Files
maternal-app/maternal-web/components/voice/VoiceFloatingButton.tsx
Andrei 962d0fb5ed
Some checks failed
CI/CD Pipeline / Lint and Test (push) Has been cancelled
CI/CD Pipeline / E2E Tests (push) Has been cancelled
CI/CD Pipeline / Build Application (push) Has been cancelled
fix: Transform sleep voice command duration to startTime/endTime format
Voice classification extracts duration in minutes, but sleep tracker expects
startTime/endTime. Added transformation logic to convert duration to proper
time range for sleep activities.

- Convert duration (minutes) to startTime + endTime timestamps
- Set default quality='good' and location='crib' if not specified
- Remove duration field after transformation

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-10-04 08:00:11 +00:00

567 lines
20 KiB
TypeScript

'use client';
import React, { useState } from 'react';
import {
Fab,
Tooltip,
Snackbar,
Alert,
Dialog,
DialogTitle,
DialogContent,
DialogActions,
Button,
Box,
Typography,
CircularProgress,
Chip,
IconButton,
Select,
MenuItem,
FormControl,
InputLabel,
} from '@mui/material';
import MicIcon from '@mui/icons-material/Mic';
import MicOffIcon from '@mui/icons-material/MicOff';
import AddIcon from '@mui/icons-material/Add';
import { useRouter } from 'next/navigation';
import { useVoiceInput } from '@/hooks/useVoiceInput';
import { useAuth } from '@/lib/auth/AuthContext';
import { trackingApi } from '@/lib/api/tracking';
import { childrenApi } from '@/lib/api/children';
import { voiceApi } from '@/lib/api/voice';
import { VoiceActivityReview } from './VoiceActivityReview';
/**
* Floating voice input button
*
* Always visible floating action button for quick voice commands.
* Positioned in bottom-right corner for easy thumb access.
*/
export function VoiceFloatingButton() {
const router = useRouter();
const { user } = useAuth();
const [open, setOpen] = useState(false);
const [isProcessing, setIsProcessing] = useState(false);
const [processingStatus, setProcessingStatus] = useState<'listening' | 'understanding' | 'saving' | null>(null);
const [identifiedActivity, setIdentifiedActivity] = useState<string>('');
const [classificationResult, setClassificationResult] = useState<any>(null);
const [processedClassificationId, setProcessedClassificationId] = useState<string | null>(null);
const [showReview, setShowReview] = useState(false);
const [showUnknownDialog, setShowUnknownDialog] = useState(false);
const [manualTrackingType, setManualTrackingType] = useState<string>('feeding');
const [snackbar, setSnackbar] = useState<{
open: boolean;
message: string;
severity: 'success' | 'info' | 'warning' | 'error';
}>({
open: false,
message: '',
severity: 'info',
});
const familyId = user?.families?.[0]?.familyId;
const { isListening, isSupported, transcript, classification, error, usesFallback, startListening, stopListening, reset } =
useVoiceInput();
// Set status when listening starts/stops
React.useEffect(() => {
console.log('[VoiceFloatingButton] isListening changed:', isListening, 'processingStatus:', processingStatus);
if (isListening && processingStatus !== 'listening') {
console.log('[VoiceFloatingButton] Setting status to listening');
setProcessingStatus('listening');
} else if (!isListening && processingStatus === 'listening') {
// When listening stops, transition to understanding
console.log('[VoiceFloatingButton] Setting status to understanding');
setProcessingStatus('understanding');
}
}, [isListening, processingStatus]);
// Auto-use classification from backend when transcription completes
// MediaRecorder sends audio to backend, which transcribes + classifies in one call
React.useEffect(() => {
// Create a unique ID for this classification based on transcript + type + timestamp
const classificationId = classification
? `${transcript}-${classification.type}-${classification.timestamp}`
: null;
// Only process if we haven't already processed this exact classification
if (classification && !isListening && !isProcessing && open && classificationId !== processedClassificationId) {
console.log('[Voice] New classification detected, showing review...', classificationId);
setProcessedClassificationId(classificationId);
setClassificationResult(classification);
// Show review dialog instead of immediately creating activity
if (classification.type !== 'unknown' && classification.confidence >= 0.3) {
setProcessingStatus(null);
