mirror of
https://github.com/R0m1k3/noteflow.git
synced 2026-10-15 19:27:53 +02:00
feat: add smart tag auto-suggestions for notes
- Create smartTags utility with keyword extraction algorithm - Filter out 100+ French stop words for better keyword quality - Implement category-based tag suggestions (travail, dev, design, etc.) - Suggest tags from similar notes based on content similarity - Display smart suggestions with one-click add in note editor - Show sparkle icon to indicate AI-assisted suggestions - Combine multiple suggestion strategies for best results - Limit to top 8 most relevant suggestions per note
This commit is contained in:
2 files changed
+209
-1
No files matched your search
@@ -0,0 +1,179 @@
|
||||
import { Note } from "@/services/NotesService";
|
||||
|
||||
// Common French stop words to filter out
|
||||
const FRENCH_STOP_WORDS = new Set([
|
||||
'le', 'la', 'les', 'un', 'une', 'des', 'de', 'du', 'et', 'ou', 'mais', 'donc',
|
||||
'or', 'ni', 'car', 'est', 'sont', 'être', 'avoir', 'a', 'dans', 'sur', 'pour',
|
||||
'par', 'avec', 'sans', 'sous', 'vers', 'chez', 'depuis', 'pendant', 'avant',
|
||||
'après', 'entre', 'parmi', 'selon', 'dont', 'que', 'qui', 'quoi', 'où', 'quand',
|
||||
'comment', 'pourquoi', 'ce', 'cet', 'cette', 'ces', 'mon', 'ton', 'son', 'ma',
|
||||
'ta', 'sa', 'mes', 'tes', 'ses', 'notre', 'votre', 'leur', 'nos', 'vos', 'leurs',
|
||||
'je', 'tu', 'il', 'elle', 'nous', 'vous', 'ils', 'elles', 'on', 'me', 'te', 'se',
|
||||
'lui', 'leur', 'y', 'en', 'ne', 'pas', 'plus', 'moins', 'très', 'trop', 'assez',
|
||||
'beaucoup', 'peu', 'bien', 'mal', 'tout', 'toute', 'tous', 'toutes', 'même',
|
||||
'autre', 'autres', 'tel', 'telle', 'quel', 'quelle', 'quelque', 'plusieurs',
|
||||
'chaque', 'aucun', 'aucune', 'nul', 'nulle', 'certain', 'certaine'
|
||||
]);
|
||||
|
||||
// Category keywords for auto-tagging
|
||||
const CATEGORY_KEYWORDS: Record<string, string[]> = {
|
||||
'travail': ['réunion', 'projet', 'équipe', 'client', 'présentation', 'deadline', 'objectif', 'tâche', 'mission', 'entreprise', 'bureau', 'collègue', 'manager'],
|
||||
'personnel': ['maison', 'famille', 'ami', 'vacances', 'loisir', 'hobby', 'détente', 'week-end', 'anniversaire', 'fête'],
|
||||
'développement': ['code', 'bug', 'feature', 'api', 'database', 'frontend', 'backend', 'git', 'deploy', 'test', 'debug', 'javascript', 'python', 'react', 'node'],
|
||||
'design': ['ui', 'ux', 'interface', 'maquette', 'prototype', 'couleur', 'police', 'layout', 'wireframe', 'mockup', 'figma', 'sketch'],
|
||||
'finance': ['budget', 'dépense', 'économie', 'investissement', 'facture', 'paiement', 'compte', 'banque', 'argent', 'revenu'],
|
||||
'santé': ['sport', 'exercice', 'nutrition', 'médecin', 'santé', 'bien-être', 'fitness', 'régime', 'yoga', 'course'],
|
||||
'formation': ['cours', 'apprentissage', 'étude', 'formation', 'certifica', 'examen', 'leçon', 'tutoriel', 'documentation', 'lecture'],
|
||||
'idée': ['brainstorming', 'innovation', 'créativité', 'concept', 'inspiration', 'réflexion', 'hypothèse', 'proposition'],
|
||||
'urgent': ['urgent', 'important', 'prioritaire', 'asap', 'deadline', 'critique', 'essentiel'],
|
||||
'meeting': ['réunion', 'rendez-vous', 'appel', 'visio', 'conférence', 'présentation', 'démonstration']
|
||||
};
|
||||
|
||||
/**
|
||||
* Extract keywords from text content
|
||||
*/
|
||||
export function extractKeywords(text: string, limit = 10): string[] {
|
||||
if (!text) return [];
|
||||
|
||||
// Remove HTML tags
|
||||
const cleanText = text.replace(/<[^>]*>/g, ' ');
|
||||
|
||||
// Convert to lowercase and split into words
|
||||
const words = cleanText
|
||||
.toLowerCase()
|
||||
.replace(/[^\w\sàâäæçéèêëïîôùûüÿœ]/g, ' ')
|
||||
.split(/\s+/)
|
||||
.filter(word => word.length > 3) // Min 4 chars
|
||||
.filter(word => !FRENCH_STOP_WORDS.has(word));
|
||||
|
||||
// Count word frequency
|
||||
const frequency = new Map<string, number>();
|
||||
words.forEach(word => {
|
||||
frequency.set(word, (frequency.get(word) || 0) + 1);
|
||||
});
|
||||
|
