alphaear-predictor

1# AlphaEar Predictor Skill
2 
3## Overview
4 
5This skill utilizes the Kronos model (via `KronosPredictorUtility`) to perform time-series forecasting and adjust predictions based on news sentiment.
6 
7## Capabilities
8 
9### 1. Forecast Market Trends
10 
11### 1. Forecast Market Trends
12 
13**Workflow:**
141.  **Generate Base Forecast**: Use `scripts/kronos_predictor.py` (via `KronosPredictorUtility`) to generate the technical/quantitative forecast.
152.  **Adjust Forecast (Agentic)**: Use the **Forecast Adjustment Prompt** in `references/PROMPTS.md` to subjectively adjust the numbers based on latest news/logic.
16 
17**Key Tools:**
18-   `KronosPredictorUtility.get_base_forecast(df, lookback, pred_len, news_text)`: Returns `List[KLinePoint]`.
19 
20**Example Usage (Python):**
21 
22```python
23from scripts.utils.kronos_predictor import KronosPredictorUtility
24from scripts.utils.database_manager import DatabaseManager
25 
26db = DatabaseManager()
27predictor = KronosPredictorUtility()
28 
29# Forecast
30forecast = predictor.predict("600519", horizon="7d")
31print(forecast)
32```
33 
34 
35## Configuration
36 
37This skill requires the **Kronos** model and an embedding model.
38 
391.  **Kronos Model**:
40    -   Ensure `exports/models` directory exists in the project root.
41    -   Place trained news projector weights (e.g., `kronos_news_v1.pt`) in `exports/models/`.
42    -   Or depend on the base model (automatically downloaded).
43 
44> [!CAUTION]
45…