setShowReview(true);
} else {
// For unknown or low confidence, show unknown dialog
setProcessingStatus(null);
setOpen(false);
setShowUnknownDialog(true);
}
}
}, [classification, isListening, isProcessing, open, transcript, processedClassificationId]);
const handleOpen = () => {
if (!isSupported) {
setSnackbar({
open: true,
message: 'Voice input not supported in this browser. Please use Chrome, Edge, or Safari.',
severity: 'error',
});
return;
}
setOpen(true);
reset();
setClassificationResult(null);
setProcessingStatus(null);
setIdentifiedActivity('');
setProcessedClassificationId(null);
// Auto-start listening after dialog opens
setTimeout(() => {
startListening();
}, 300);
};
const handleClose = () => {
if (isListening) {
stopListening();
}
setOpen(false);
reset();
setClassificationResult(null);
setProcessingStatus(null);
setIdentifiedActivity('');
setProcessedClassificationId(null);
};
const handleStartListening = () => {
reset();
setClassificationResult(null);
setProcessedClassificationId(null);
startListening();
};
const handleStopListening = () => {
stopListening();
};
const createActivity = async (activityType: string, activityDetails: Record<string, any>, activityTimestamp?: Date) => {
// Get the first child from the family
if (!familyId) {
console.log('[Voice] No familyId found');
throw new Error('No family found. Please set up your profile first.');
}
console.log('[Voice] Family ID:', familyId);
// Fetch children
console.log('[Voice] Fetching children for family:', familyId);
const children = await childrenApi.getChildren(familyId);
console.log('[Voice] Children found:', children.length, children);
if (children.length === 0) {
throw new Error('No children found. Please add a child first.');
}
// Use the first child
const childId = children[0].id;
console.log('[Voice] Using child ID:', childId);
// Transform sleep activity data - convert duration to startTime/endTime
let transformedDetails = { ...activityDetails };
const startTimestamp = activityTimestamp ? new Date(activityTimestamp) : new Date();
if (activityType === 'sleep' && activityDetails.duration) {
const durationMinutes = Number(activityDetails.duration);
const endTime = new Date(startTimestamp.getTime() + durationMinutes * 60 * 1000);
transformedDetails = {
...activityDetails,
startTime: startTimestamp.toISOString(),
endTime: endTime.toISOString(),
quality: activityDetails.quality || 'good',
location: activityDetails.location || 'crib',
};
// Remove duration as it's now represented by startTime/endTime
delete transformedDetails.duration;
console.log('[Voice] Transformed sleep data - duration', durationMinutes, 'minutes to startTime/endTime');
}
// Create the activity - use frontend API format (trackingApi transforms to backend DTO)
const activityData = {
type: activityType,
timestamp: startTimestamp.toISOString(),
data: transformedDetails,
notes: activityDetails.notes || undefined,
};
console.log('[Voice] Creating activity with data:', JSON.stringify(activityData, null, 2));
const createdActivity = await trackingApi.createActivity(childId, activityData);
console.log('[Voice] Activity created successfully:', createdActivity);
return { childId, activity: createdActivity };
};
const saveFeedback = async (action: 'approved' | 'edited' | 'rejected', finalData?: Record<string, any>, childId?: string, activityId?: string) => {
try {
await voiceApi.saveFeedback({
childId,
activityId,
transcript,
language: classificationResult?.language || 'en',
extractedType: classificationResult.type,
extractedData: classificationResult.details,
confidence: classificationResult.confidence,
action,
finalType: action === 'edited' ? classificationResult.type : undefined,
finalData: action === 'edited' ? finalData : undefined,
});
console.log(`[Voice] Feedback saved: ${action}`);
} catch (error) {
console.error('[Voice] Failed to save feedback:', error);
// Don't throw - feedback is nice-to-have, not critical
}
};
const handleApprove = async (data: any) => {
try {
setIsProcessing(true);
setIdentifiedActivity(data.type); // Set the activity type for display
setProcessingStatus('saving');
setShowReview(false);
const { childId, activity } = await createActivity(data.type, data.details, data.timestamp);
// Save feedback
await saveFeedback('approved', undefined, childId, activity.id);
// Show success message
const activityLabel = data.type.charAt(0).toUpperCase() + data.type.slice(1);
