technical\n85 primitives"] PATTERN["Pattern\n35 primitives"] end subgraph AlternativeData["Alternative Data"] SENT["Sentiment\n36 primitives"] MACRO["Macro\n24 primitives"] PRED["Prediction Markets\n12 primitives"] end subgraph Derivatives["Derivatives"] OPT["Options\n18 primitives"] FUND["Fundamental\n22 primitives"] end subgraph AI["AI/ML"] PINN["PINN\n60+ primitives"] LLM["LLM\n8 primitives"] end subgraph Trading["Trading"] POS["Position\n15 primitives"] SIZE["Sizing\n12 primitives"] COMP["Composer\n10 primitives"] WEATHER["Weather\n6 primitives"] end

Category Details

Category	Count	Description	Key Primitives
Price	32	Real-time and historical price, volume, VWAP data	`price()`, `volume()`, `vwap()`, `price_change()`
Technical	85	Technical indicators: RSI, MACD, moving averages, Bollinger Bands, GARCH, volatility models	`rsi(14)`, `macd()`, `sma(50)`, `garch_volatility()`
Sentiment	36	News, social media, analyst ratings, fear/greed, insider activity, mailbox/email intelligence	`sentiment("reddit", "24h")`, `fear_greed_index()`, `mailbox_sentiment()`, `mailbox_consensus_score()`
Macro	24	VIX, yield curve, Fed funds, regime detection, sector rotation	`vix()`, `yield_curve_slope()`, `markov_regime()`
Pattern	35	FVG zones, candlestick patterns, support/resistance, DTW templates	`fvg_bullish()`, `hammer()`, `dtw_template_match()`
Options	18	Implied volatility, put/call ratio, gamma exposure, IV term structure	`iv_rank()`, `put_call_ratio()`, `vol_selling_signal()`
PINN	60+	Physics-Informed Neural Network predictions: direction, volatility, sizing, correlation, regime	`pinn_direction()`, `pinn_confidence()`, `pinn_optimal_size()`
Fundamental	45+	P/E ratio, earnings, institutional holdings, activist investors, crisis resilience	`pe_ratio()`, `institutional_ownership_pct()`, `resilience_score()`
Position	20+	Current position info, P&L, drawdown, factor VaR, stress testing	`position_pnl_pct()`, `drawdown_from_peak()`, `factor_var()`
Sizing	12	Kelly criterion, volatility-adjusted sizing, risk parity	`kelly(0.04, 5000)`, `risk_parity_weight()`
Prediction Markets	17	Polymarket, Kalshi probabilities, wallet clustering, fraud detection	`prediction_market_prob("fed-rate-cut")`, `polymarket_manipulation_risk()`
Composer	3	Combine signals: all, any, xor	`all(signals)`, `any(signals)`, `xor(a, b)`
Weather	20+	El Nino, hurricanes, temperature, climate indicators	`el_nino_index()`, `hurricane_season()`, `climate_risk()`
LLM	1	AI-powered analysis and confirmation	`llm_confirms("claim", 0.7)`

3. User Flow

Strategy Creation User Journey

flowchart TB subgraph Entry["Entry Points"] ALG["Trading > Algorithms\nPrimitives Tab"] WIZ["AI Trader Wizard\nGuided Builder"] SMART["Smart Portfolio\nStep 3 Strategies"] end subgraph Browse["Browse Primitives"] SEARCH["Search by name\nor description"] FILTER["Filter by\ncategory"] EXPAND["Expand category\nview details"] end subgraph Learn["Learn Primitive"] DESC["Read description"] ARGS["See arguments\nand return type"] EXAMPLE["Copy example\nusage"] end subgraph Build["Build Strategy"] ENTRY["Define entry\nconditions"] EXIT["Define exit\nconditions"] RISK["Set risk\nparameters"] end subgraph Test["Test Strategy"] BACK["Run backtest\n90 days"] WALK["Walk-forward\nvalidation"] STRESS["Regime stress\ntest"] end subgraph Deploy["Deploy"] PAPER["Start paper\ntrading"] LIVE["Go live\nafter graduation"] end ALG & WIZ & SMART --> SEARCH & FILTER SEARCH & FILTER --> EXPAND EXPAND --> DESC & ARGS & EXAMPLE DESC & ARGS & EXAMPLE --> ENTRY ENTRY --> EXIT --> RISK RISK --> BACK --> WALK --> STRESS STRESS --> PAPER --> LIVE

