File size: 7,180 Bytes
82a3b34
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
"""

Comparison Table Component



Gradio component for displaying model comparison tables.

"""

import logging
from typing import Any, List, Optional, Tuple

import gradio as gr
import plotly.graph_objects as go

from dashboard.schemas import ComparisonData, DeltaRobustnessData
from dashboard.utils import format_score, log_dashboard_event

logger = logging.getLogger(__name__)


def create_comparison_table() -> Tuple[gr.DataFrame, gr.Plot]:
    """

    Create comparison table component.

    

    Returns:

        Tuple of (table, plot) components

    """
    
    # Comparison table
    table = gr.DataFrame(
        headers=[
            "Model",
            "Hallucination",
            "Toxicity",
            "Bias",
            "Confidence",
            "Composite Score",
            "Sample Count",
        ],
        label="Model Comparison",
        interactive=False,
    )
    
    # Delta chart
    delta_plot = gr.Plot(
        label="Delta Robustness",
    )
    
    return table, delta_plot


def update_comparison_table(

    comparison_data: Optional[ComparisonData],

) -> List[List[str]]:
    """

    Update comparison table with data.

    

    Args:

        comparison_data: Comparison data

        

    Returns:

        Table data as list of rows

    """
    if comparison_data is None or not comparison_data.models:
        return [["N/A", "N/A", "N/A", "N/A", "N/A", "N/A", "0"]]
    
    table_data = []
    for i, model in enumerate(comparison_data.models):
        table_data.append([
            model,
            format_score(comparison_data.hallucination[i], 4),
            format_score(comparison_data.toxicity[i], 4),
            format_score(comparison_data.bias[i], 4),
            format_score(comparison_data.confidence[i], 4),
            format_score(comparison_data.composite_score[i], 4),
            str(comparison_data.sample_count[i]),
        ])
    
    return table_data


def update_delta_chart(

    delta_data: List[DeltaRobustnessData],

) -> Any:
    """

    Update delta chart with data.

    

    Args:

        delta_data: List of delta robustness data

        

    Returns:

        Plotly figure

    """
    if not delta_data:
        # Return empty chart
        fig = go.Figure()
        fig.update_layout(
            title="No Data Available",
            xaxis=dict(title="Model"),
            yaxis=dict(title="Delta Robustness"),
        )
        return fig
    
    log_dashboard_event("DASHBOARD_COMPARE_MODELS")
    
    models = [d.model_name for d in delta_data]
    deltas = [d.delta_robustness for d in delta_data]
    composites = [d.composite_score for d in delta_data]
    
    # Color based on delta (green for positive, red for negative)
    colors = ["#22c55e" if d >= 0 else "#ef4444" for d in deltas]
    
    fig = go.Figure(
        data=go.Bar(
            x=models,
            y=deltas,
            marker_color=colors,
            text=[f"Δ={c:.3f}" for c in composites],
            textposition="auto",
        )
    )
    
    fig.update_layout(
        title="Delta Robustness Comparison",
        xaxis=dict(title="Model"),
        yaxis=dict(
            title="Delta Robustness",
            range=[min(min(deltas) - 0.1, -0.1), max(max(deltas) + 0.1, 0.1)],
        ),
        height=400,
        width=600,
    )
    
    return fig


def create_ranking_table(

    delta_data: List[DeltaRobustnessData],

) -> List[List[str]]:
    """

    Create ranking table from delta data.

    

    Args:

        delta_data: List of delta robustness data

        

    Returns:

        Table data as list of rows

    """
    if not delta_data:
        return [["N/A", "N/A", "N/A", "N/A"]]
    
    # Sort by composite score descending
    sorted_data = sorted(delta_data, key=lambda x: x.composite_score, reverse=True)
    
    table_data = []
    for rank, data in enumerate(sorted_data, 1):
        table_data.append([
            str(rank),
            data.model_name,
            format_score(data.composite_score, 4),
            format_score(data.delta_robustness, 4),
        ])
    
    return table_data


class ComparisonTable:
    """

    Comparison table component with state management.

    """
    
    def __init__(self):
        """Initialize comparison table."""
        self._comparison_data: Optional[ComparisonData] = None
        self._delta_data: List[DeltaRobustnessData] = []
    
    def set_comparison_data(self, data: ComparisonData) -> None:
        """

        Set comparison data.

        

        Args:

            data: Comparison data

        """
        self._comparison_data = data
    
    def set_delta_data(self, data: List[DeltaRobustnessData]) -> None:
        """

        Set delta data.

        

        Args:

            data: List of delta robustness data

        """
        self._delta_data = data
    
    def get_comparison_data(self) -> Optional[ComparisonData]:
        """

        Get comparison data.

        

        Returns:

            Comparison data or None

        """
        return self._comparison_data
    
    def get_delta_data(self) -> List[DeltaRobustnessData]:
        """

        Get delta data.

        

        Returns:

            List of delta robustness data

        """
        return self._delta_data
    
    def get_table_data(self) -> List[List[str]]:
        """

        Get table data for display.

        

        Returns:

            Table data as list of rows

        """
        return update_comparison_table(self._comparison_data)
    
    def get_delta_figure(self) -> Any:
        """

        Get delta chart figure.

        

        Returns:

            Plotly figure

        """
        return update_delta_chart(self._delta_data)


def create_model_selector_multi() -> gr.Component:
    """

    Create multi-select model selector.

    

    Returns:

        Gradio multi-select component

    """
    selector = gr.Dropdown(
        label="Select Models to Compare",
        choices=[],
        interactive=True,
        multiselect=True,
    )
    
    return selector


def create_benchmark_comparison(

    benchmark_runs: List[dict],

) -> gr.DataFrame:
    """

    Create benchmark comparison table.

    

    Args:

        benchmark_runs: List of benchmark run data

        

    Returns:

        DataFrame component

    """
    headers = ["Benchmark", "Model", "Score", "Status", "Date"]
    
    table_data = []
    for run in benchmark_runs:
        table_data.append([
            run.get("benchmark_name", "N/A"),
            run.get("model_name", "N/A"),
            format_score(run.get("composite_score"), 4),
            run.get("status", "N/A"),
            run.get("timestamp", "N/A")[:10],
        ])
    
    table = gr.DataFrame(
        headers=headers,
        value=table_data if table_data else [["N/A"] * len(headers)],
        label="Benchmark Comparison",
        interactive=False,
    )
    
    return table