prediction_sdk 0.1.2

Advanced cryptocurrency forecasting SDK combining statistical models, machine learning (SVR), and technical analysis with intelligent caching and rate limiting
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141

# Inferenco Predictions SDK

A Rust library for cryptocurrency price forecasting that combines statistical methods, technical analysis, and machine learning.

## Projects using this SDK

- [DeFiCalc](https://deficalc.io/) integrates the SDK for price-aware DeFi calculations.
- [Inferenco Nova](https://inferenco.com/app.html#nova) uses the SDK to power AI-driven blockchain conversations and tooling.

## Features

- **Hybrid Forecasting Engine**: Ensemble of statistical models, ML (SVR), and technical indicators
- **Multi-Horizon Forecasts**: From 15 minutes to 4 years
- **Two-Tier Caching**: Smart TTL-based caching for API responses and forecasts
- **Rate Limiting**: Built-in token bucket + exponential backoff for CoinGecko API
- **Technical Analysis**: RSI, MACD, Bollinger Bands
- **Local AI**: On-the-fly SmartCore SVR training (no external models needed)

## Installation

Add to your `Cargo.toml`:

```toml
[dependencies]
prediction_sdk = "0.1.0"
tokio = { version = "1", features = ["full"] }
```

## Quick Start

```bash
cd prediction_sdk
cargo build
cargo test
```

## Usage

```rust
use prediction_sdk::{PredictionSdk, ForecastHorizon, ShortForecastHorizon, SentimentSnapshot};

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let sdk = PredictionSdk::new()?;
    
    let sentiment = SentimentSnapshot {
        news_score: 0.65,
        social_score: 0.80,
    };
    
    let result = sdk.forecast_with_fetch(
        "bitcoin",
        "usd",
        ForecastHorizon::Short(ShortForecastHorizon::OneHour),
        Some(sentiment),
    ).await?;
    
    println!("Forecast: {:?}", result);
    Ok(())
}
```

### Fetching Charts

To retrieve historical candles and (for long horizons) projection bands, set the `chart` flag to `true` in your request. This is best handled via the `run_prediction_handler` helper or by manually calling `fetch_chart_candles`.

```rust
use prediction_sdk::{ForecastRequest, ForecastHorizon, ShortForecastHorizon, run_prediction_handler};

// ... inside async fn
let request = ForecastRequest {
    asset_id: "bitcoin".to_string(),
    vs_currency: "usd".to_string(),
    horizon: ForecastHorizon::Short(ShortForecastHorizon::OneHour),
    sentiment: None,
    chart: true, // <--- Enable chart data
};

// Returns a JSON string containing both "forecast" and "chart" fields
let json_response = run_prediction_handler(request).await?;
println!("{}", json_response);
```

**Response Structure:**

```json
{
  "forecast": {
    "type": "short",
    "value": {
      "expected_price": 42350.0,
      "confidence": 0.85,
      // ... other forecast fields
    }
  },
  "chart": {
    "history": [
      {
        "timestamp": "2024-01-01T00:00:00Z",
        "open": 42000.0,
        "high": 42500.0,
        "low": 41800.0,
        "close": 42350.0,
        "volume": 1823.4
      }
      // ... more candles
    ],
    "projection": [
      // Only present for Long horizons
      {
        "timestamp": "2024-01-02T00:00:00Z", 
        "percentile_10": 40000.0, 
        "mean": 43000.0, 
        "percentile_90": 46000.0
      }
    ]
  }
}
```

## Run the Example

```bash
cargo run --example real_world_test
```

## Documentation

See [DOCUMENTATION.md](DOCUMENTATION.md) for detailed architecture, API reference, and integration guides.

## CI/CD

GitHub Actions runs:
- `cargo fmt --check`
- `cargo clippy`
- `cargo build`
- `cargo test` (unit + integration + doc tests)

## License

MIT