4. Integration Points

The DSL Primitive Browser and Simple Strategy Builder are integrated at three key locations in the platform:

Integration Architecture

flowchart TB subgraph Components["Shared Components"] BROWSER["DSLPrimitiveBrowser\npackages/fe/components/algorithms/"] BUILDER["SimpleStrategyBuilder\npackages/fe/components/algorithms/"] end subgraph Location1["Trading > Algorithms"] TAB1["Primitives Tab"] VIEW1["Full browser view\nwith categories"] end subgraph Location2["AI Trader Wizard"] STEP1["Welcome Step\nGuided Builder option"] STEP2["Create Step\nCollapsible browser"] end subgraph Location3["Smart Portfolio Wizard"] STEP3["Step 3 Strategies\nExplore DSL button"] BUILD3["Build Custom\nbutton"] end BROWSER --> TAB1 & STEP2 & STEP3 BUILDER --> STEP2 & BUILD3 TAB1 --> VIEW1 STEP1 --> STEP2 STEP3 --> BUILD3

Integration Details

Location	Component	Mode	Purpose
/algorithms (Primitives tab)	DSLPrimitiveBrowser	Full	Browse all 417 primitives with full descriptions
AI Trader Wizard (Create step)	DSLPrimitiveBrowser + SimpleStrategyBuilder	Collapsible	Reference primitives while building strategy
Smart Portfolio Wizard (Step 3)	DSLPrimitiveBrowser + SimpleStrategyBuilder	Compact	Explore and customize generated strategies

5. Primitive Browser Component

Primitive Browser UI Structure

flowchart TB subgraph Header["Header Section"] HELP["What are DSL Primitives?\nExplanation with example"] end subgraph Search["Search & Filter"] INPUT["Search input\nFilter by name/description"] CATS["Category chips\nClick to filter"] COUNT["Result count\nX primitives found"] end subgraph CategoryView["Category Accordion"] CAT1["Price (32)\nExpand/collapse"] CAT2["Technical (85)\nExpand/collapse"] CAT3["...\n14 categories total"] end subgraph PrimitiveCard["Primitive Details"] NAME["Primitive name\nwith arguments"] TYPE["Return type\nnumber/boolean/string"] DESC["Description"] BTN["Copy button\nInsert into editor"] end subgraph Footer["Quick Reference"] OPS["Operators\n< > <= >= == != && ||"] LOGIC["Logic\nall any count threshold"] DUR["Durations\n1h 4h 24h 7d"] end Header --> Search --> CategoryView CAT1 & CAT2 & CAT3 --> PrimitiveCard PrimitiveCard --> Footer

Component Props

interface DSLPrimitiveBrowserProps {
  primitives: PrimitiveSpec[];      // Array of all primitives
  onInsert?: (text: string) => void; // Insert into editor callback
  onSelectPrimitive?: (name: string) => void; // Selection callback
  className?: string;                // Additional CSS classes
  compact?: boolean;                 // Hide header/footer for embedding
}

6. Simple Strategy Builder

Strategy Builder Flow

flowchart TB subgraph Templates["Quick Start Templates"] T1["RSI Mean Reversion"] T2["Golden Cross"] T3["Sentiment Breakout"] T4["Fear & Greed Contrarian"] end subgraph Config["Basic Configuration"] NAME["Strategy Name"] SYMBOL["Symbol\nAAPL, BTC-USD, etc"] end subgraph Entry["Entry Conditions (AND)"] E1["Primitive dropdown\nrsi, sma, sentiment..."] E2["Operator\n< > <= >= == !="] E3["Value\nnumber or primitive"] E4["+ Add condition"] end subgraph Exit["Exit Conditions (ANY)"] X1["Primitive dropdown"] X2["Operator"] X3["Value"] X4["+ Add condition"] end subgraph Risk["Risk Settings"] R1["Max Position USD"] R2["Stop Loss %"] end subgraph Output["Generated DSL"] YAML["YAML output\nReady to use"] COPY["Copy button"] CREATE["Create Strategy button"] end T1 & T2 & T3 & T4 --> Config Config --> Entry --> Exit --> Risk --> Output

