audex 1.0.7a3__py3-none-any.whl

audex/view/pages/recording.py

@@ -0,0 +1,348 @@
+from __future__ import annotations
+
+import asyncio
+import typing as t
+
+from dependency_injector.wiring import Provide
+from dependency_injector.wiring import inject
+from fastapi import Depends
+from fastapi import Query
+from nicegui import ui
+
+from audex.container import Container
+from audex.service.doctor import DoctorService
+from audex.service.session import SessionService
+from audex.service.session.types import CreateSessionCommand
+from audex.service.session.types import Delta
+from audex.service.session.types import Done
+from audex.service.session.types import Start
+from audex.view.decorators import handle_errors
+
+
+@ui.page("/recording")
+@handle_errors
+@inject
+async def render(
+    doctor_service: DoctorService = Depends(Provide[Container.service.doctor]),
+    session_service: SessionService = Depends(Provide[Container.service.session]),
+    session_id: str | None = Query(default=None),
+) -> None:
+    """Render recording session page with lyrics-style scrolling."""
+
+    doctor = await doctor_service.current_doctor()
+    has_vp = await doctor_service.has_voiceprint()
+
+    if not has_vp:
+        ui.notify("请先注册声纹后再使用录音功能", type="warning", position="top")
+        ui.navigate.to("/voiceprint/enroll")
+        return
+
+    ui.add_head_html('<link rel="stylesheet" href="/static/css/recording.css">')
+
+    # Global auto-scroll JavaScript
+    ui.add_head_html('<script src="/static/js/recording.js"></script>')
+
+    # Tasks
+    asyncio_tasks: dict[str, asyncio.Task] = {}
+
+    # State variables
+    session_id_state: dict[str, str | None] = {"value": session_id}
+    is_recording = {"value": False}
+    session_context: dict[str, t.Any] = {"value": None}
+    is_session_completed = {"value": False}
+
+    current_utterance_element: dict[str, t.Any] = {"element": None}
+    current_sequence: dict[str, int] = {"value": 0}
+
+    # Header with glass effect
+    with (
+        ui.header().classes("header-glass items-center justify-between px-6 py-3"),
+        ui.row().classes("items-center gap-3"),
+    ):
+        ui.button(icon="arrow_back", on_click=lambda: ui.navigate.to("/")).props(
+            "flat round"
+        ).tooltip("返回主面板")
+        ui.label("录音会话").classes("text-h6 font-semibold text-grey-8")
+
+    # Scroll to bottom button
+    (
+        ui.button(
+            icon="keyboard_arrow_down", on_click=lambda: ui.run_javascript("scrollToBottom()")
+        )
+        .props("round color=white text-color=grey-8")
+        .classes("scroll-bottom-btn")
+        .tooltip("滚动到底部")
+    )
+
+    # Main scrollable container
+    lyrics_container = ui.element("div").classes("lyrics-container").props('id="lyrics-container"')
+    with lyrics_container:
+        lyrics_column = ui.column().classes("w-full items-center")
+
+    # Footer glass overlay
+    ui.element("div").classes("footer-glass")
+
+    async def load_existing_utterances():
+        """Load existing conversation history."""
+        if not session_id_state["value"]:
+            return
+
+        try:
+            utterances = await session_service.get_utterances(session_id_state["value"])
+
+            for utterance in utterances:
+                with lyrics_column:
+                    elem = ui.element("div").classes("utterance-final")
+
+                    if utterance.speaker.value == "doctor":
+                        elem.classes(add="is-doctor")
+                        with elem:
+                            ui.label("医生").classes("speaker-label")
+                            ui.html(
+                                f'<div class="bubble-doctor">{utterance.text}</div>', sanitize=False
+                            )
+                    else:
+                        elem.classes(add="is-patient")
+                        with elem:
+                            ui.label("患者").classes("speaker-label")
+                            ui.html(
+                                f'<div class="bubble-patient">{utterance.text}</div>',
+                                sanitize=False,
+                            )
+
+                current_sequence["value"] = max(current_sequence["value"], utterance.sequence)
+
+            # JavaScript handles auto-scroll via MutationObserver
+
+        except Exception as e:
+            print(f"Failed to load utterances: {e}")
+
+    async def wait_for_vpr_and_render(
+        elem: ui.element, sequence: int, text: str, max_retries: int = 30
+    ):
+        """Wait for VPR completion and render final bubble."""