||||
// Sort by frequency and take top N
|
||||
return Array.from(frequency.entries())
|
||||
.sort((a, b) => b[1] - a[1])
|
||||
.slice(0, limit)
|
||||
.map(([word]) => word);
|
||||
}
|
||||
|
||||
/**
|
||||
* Suggest category tags based on content
|
||||
*/
|
||||
export function suggestCategoryTags(text: string): string[] {
|
||||
if (!text) return [];
|
||||
|
||||
const cleanText = text.toLowerCase().replace(/<[^>]*>/g, ' ');
|
||||
const suggestions: string[] = [];
|
||||
|
||||
for (const [category, keywords] of Object.entries(CATEGORY_KEYWORDS)) {
|
||||
// Check if any category keyword appears in the text
|
||||
const matches = keywords.filter(keyword =>
|
||||
cleanText.includes(keyword.toLowerCase())
|
||||
);
|
||||
|
||||
// If multiple matches, suggest this category
|
||||
if (matches.length >= 2) {
|
||||
suggestions.push(category);
|
||||
}
|
||||
}
|
||||
|
||||
return suggestions;
|
||||
}
|
||||
|
||||
/**
|
||||
* Suggest tags based on existing tags from similar notes
|
||||
*/
|
||||
export function suggestFromSimilarNotes(
|
||||
currentNote: Note,
|
||||
allNotes: Note[],
|
||||
limit = 5
|
||||
): string[] {
|
||||
if (!currentNote.title && !currentNote.content) return [];
|
||||
|
||||
const currentKeywords = extractKeywords(
|
||||
`${currentNote.title || ''} ${currentNote.content || ''}`,
|
||||
15
|
||||
);
|
||||
|
||||
if (currentKeywords.length === 0) return [];
|
||||
|
||||
// Find notes with similar keywords
|
||||
const noteSimilarity = allNotes
|
||||
.filter(note => note.id !== currentNote.id && note.tags && note.tags.length > 0)
|
||||
.map(note => {
|
||||
const noteKeywords = extractKeywords(
|
||||
`${note.title || ''} ${note.content || ''}`,
|
||||
15
|
||||
);
|
||||
|
||||
// Calculate similarity (common keywords / total unique keywords)
|
||||
const commonKeywords = currentKeywords.filter(kw =>
|
||||
noteKeywords.includes(kw)
|
||||
).length;
|
||||
|
||||
const totalKeywords = new Set([...currentKeywords, ...noteKeywords]).size;
|
||||
const similarity = totalKeywords > 0 ? commonKeywords / totalKeywords : 0;
|
||||
|
||||
return {
|
||||
note,
|
||||
similarity
|
||||
};
|
||||
})
|
||||
.filter(item => item.similarity > 0.2) // Min 20% similarity
|
||||
.sort((a, b) => b.similarity - a.similarity)
|
||||
.slice(0, 5); // Top 5 similar notes
|
||||
|
||||
// Extract tags from similar notes
|
||||
const tagFrequency = new Map<string, number>();
|
||||
noteSimilarity.forEach(({ note }) => {
|
||||
note.tags?.forEach(tag => {
|
||||
const tagName = tag.name.toLowerCase();
|
||||
tagFrequency.set(tagName, (tagFrequency.get(tagName) || 0) + 1);
|
||||
});
|
||||
});
|
||||
|
||||
// Sort by frequency and return top tags
|
||||
return Array.from(tagFrequency.entries())
|
||||
.sort((a, b) => b[1] - a[1])
|
||||
.slice(0, limit)
|
||||
.map(([tag]) => tag);
|
||||
}
|
||||
|
||||
/**
|
||||
* Get all tag suggestions for a note
|
||||
*/
|
||||
export function getSmartTagSuggestions(
|
||||
note: Note,
|
||||
allNotes: Note[],
|
||||
existingTags: string[] = []
|
||||
): string[] {
|
||||
const content = `${note.title || ''} ${note.content || ''}`;
|
||||
|
||||
// Get category suggestions
|
||||
const categoryTags = suggestCategoryTags(content);
|
||||
|
||||
// Get suggestions from similar notes
|
||||
const similarNoteTags = suggestFromSimilarNotes(note, allNotes);
|
||||
|
||||
// Get keyword-based suggestions
|
||||
const keywords = extractKeywords(content, 5);
|
||||
|
||||
// Combine all suggestions
|
||||
const allSuggestions = [
|
||||
...categoryTags,
|
||||
...similarNoteTags,
|
||||
...keywords
|
||||
];
|
||||
|
||||
// Filter out existing tags and duplicates
|
||||
const existingTagsLower = existingTags.map(t => t.toLowerCase());
|
||||
const uniqueSuggestions = Array.from(
|
||||
new Set(allSuggestions.map(t => t.toLowerCase()))
|
||||
).filter(tag => !existingTagsLower.includes(tag));
|
||||
|
||||
// Return top 8 suggestions
|
||||
return uniqueSuggestions.slice(0, 8);
|
||||
}
|
||||
Reference in new issue
Block a user