setSnackbar({
open: true,
message: `${activityLabel} activity saved successfully!`,
severity: 'success',
});
// Auto-close dialog
setTimeout(() => {
handleClose();
}, 1500);
} catch (error: any) {
console.error('[Voice] Failed to create activity:', error);
setSnackbar({
open: true,
message: error.message || 'Failed to save activity. Please try again.',
severity: 'error',
});
} finally {
setIsProcessing(false);
setProcessingStatus(null);
}
};
const handleEdit = async (editedData: any) => {
try {
setIsProcessing(true);
setIdentifiedActivity(editedData.type); // Set the activity type for display
setProcessingStatus('saving');
setShowReview(false);
const { childId, activity } = await createActivity(editedData.type, editedData.details, editedData.timestamp);
// Save feedback with edited data
await saveFeedback('edited', editedData.details, childId, activity.id);
// Show success message
const activityLabel = editedData.type.charAt(0).toUpperCase() + editedData.type.slice(1);
setSnackbar({
open: true,
message: `${activityLabel} activity saved with your edits!`,
severity: 'success',
});
// Auto-close dialog
setTimeout(() => {
handleClose();
}, 1500);
} catch (error: any) {
console.error('[Voice] Failed to create edited activity:', error);
setSnackbar({
open: true,
message: error.message || 'Failed to save activity. Please try again.',
severity: 'error',
});
} finally {
setIsProcessing(false);
setProcessingStatus(null);
}
};
const handleReject = async () => {
try {
setShowReview(false);
// Save feedback
await saveFeedback('rejected');
setSnackbar({
open: true,
message: 'Voice command rejected. Try again or use manual entry.',
severity: 'info',
});
} catch (error) {
console.error('[Voice] Failed to save rejection feedback:', error);
}
};
const handleCloseSnackbar = () => {
setSnackbar(prev => ({ ...prev, open: false }));
};
const handleRetry = () => {
setShowUnknownDialog(false);
setOpen(true);
reset();
setClassificationResult(null);
setProcessingStatus(null);
setProcessedClassificationId(null);
// Auto-start listening
setTimeout(() => {
startListening();
}, 300);
};
const handleManualTracking = () => {
setShowUnknownDialog(false);
router.push(`/track/${manualTrackingType}`);
};
return (
<>
{/* Floating button positioned in bottom-right - Hidden on desktop since we have TabBar center button */}
<Tooltip title="Voice Command (Beta)" placement="left">
<Fab
color="primary"
aria-label="voice input"
onClick={handleOpen}
disabled={!isSupported}
sx={{
position: 'fixed',
bottom: 24,
right: 24,
zIndex: 1000,
display: 'none', // Hidden - using TabBar center button instead
}}
>
<MicIcon />
</Fab>
</Tooltip>
{/* Voice input dialog */}
<Dialog
open={open}
onClose={handleClose}
maxWidth="sm"
fullWidth
aria-labelledby="voice-dialog-title"
aria-describedby="voice-dialog-status"
>
<DialogTitle id="voice-dialog-title">
Voice Command
{classificationResult && !classificationResult.error && (
<Chip
label={`${classificationResult.type || classificationResult.intent} (${classificationResult.confidenceLevel || Math.round((classificationResult.confidence || 0) * 100) + '%'})`}
color="success"
size="small"
sx={{ ml: 2 }}
aria-label={`Detected activity: ${classificationResult.type || classificationResult.intent}, confidence ${classificationResult.confidenceLevel || Math.round((classificationResult.confidence || 0) * 100) + ' percent'}`}
/>
)}
</DialogTitle>
<DialogContent>
<Box sx={{ textAlign: 'center', py: 3 }}>
{/* Microphone animation */}
<Box sx={{ position: 'relative', display: 'inline-block', mb: 3 }}>
<IconButton
color={isListening ? 'error' : 'primary'}
onClick={isListening ? handleStopListening : handleStartListening}
aria-label={isListening ? 'Stop listening' : 'Start listening'}
aria-pressed={isListening}
sx={{
width: 80,
height: 80,
bgcolor: isListening ? 'error.light' : 'primary.light',
'&:hover': {
bgcolor: isListening ? 'error.main' : 'primary.main',
},
animation: isListening ? 'pulse 1.5s infinite' : 'none',
'@keyframes pulse': {
'0%': { transform: 'scale(1)', opacity: 1 },
'50%': { transform: 'scale(1.1)', opacity: 0.8 },
'100%': { transform: 'scale(1)', opacity: 1 },
},
}}
>
{isListening ? <MicIcon sx={{ fontSize: 48 }} aria-hidden="true" /> : <MicOffIcon sx={{ fontSize: 48 }} aria-hidden="true" />}
</IconButton>
</Box>
{/* Status text with detailed processing stages */}