+        ctx = session_context["value"]
+
+        for _ in range(max_retries):
+            is_doctor = ctx._vpr_results.get(sequence)
+
+            if is_doctor is not None:
+                elem.classes(remove="utterance-item utterance-current utterance-past")
+                elem.classes(add="utterance-final")
+
+                if is_doctor:
+                    elem.classes(add="is-doctor")
+                else:
+                    elem.classes(add="is-patient")
+
+                elem.clear()
+                with elem:
+                    ui.label("医生" if is_doctor else "患者").classes("speaker-label")
+                    ui.html(
+                        f'<div class="bubble-{"doctor" if is_doctor else "patient"}">{text}</div>',
+                        sanitize=False,
+                    )
+
+                # JavaScript MutationObserver will handle auto-scroll
+                return
+
+            await asyncio.sleep(0.1)
+
+        # Timeout - default to patient
+        elem.classes(remove="utterance-item utterance-current utterance-past")
+        elem.classes(add="utterance-final is-patient")
+        elem.clear()
+        with elem:
+            ui.label("患者").classes("speaker-label")
+            ui.html(f'<div class="bubble-patient">{text}</div>', sanitize=False)
+
+    async def process_transcription():
+        """Process transcription events."""
+        ctx = session_context["value"]
+
+        async for event in ctx.receive():
+            if isinstance(event, Start):
+                with lyrics_column:
+                    elem = ui.element("div").classes("utterance-item utterance-current slide-up")
+                    with elem:
+                        ui.label("")
+                current_utterance_element["element"] = elem
+
+            elif isinstance(event, Delta):
+                elem = current_utterance_element["element"]
+
+                if elem is None:
+                    with lyrics_column:
+                        elem = ui.element("div").classes(
+                            "utterance-item utterance-current slide-up"
+                        )
+                        with elem:
+                            ui.label("")
+                    current_utterance_element["element"] = elem
+
+                elem.clear()
+                with elem:
+                    ui.label(event.text)
+
+                if not event.interim and event.sequence is not None:
+                    current_sequence["value"] = event.sequence
+                    elem.classes(remove="utterance-current")
+                    elem.classes(add="utterance-past")
+
+                # JavaScript MutationObserver handles auto-scroll
+
+            elif isinstance(event, Done):
+                elem = current_utterance_element["element"]
+                if elem and current_sequence["value"] > 0:
+                    vpr_and_render_task = asyncio.create_task(
+                        wait_for_vpr_and_render(elem, current_sequence["value"], event.full_text)
+                    )
+                    asyncio_tasks.setdefault(
+                        f"vpr_render_{current_sequence['value']}", vpr_and_render_task
+                    )
+                    current_utterance_element["element"] = None
+
+    @handle_errors
+    async def toggle_recording():
+        """Toggle recording state."""
+        if not is_recording["value"]:
+            if is_session_completed["value"] and session_id_state["value"]:
+                # Continue existing session
+                record_btn.props("loading")
+                try:
+                    ctx = await session_service.session(session_id_state["value"])
+                    session_context["value"] = ctx
+                    await ctx.start()
+
+                    is_recording["value"] = True
+                    is_session_completed["value"] = False
+
+                    record_btn.props("icon=stop color=negative")
+                    record_btn.classes(add="recording-pulse")
+
+                    ui.notify("继续录音", type="positive")
+
+                    transcription_task = asyncio.create_task(process_transcription())
+                    asyncio_tasks.setdefault("transcription", transcription_task)
+
+                finally:
+                    record_btn.props(remove="loading")
+
+            else:
+                # Create new session
+                with (
+                    ui.dialog() as dialog,
+                    ui.card().style("width: 550px; padding: 32px; border-radius: 20px;"),
+                ):
+                    ui.label("创建录音会话").classes("text-h5 font-bold mb-2 text-grey-9")
+                    ui.label("填写患者信息以开始录音").classes("text-body2 text-grey-7 mb-6")
+
+                    with ui.row().classes("w-full gap-4"):
+                        patient_name = (
+                            ui.input("", placeholder="患者姓名")
+                            .classes("flex-1 clean-input")
+                            .props("standout dense hide-bottom-space")
+                        )
+
+                        clinic_number = (
+                            ui.input("", placeholder="门诊号")
