File size: 7,492 Bytes
54b5a2a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1d14fcd
54b5a2a
1d14fcd
 
54b5a2a
 
 
 
 
1d14fcd
54b5a2a
 
1d14fcd
54b5a2a
 
 
 
 
 
 
 
 
 
 
 
1d14fcd
 
54b5a2a
 
1d14fcd
 
 
 
 
 
 
 
 
 
 
 
 
 
54b5a2a
 
 
 
 
 
 
 
 
 
 
 
 
 
1d14fcd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
54b5a2a
 
1d14fcd
8a40c79
1d14fcd
 
54b5a2a
 
 
 
 
 
 
1d14fcd
54b5a2a
1d14fcd
 
 
54b5a2a
 
 
 
 
 
1d14fcd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
"""Pydantic schemas for key-point matching prediction endpoints"""

from pydantic import BaseModel, Field, ConfigDict
from typing import List, Optional, Dict


class PredictionRequest(BaseModel):
    """Request model for single key-point/argument prediction"""
    model_config = ConfigDict(
        json_schema_extra={
            "example": {
                "argument": "Climate change is accelerating due to industrial emissions.",
                "key_point": "Human industry contributes significantly to global warming."
            }
        }
    )

    argument: str = Field(
        ..., min_length=5, max_length=1000,
        description="The argument text to evaluate"
    )
    key_point: str = Field(
        ..., min_length=5, max_length=500,
        description="The key point used for comparison"
    )


class PredictionResponse(BaseModel):
    """Response model for single prediction"""
    model_config = ConfigDict(
        json_schema_extra={
            "example": {
                "prediction": 1,
                "confidence": 0.874,
                "label": "apparie",
                "probabilities": {
                    "non_apparie": 0.126,
                    "apparie": 0.874
                }
            }
        }
    )

    prediction: int = Field(..., description="1 = apparie, 0 = non_apparie")
    confidence: float = Field(..., ge=0.0, le=1.0,
                              description="Confidence score of the prediction")
    label: str = Field(..., description="apparie or non_apparie")
    probabilities: Dict[str, float] = Field(
        ..., description="Dictionary of class probabilities"
    )


class BatchPredictionRequest(BaseModel):
    """Request model for batch predictions"""
    model_config = ConfigDict(
        json_schema_extra={
            "example": {
                "pairs": [
                    {
                        "argument": "School uniforms limit students' self-expression and creativity",
                        "key_point": "Uniforms restrict personal freedom and individuality"
                    },
                    {
                        "argument": "Renewable energy creates more jobs than fossil fuel industries",
                        "key_point": "Green energy generates employment opportunities"
                    },
                    {
                        "argument": "We should invest more in renewable energy to combat climate change",
                        "key_point": "Capital punishment violates human rights"
                    },
                    {
                        "argument": "Online education provides flexibility and accessibility for all students",
                        "key_point": "Digital learning offers convenient and inclusive education"
                    },
                    {
                        "argument": "Vaccinations are essential for public health and disease prevention",
                        "key_point": "Space exploration leads to scientific discoveries"
                    }
                ]
            }
        }
    )

    pairs: List[PredictionRequest] = Field(
        ..., max_length=100,
        description="List of argument-keypoint pairs (max 100)"
    )


class BatchPredictionResponse(BaseModel):
    """Response model for batch key-point predictions"""
    model_config = ConfigDict(
        json_schema_extra={
            "example": {
                "predictions": [
                    {
                        "prediction": 1,
                        "confidence": 0.956,
                        "label": "apparie",
                        "probabilities": {
                            "non_apparie": 0.044,
                            "apparie": 0.956
                        }
                    },
                    {
                        "prediction": 1,
                        "confidence": 0.892,
                        "label": "apparie",
                        "probabilities": {
                            "non_apparie": 0.108,
                            "apparie": 0.892
                        }
                    },
                    {
                        "prediction": 0,
                        "confidence": 0.934,
                        "label": "non_apparie",
                        "probabilities": {
                            "non_apparie": 0.934,
                            "apparie": 0.066
                        }
                    },
                    {
                        "prediction": 1,
                        "confidence": 0.878,
                        "label": "apparie",
                        "probabilities": {
                            "non_apparie": 0.122,
                            "apparie": 0.878
                        }
                    },
                    {
                        "prediction": 0,
                        "confidence": 0.967,
                        "label": "non_apparie",
                        "probabilities": {
                            "non_apparie": 0.967,
                            "apparie": 0.033
                        }
                    }
                ],
                "total_processed": 5,
                "summary": {
                    "total_apparie": 3,
                    "total_non_apparie": 2,
                    "average_confidence": 0.9254
                }
            }
        }
    )

    predictions: List[PredictionResponse]
    total_processed: int = Field(..., description="Number of processed items")
    summary: Dict[str, float] = Field(
        default_factory=dict,
        description="Summary statistics of the batch prediction"
    )


class HealthResponse(BaseModel):
    """Health check model for the API"""
    model_config = ConfigDict(
        json_schema_extra={
            "example": {
                "status": "healthy",
                "model_loaded": True,
                "device": "cpu",
                "model_name": "NLP-Debater-Project/destlibert-keypoint-matching",
                "timestamp": "2024-01-01T12:00:00Z"
            }
        }
    )
    
    status: str = Field(..., description="API health status")
    model_loaded: bool = Field(..., description="Whether the model is loaded")
    device: str = Field(..., description="Device used for inference (cpu/cuda)")
    model_name: Optional[str] = Field(None, description="Name of the loaded model")
    timestamp: str = Field(..., description="Timestamp of the health check")


class ModelInfoResponse(BaseModel):
    """Detailed model information response"""
    model_config = ConfigDict(
        json_schema_extra={
            "example": {
                "model_name": "NLP-Debater-Project/destlibert-keypoint-matching",
                "device": "cpu",
                "max_length": 256,
                "num_labels": 2,
                "loaded": True,
                "performance": {
                    "accuracy": 0.9285,
                    "f1_score": 0.8836,
                    "f1_apparie": 0.8113,
                    "f1_non_apparie": 0.9559
                },
                "description": "DistilBERT model for key point - argument semantic matching"
            }
        }
    )

    model_name: str
    device: str
    max_length: int
    num_labels: int
    loaded: bool
    performance: Dict[str, float] = Field(
        ..., description="Model performance metrics"
    )
    description: str = Field(..., description="Model description")