Default Templates

Template	Entry Condition	Exit Condition
RSI Mean Reversion	`rsi(14) < 30`	`rsi(14) > 70`
Golden Cross	`sma(50) > sma(200) AND rsi(14) < 70`	`sma(50) < sma(200)`
Sentiment Breakout	`sentiment("reddit", "24h") > 0.6 AND vix() < 20`	`sentiment("reddit", "24h") < 0.3`
Fear & Greed Contrarian	`fear_greed_index() < 25`	`fear_greed_index() > 75`

7. Strategy Examples

Example 1: RSI + Moving Average Strategy

strategy: "RSI + MA Crossover"
mode: paper
universe: [AAPL]

signals:
  entry:
    - rsi(14) < 30
    - sma(50) > sma(200)
  exit:
    - rsi(14) > 70

entry:
  when: all(entry)
  size: kelly(0.04, 5000)

exit:
  when: any(exit)

risk:
  max_position_usd: 5000
  stop_loss_pct: 3
  max_daily_loss_pct: 2

Example 2: PINN + Sentiment Strategy

strategy: "AI-Enhanced Momentum"
mode: paper
universe: [SPY]

signals:
  entry:
    - pinn_is_bullish("SPY", 0.6)
    - pinn_crisis_warning() < 0.3
    - sentiment("news", "24h") > 0.5
  exit:
    - pinn_is_bearish("SPY", 0.5)
    - position_pnl_pct() > 5

entry:
  when: all(entry)
  size: pinn_optimal_size("SPY", 30)

exit:
  when: any(exit)

risk:
  max_position_usd: 10000
  stop_loss_pct: 5
  trailing_stop_pct: 3

Example 3: Pattern + Macro Strategy

strategy: "Pattern Breakout"
mode: paper
universe: [QQQ]

signals:
  entry:
    - fvg_bullish()
    - vix() < 25
    - volume_z("", "20d") > 1.5
  exit:
    - fvg_bearish()
    - vix() > 35

entry:
  when: all(entry)
  size: volatility_size(2, 0.02)

exit:
  when: any(exit)

risk:
  max_position_usd: 8000
  stop_loss_pct: 4

8. API Reference

Endpoints

Method	Path	Description
GET	`/api/predict/v1/algorithms/primitives`	Get all 417 primitives with metadata
POST	`/api/predict/v1/algorithms/from-english`	Translate natural language to DSL
POST	`/api/predict/v1/algorithms/validate`	Validate DSL syntax and primitives
POST	`/api/predict/v1/algorithms`	Create algorithm from DSL

File References

Component	Path	Lines
DSLPrimitiveBrowser	`packages/fe/src/components/algorithms/DSLPrimitiveBrowser.tsx`	~450
SimpleStrategyBuilder	`packages/fe/src/components/algorithms/SimpleStrategyBuilder.tsx`	~413
DSL Evaluator	`packages/be/src/algorithms/dsl/evaluator.ts`	~2000
DSL Types	`packages/be/src/algorithms/dsl/types.ts`	~2500
Primitives API	`packages/be/src/algorithms/api/handlers.ts`	~600

9. Complete Primitive Reference (417)

Below is the complete reference of all 417 DSL primitives, organized by category. Each primitive shows its name, arguments, return type, and description.

Loading primitives...

DSL Primitive Browser Documentation - Agencio Predict
Generated: 2026-05-20 | 417 Primitives | 14 Categories
Integrated with AI Trader Wizard and Smart Portfolio Wizard