+                            .classes("flex-1 clean-input")
+                            .props("standout dense hide-bottom-space")
+                        )
+
+                    with ui.row().classes("w-full gap-4 mt-3"):
+                        medical_record = (
+                            ui.input("", placeholder="病历号")
+                            .classes("flex-1 clean-input")
+                            .props("standout dense hide-bottom-space")
+                        )
+
+                        diagnosis = (
+                            ui.input("", placeholder="诊断")
+                            .classes("flex-1 clean-input")
+                            .props("standout dense hide-bottom-space")
+                        )
+
+                    notes = (
+                        ui.textarea("", placeholder="备注")
+                        .classes("w-full mt-3 clean-input notes-textarea")
+                        .props("standout hide-bottom-space")
+                    )
+
+                    @handle_errors
+                    async def do_start():
+                        dialog.close()
+                        record_btn.props("loading")
+
+                        try:
+                            session = await session_service.create(
+                                CreateSessionCommand(
+                                    doctor_id=doctor.id,
+                                    patient_name=patient_name.value.strip() or None,
+                                    clinic_number=clinic_number.value.strip() or None,
+                                    medical_record_number=medical_record.value.strip() or None,
+                                    diagnosis=diagnosis.value.strip() or None,
+                                    notes=notes.value.strip() or None,
+                                )
+                            )
+
+                            session_id_state["value"] = session.id
+                            is_session_completed["value"] = False
+
+                            ctx = await session_service.session(session_id_state["value"])
+                            session_context["value"] = ctx
+                            await ctx.start()
+
+                            is_recording["value"] = True
+
+                            record_btn.props("icon=stop color=negative")
+                            record_btn.classes(add="recording-pulse")
+
+                            ui.notify("开始录音", type="positive")
+
+                            transcription_task = asyncio.create_task(process_transcription())
+                            asyncio_tasks.setdefault("transcription", transcription_task)
+
+                        finally:
+                            record_btn.props(remove="loading")
+
+                    with ui.row().classes("w-full justify-end gap-2 mt-6"):
+                        ui.button("取消", on_click=dialog.close).props("flat no-caps").classes(
+                            "action-button"
+                        )
+                        ui.button("开始录音", on_click=do_start).props(
+                            "unelevated color=primary size=lg no-caps"
+                        ).classes("action-button").style("height: 48px;")
+
+                dialog.open()
+
+        else:
+            # Stop recording
+            record_btn.props("loading")
+
+            try:
+                ctx = session_context["value"]
+                await ctx.close()
+
+                await session_service.complete(session_id_state["value"])
+
+                is_recording["value"] = False
+                is_session_completed["value"] = True
+
+                record_btn.props("icon=mic color=primary")
+                record_btn.classes(remove="recording-pulse")
+
+                ui.notify("录音已保存，可继续添加内容", type="positive", timeout=3000)
+
+            finally:
+                record_btn.props(remove="loading")
+
+    # Recording button
+    record_btn = (
+        ui.button(icon="mic", on_click=toggle_recording)
+        .props("round unelevated color=primary size=xl")
+        .classes("record-btn")
+    )
+
+    # Load existing utterances on page load
+    await load_existing_utterances()

audex/view/pages/register.py

@@ -0,0 +1,202 @@
+from __future__ import annotations
+
+from dependency_injector.wiring import Provide
+from dependency_injector.wiring import inject
+from fastapi import Depends
+from nicegui import ui
+
+from audex.config import Config
+from audex.container import Container
+from audex.exceptions import PermissionDeniedError
+from audex.exceptions import ValidationError
+from audex.service.doctor import DoctorService
+from audex.service.doctor.types import RegisterCommand
+from audex.valueobj.common.auth import Password
+from audex.valueobj.common.email import Email
+from audex.valueobj.common.phone import CNPhone
+from audex.view.decorators import handle_errors
+
+
+@ui.page("/register")
+@handle_errors
+@inject
+async def render(
+    doctor_service: DoctorService = Depends(Provide[Container.service.doctor]),
+    config: Config = Depends(Provide[Container.config]),
+) -> None:
+    """Render registration page with clean design."""
