File size: 2,438 Bytes
2b21f2d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
"""Renders a call's full transcript as attention-shaded HTML, driven by the per-turn
attention_weight values exported by src/evaluation/export_dashboard_data.py."""

import html

from theme import BORDER_HAIRLINE, TEXT_SECONDARY

SECTION_LABELS = {
    "operator_transition": "Operator",
    "prepared_remarks": "Prepared Remarks",
    "qa": "Q&A",
}


def _intensity_color(weight: float, max_weight: float) -> str:
    """Accent-tinted background, opacity scaled by relative attention weight."""
    if max_weight <= 0:
        return "transparent"
    frac = min(weight / max_weight, 1.0)
    alpha = 0.05 + 0.30 * frac
    r, g, b = 255, 87, 34
    return f"rgba({r},{g},{b},{alpha:.3f})"


def render_full_transcript(turns: list[dict], selected_turn: int | None = None) -> str:
    """Full ordered transcript, grouped into sections, each turn background-shaded by
    its own attention weight relative to the call's max.

    `selected_turn`, when set, adds the `fl-timeline-selected` class to that turn's
    block (used by the Workspace tab to sync highlighting with the Timeline/Evidence
    panel)."""
    if not turns:
        return "<p>No transcript available.</p>"
    max_w = max(t["attention_weight"] for t in turns)

    blocks = []
    current_section = None
    for t in turns:
        section = SECTION_LABELS.get(t["segment_type"], t["segment_type"])
        if section != current_section:
            blocks.append(f'<div class="fl-section-header">{html.escape(section)}</div>')
            current_section = section
        text = html.escape(t["text"])
        speaker = html.escape(t.get("speaker_name") or t["speaker_role"])
        bg = _intensity_color(t["attention_weight"], max_w)
        selected_cls = " fl-timeline-selected" if t["turn_index"] == selected_turn else ""
        xray_cls = " fl-xray-highlight" if t.get("is_risk_event") else ""
        blocks.append(
            f'<div class="fl-turn{selected_cls}{xray_cls}" id="turn-{t["turn_index"]}" data-turn-index="{t["turn_index"]}" '
            f'style="background:{bg}; border-left-color:{TEXT_SECONDARY};">'
            f'<div class="fl-turn-meta">{speaker} ({t["speaker_role"]})</div>'
            f'<div class="fl-turn-text">{text}</div>'
            f"</div>"
        )
    return (
        f'<div style="border:1px solid {BORDER_HAIRLINE}; padding:12px; max-height:520px; overflow-y:auto;">'
        + "".join(blocks)
        + "</div>"
    )