File size: 1,393 Bytes
d2f66eb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
"""
OpenEnv Graders for StockSense-AI.
These functions provide programmatic evaluation of the environment tasks.
Each grader returns a float score in the range (0.01, 0.99).
"""

import json

def _clamp_score(val: float) -> float:
    """Ensure score is strictly within the open interval (0, 1)."""
    return max(0.01, min(0.99, val))

def grade_prediction(agent_output: str, **kwargs) -> float:
    """
    Grades the market direction prediction task.
    Checks for presence of direction and a valid score.
    """
    try:
        # Standard OpenEnv inference outputs are often JSON
        data = json.loads(agent_output)
        score = float(data.get("score", 0.5))
        return _clamp_score(score)
    except:
        # If output is raw text or invalid, return a safe neutral score
        return 0.5

def grade_sentiment(agent_output: str, **kwargs) -> float:
    """
    Grades the news sentiment analysis task.
    """
    try:
        if "sentiment" in agent_output.lower() or "news" in agent_output.lower():
            return 0.85
        return 0.5
    except:
        return 0.5

def grade_technical(agent_output: str, **kwargs) -> float:
    """
    Grades the technical indicator interpretation task.
    """
    try:
        if "indicator" in agent_output.lower() or "rsi" in agent_output.lower():
            return 0.85
        return 0.5
    except:
        return 0.5