+
+    # Check if already logged in
+    try:
+        await doctor_service.current_doctor()
+        ui.navigate.to("/")
+        return
+    except PermissionDeniedError:
+        pass  # Not logged in, continue
+
+    # Add consistent CSS (same as login)
+    ui.add_head_html('<link rel="stylesheet" href="/static/css/register.css">')
+
+    # Full screen container - allow scrolling
+    with (
+        (
+            ui.element("div")
+            .classes("w-full bg-white")
+            .style(
+                "min-height: 100vh; display: flex; align-items: center; justify-content: center; padding: 40px 20px;"
+            )
+        ),
+        ui.card().classes("register-card").style("width: 480px; padding: 32px 36px;"),
+    ):  # Register card
+        # Logo
+        ui.image("/static/images/logo.svg").classes("mx-auto mb-3 register-logo")
+
+        # Title
+        ui.label("创建账号").classes("gradient-title text-h5 text-center w-full mb-1")
+        ui.label(f"注册 {config.core.app.app_name} 医生账号").classes(
+            "text-sm text-grey-7 text-center w-full mb-5"
+        )
+
+        # Required fields section
+        ui.label("基本信息").classes("text-subtitle2 font-semibold text-grey-8 mb-3")
+
+        eid_input = (
+            ui.input("", placeholder="工号 *")
+            .classes("w-full mb-3 clean-input")
+            .props("standout dense hide-bottom-space")
+        )
+
+        name_input = (
+            ui.input("", placeholder="姓名 *")
+            .classes("w-full mb-3 clean-input")
+            .props("standout dense hide-bottom-space")
+        )
+
+        password_input = (
+            ui.input(
+                "",
+                placeholder="密码 *",
+                password=True,
+                password_toggle_button=True,
+            )
+            .classes("w-full mb-3 clean-input")
+            .props("standout dense hide-bottom-space")
+        )
+
+        password2_input = (
+            ui.input(
+                "",
+                placeholder="确认密码 *",
+                password=True,
+                password_toggle_button=True,
+            )
+            .classes("w-full mb-3 clean-input")
+            .props("standout dense hide-bottom-space")
+        )
+
+        # Optional fields section
+        ui.separator().classes("my-4")
+        ui.label("其他信息（选填）").classes("text-subtitle2 font-semibold text-grey-8 mb-3")
+
+        department_input = (
+            ui.input("", placeholder="科室")
+            .classes("w-full mb-3 clean-input")
+            .props("standout dense hide-bottom-space")
+        )
+
+        title_input = (
+            ui.input("", placeholder="职称")
+            .classes("w-full mb-3 clean-input")
+            .props("standout dense hide-bottom-space")
+        )
+
+        hospital_input = (
+            ui.input("", placeholder="医院")
+            .classes("w-full mb-3 clean-input")
+            .props("standout dense hide-bottom-space")
+        )
+
+        phone_input = (
+            ui.input("", placeholder="手机号")
+            .classes("w-full mb-3 clean-input")
+            .props("standout dense hide-bottom-space")
+        )
+
+        email_input = (
+            ui.input("", placeholder="邮箱")
+            .classes("w-full mb-3 clean-input")
+            .props("standout dense hide-bottom-space")
+        )
+
+        @handle_errors
+        async def do_register() -> None:
+            """Handle registration action."""
+            # Validate required fields
+            if not eid_input.value.strip():
+                ui.notify("请输入工号", type="warning", position="top")
+                return
+
+            if not name_input.value.strip():
+                ui.notify("请输入姓名", type="warning", position="top")
+                return
+
+            if not password_input.value:
+                ui.notify("请输入密码", type="warning", position="top")
+                return
+
+            if password_input.value != password2_input.value:
+                ui.notify("两次密码不一致", type="warning", position="top")
+                return
+
+            # Parse optional fields
+            phone = None
+            if phone_input.value.strip():
+                try:
+                    phone_str: str = phone_input.value.strip()
+                    if not phone_str.startswith("+86 "):
+                        phone_str = "+86 " + phone_str
+                    phone = CNPhone.parse(phone_str)
+                except ValidationError:
+                    ui.notify("手机号格式不正确", type="warning", position="top")
+                    return
+
+            email = None
+            if email_input.value.strip():
+                try:
+                    email = Email.parse(email_input.value.strip())
+                except ValidationError:
+                    ui.notify("邮箱格式不正确", type="warning", position="top")
+                    return
+
+            # Call service to register
+            await doctor_service.register(
+                RegisterCommand(
+                    eid=eid_input.value.strip(),
+                    password=Password.parse(password_input.value),
+                    name=name_input.value.strip(),
+                    department=department_input.value.strip() or None,
+                    title=title_input.value.strip() or None,
+                    hospital=hospital_input.value.strip() or None,
+                    phone=phone,
+                    email=email,
+                )
+            )
+
+            ui.notify("注册成功", type="positive", position="top")
+            ui.navigate.to("/")
+
+        # Register button
+        ui.button("注册", on_click=do_register).props(
+            "unelevated color=primary size=lg no-caps"
+        ).classes("w-full register-button mt-4").style("height: 48px;")
+
+        # Divider
+        with ui.row().classes("w-full items-center gap-4 my-3"):
+            ui.separator().classes("flex-1")
+            ui.label("或").classes("text-xs text-grey-6")
+            ui.separator().classes("flex-1")
+
+        # Back to login button
+        ui.button("返回登录", on_click=lambda: ui.navigate.to("/login")).props(
+            "flat size=lg no-caps"
+        ).classes("w-full back-button").style("height: 48px;")