<Typography variant="body1" color="text.secondary" gutterBottom id="voice-dialog-status" role="status" aria-live="polite">
{processingStatus === 'listening' && 'Listening... Speak now'}
{processingStatus === 'understanding' && 'Understanding your request...'}
{processingStatus === 'saving' && identifiedActivity && `Adding to ${identifiedActivity.charAt(0).toUpperCase() + identifiedActivity.slice(1)} tracker...`}
{!processingStatus && !isListening && 'Click the microphone to start'}
</Typography>
{/* Transcript */}
{transcript && (
<Box sx={{ mt: 3, p: 2, bgcolor: 'grey.100', borderRadius: 1 }}>
<Typography variant="body2" color="text.secondary" gutterBottom>
Transcript:
</Typography>
<Typography variant="body1">{transcript}</Typography>
</Box>
)}
{/* Processing indicator with status */}
{processingStatus && (
<Box sx={{ mt: 2, display: 'flex', alignItems: 'center', justifyContent: 'center' }} role="status" aria-live="polite">
<CircularProgress size={20} sx={{ mr: 1 }} aria-hidden="true" />
<Typography variant="body2" color="text.secondary">
{processingStatus === 'listening' && 'Listening...'}
{processingStatus === 'understanding' && 'Understanding...'}
{processingStatus === 'saving' && 'Saving...'}
</Typography>
</Box>
)}
{/* Classification result */}
{classificationResult && !classificationResult.error && (
<Alert severity="success" sx={{ mt: 2 }} role="status">
<Typography variant="body2" gutterBottom>
<strong>Understood:</strong> {classificationResult.type || classificationResult.intent}
</Typography>
</Alert>
)}
{/* Error messages */}
{(error || (classificationResult && classificationResult.error)) && (
<Alert severity="error" sx={{ mt: 2 }} role="alert">
{error || classificationResult.message}
</Alert>
)}
{/* Examples */}
{!transcript && !isListening && (
<Box sx={{ mt: 3, textAlign: 'left' }}>
<Typography variant="caption" color="text.secondary" gutterBottom display="block">
Example commands:
</Typography>
<Typography variant="caption" color="text.secondary" component="div">
"Fed baby 120 ml"
<br />
"Nursed on left breast for 15 minutes"
<br />
"Changed wet diaper"
<br />
"Baby napped for 45 minutes"
</Typography>
</Box>
)}
</Box>
</DialogContent>
<DialogActions>
<Button onClick={handleClose}>Cancel</Button>
</DialogActions>
</Dialog>
{/* Review Dialog */}
{showReview && classificationResult && (
<VoiceActivityReview
open={showReview}
transcript={transcript}
classification={classificationResult}
onApprove={handleApprove}
onEdit={handleEdit}
onReject={handleReject}
onClose={() => setShowReview(false)}
/>
)}
{/* Unknown Intent Dialog */}
<Dialog
open={showUnknownDialog}
onClose={() => setShowUnknownDialog(false)}
maxWidth="sm"
fullWidth
aria-labelledby="unknown-command-dialog-title"
aria-describedby="unknown-command-dialog-description"
>
<DialogTitle id="unknown-command-dialog-title">Could Not Understand Command</DialogTitle>
<DialogContent>
<Box sx={{ mb: 3 }} id="unknown-command-dialog-description">
<Typography variant="body2" color="text.secondary" gutterBottom>
You said: "{transcript}"
</Typography>
<Typography variant="body2" sx={{ mt: 2 }}>
I couldn't identify a specific activity from your command. You can either try again or manually add an activity.
</Typography>
</Box>
<FormControl fullWidth sx={{ mt: 2 }}>
<InputLabel id="activity-type-label">Activity Type</InputLabel>
<Select
value={manualTrackingType}
onChange={(e) => setManualTrackingType(e.target.value)}
label="Activity Type"
labelId="activity-type-label"
inputProps={{
'aria-label': 'Select activity type for manual tracking',
}}
>
<MenuItem value="feeding">Feeding</MenuItem>
<MenuItem value="sleep">Sleep</MenuItem>
<MenuItem value="diaper">Diaper Change</MenuItem>
<MenuItem value="medicine">Medicine</MenuItem>
<MenuItem value="activity">Activity</MenuItem>
</Select>
</FormControl>
</DialogContent>
<DialogActions>
<Button onClick={handleRetry} startIcon={<MicIcon />} color="primary">
Retry Voice Command
</Button>
<Button onClick={handleManualTracking} startIcon={<AddIcon />} variant="contained">
Add Manual Tracking
</Button>
</DialogActions>
</Dialog>
{/* Snackbar for feedback */}
<Snackbar
open={snackbar.open}
autoHideDuration={3000}
onClose={handleCloseSnackbar}
anchorOrigin={{ vertical: 'bottom', horizontal: 'center' }}
>
<Alert onClose={handleCloseSnackbar} severity={snackbar.severity} sx={{ width: '100%' }}>
{snackbar.message}
</Alert>
</Snackbar>
</>
);
}