app.py

import tempfile
import traceback
from pathlib import Path
import gradio as gr
from pipeline import get_pipeline, to_json, to_csv, to_abab_text

UPLOAD_DIR = Path(tempfile.gettempdir()) / "speech_annotation"
UPLOAD_DIR.mkdir(parents=True, exist_ok=True)
_last_segments = []

SPEAKER_COLORS = [
    ("#4F46E5", "#EEF2FF"),
    ("#059669", "#ECFDF5"),
    ("#DC2626", "#FEF2F2"),
    ("#D97706", "#FFFBEB"),
    ("#7C3AED", "#F5F3FF"),
    ("#0891B2", "#ECFEFF"),
    ("#DB2777", "#FDF2F8"),
    ("#65A30D", "#F7FEE7"),
    ("#EA580C", "#FFF7ED"),
    ("#0284C7", "#F0F9FF"),
]

def make_conversation_html(segments):
    if not segments:
        return ""
    speaker_list = list(dict.fromkeys(s.speaker for s in segments))
    color_map = {spk: SPEAKER_COLORS[i % len(SPEAKER_COLORS)] for i, spk in enumerate(speaker_list)}
    legend_items = "".join(
        f"<span style='display:inline-flex;align-items:center;gap:6px;margin-right:16px'>"
        f"<span style='width:12px;height:12px;border-radius:50%;background:{color_map[spk][0]}'></span>"
        f"<span style='font-weight:600;color:{color_map[spk][0]}'>Speaker {spk}</span></span>"
        for spk in speaker_list
    )
    legend = f"<div style='padding:12px 16px;border-bottom:1px solid #e5e7eb;display:flex;flex-wrap:wrap;gap:4px'>{legend_items}</div>"
    bubbles = ""
    for seg in segments:
        fg, bg = color_map[seg.speaker]
        align = "flex-end" if speaker_list.index(seg.speaker) % 2 == 1 else "flex-start"
        radius = "4px 16px 16px 16px" if align == "flex-start" else "16px 4px 16px 16px"
        ta = "text-align:right;" if align == "flex-end" else ""
        bubbles += (
            f"<div style='display:flex;justify-content:{align};margin:6px 12px'>"
            f"<div style='max-width:75%'>"
            f"<div style='font-size:11px;color:#6b7280;margin-bottom:3px;{ta}'>"
            f"<span style='font-weight:600;color:{fg}'>Speaker {seg.speaker}</span>"
            f" &middot; {seg.start_fmt} &rarr; {seg.end_fmt}</div>"
            f"<div style='background:{bg};border:1px solid {fg}30;color:#111827;"
            f"padding:10px 14px;border-radius:{radius};font-size:14px;line-height:1.5'>"
            f"{seg.text}</div></div></div>"
        )
    return (
        f"<div style='border:1px solid #e5e7eb;border-radius:12px;overflow:hidden;"
        f"font-family:system-ui,sans-serif;background:white'>"
        f"{legend}"
        f"<div style='height:480px;overflow-y:auto;padding:8px 0;background:#f9fafb'>"
        f"{bubbles}</div></div>"
    )

def make_table_html(segments):
    if not segments:
        return ""
    speaker_list = list(dict.fromkeys(s.speaker for s in segments))
    color_map = {spk: SPEAKER_COLORS[i % len(SPEAKER_COLORS)] for i, spk in enumerate(speaker_list)}
    rows = "".join(
        f"<tr style='border-bottom:1px solid #f3f4f6'>"
        f"<td style='padding:8px 12px'><span style='background:{color_map[s.speaker][1]};"
        f"color:{color_map[s.speaker][0]};padding:3px 10px;border-radius:99px;"
        f"font-weight:700;font-size:13px'>{s.speaker}</span></td>"
        f"<td style='padding:8px 12px;color:#6b7280;font-size:13px'>{s.start_fmt}</td>"
        f"<td style='padding:8px 12px;color:#6b7280;font-size:13px'>{s.end_fmt}</td>"
        f"<td style='padding:8px 12px;font-size:14px;color:#111827'>{s.text}</td>"
        f"</tr>"
        for s in segments
    )
    return (
        f"<div style='border:1px solid #e5e7eb;border-radius:12px;overflow:hidden;"
        f"font-family:system-ui,sans-serif'>"
        f"<table style='width:100%;border-collapse:collapse'>"
        f"<thead><tr style='background:#1e3a5f;color:white'>"
        f"<th style='padding:10px 12px;text-align:left'>Speaker</th>"
        f"<th style='padding:10px 12px;text-align:left'>Start</th>"
        f"<th style='padding:10px 12px;text-align:left'>End</th>"
        f"<th style='padding:10px 12px;text-align:left'>Transcript</th>"
        f"</tr></thead><tbody>{rows}</tbody></table></div>"
    )

def process_audio(audio_path, num_speakers):
    global _last_segments
    if audio_path is None:
        return "Please upload an audio file first.", "", ""
    try:
        pipeline = get_pipeline()
        n = int(num_speakers) if num_speakers and int(num_speakers) > 0 else 0
        segments = pipeline.process(audio_path, num_speakers=n)
    except Exception as e:
        return f"Error: {e}\n{traceback.format_exc()}", "", ""
    if not segments:
        return "No speech detected.", "", ""
    _last_segments = segments
    unique = len(set(s.speaker for s in segments))
    return f"Done - {len(segments)} segments, {unique} speaker(s) detected", make_conversation_html(segments), make_table_html(segments)

def export_json():
    if not _last_segments: return None
    out = str(UPLOAD_DIR / "annotation.json")
    to_json(_last_segments, out)
    return out

def export_csv():
    if not _last_segments: return None
    out = str(UPLOAD_DIR / "annotation.csv")
    to_csv(_last_segments, out)
    return out

with gr.Blocks(title="Speech Annotation Pipeline") as demo:
    gr.Markdown("# Speech Annotation Pipeline\n*Upload audio · Detect speakers · Export transcript*")
    with gr.Row():
        with gr.Column(scale=1):
            audio_input = gr.Audio(label="Upload Audio (.wav / .mp3 / .flac)", type="filepath")
            num_speakers = gr.Slider(minimum=0, maximum=10, step=1, value=0,
                                     label="Number of speakers (0 = auto-detect)")
            run_btn = gr.Button("Run Annotation", variant="primary")
            status_box = gr.Textbox(label="Status", value="Ready.", interactive=False)
            with gr.Row():
                json_btn = gr.Button("Export JSON")
                csv_btn  = gr.Button("Export CSV")
            json_file = gr.File(label="JSON Download")
            csv_file  = gr.File(label="CSV Download")
        with gr.Column(scale=2):
            gr.Markdown("### Conversation View")
            conversation_html = gr.HTML(value="<div style='height:480px;border:1px solid #e5e7eb;border-radius:12px;display:flex;align-items:center;justify-content:center;color:#9ca3af'>Transcript will appear here after processing</div>")
            gr.Markdown("### Segment Table")
            table_html = gr.HTML(value="")

    run_btn.click(fn=process_audio, inputs=[audio_input, num_speakers],
                  outputs=[status_box, conversation_html, table_html])
    json_btn.click(fn=export_json, inputs=[], outputs=[json_file])
    csv_btn.click(fn=export_csv, inputs=[], outputs=[csv_file])

if __name__ == "__main__":
    demo.launch(server_name="0.0.0.0", server_port=7860)