Add advanced TF-IDF search with autocomplete, query operators, facets, pagination, and accent normalization

2026-03-23 12:09:46 -04:00 · 2026-03-23 12:09:46 -04:00 · e171a0dc35
commit e171a0dc35
parent ba6271b89b
6 changed files with 1965 additions and 46 deletions
--- a/README.md
+++ b/README.md
@ -46,7 +46,10 @@

 - **🗂️ Multi-vault** : Visualisez plusieurs vaults Obsidian simultanément
 - **🌳 Navigation arborescente** : Parcourez vos dossiers et fichiers dans la sidebar
- **🔍 Recherche fulltext** : Recherche instantanée dans le contenu et les titres
+- **🔍 Recherche avancée** : Moteur TF-IDF avec normalisation des accents, snippets surlignés, facettes, pagination et tri
+- **💡 Autocomplétion intelligente** : Suggestions de fichiers, tags et historique avec navigation clavier
+- **🧩 Syntaxe de requête** : Opérateurs `tag:`, `#`, `vault:`, `title:`, `path:` avec chips visuels
+- **📜 Historique de recherche** : Persisté en localStorage (max 50 entrées, LIFO, dédupliqué)
 - **🏷️ Tag cloud** : Filtrage par tags extraits des frontmatters YAML
 - **🔗 Wikilinks** : Les `[[liens internes]]` Obsidian sont cliquables
 - **🖼️ Images Obsidian** : Support complet des syntaxes d'images Obsidian avec résolution intelligente
@ -347,7 +350,10 @@ ObsiGate expose une API REST complète :
 | `/api/file/{vault}/download?path=` | Téléchargement d'un fichier | GET |
 | `/api/file/{vault}/save?path=` | Sauvegarder un fichier | PUT |
 | `/api/file/{vault}?path=` | Supprimer un fichier | DELETE |
-| `/api/search?q=&vault=&tag=` | Recherche fulltext | GET |
+| `/api/search?q=&vault=&tag=` | Recherche fulltext (legacy) | GET |
+| `/api/search/advanced?q=&vault=&tag=&limit=&offset=&sort=` | Recherche avancée TF-IDF avec facettes et pagination | GET |
+| `/api/suggest?q=&vault=&limit=` | Suggestions de titres de fichiers (autocomplétion) | GET |
+| `/api/tags/suggest?q=&vault=&limit=` | Suggestions de tags (autocomplétion) | GET |
 | `/api/tags?vault=` | Tags uniques avec compteurs | GET |
 | `/api/index/reload` | Force un re-scan des vaults | GET |
 | `/api/image/{vault}?path=` | Servir une image avec MIME type approprié | GET |
@ -364,15 +370,63 @@ curl http://localhost:2020/api/health
 # Lister les vaults
 curl http://localhost:2020/api/vaults

-# Rechercher
+# Recherche simple (legacy)
 curl "http://localhost:2020/api/search?q=recette&vault=all"

+# Recherche avancée avec TF-IDF, facettes et pagination
+curl "http://localhost:2020/api/search/advanced?q=recette%20tag:cuisine&vault=all&limit=20&offset=0&sort=relevance"
+
+# Autocomplétion de titres
+curl "http://localhost:2020/api/suggest?q=piz&vault=all"
+
+# Autocomplétion de tags
+curl "http://localhost:2020/api/tags/suggest?q=rec&vault=all"
+
 # Obtenir un fichier
 curl "http://localhost:2020/api/file/Recettes?path=pizza.md"
 ```

 ---

+## 🔍 Recherche avancée
+
+### Syntaxe de requête
+
+| Opérateur | Description | Exemple |
+|-----------|-------------|---------|
+| `tag:<nom>` | Filtrer par tag | `tag:recette docker` |
+| `#<nom>` | Raccourci tag | `#linux serveur` |
+| `vault:<nom>` | Filtrer par vault | `vault:IT kubernetes` |
+| `title:<texte>` | Filtrer par titre | `title:pizza` |
+| `path:<texte>` | Filtrer par chemin | `path:recettes/soupes` |
+| `"phrase exacte"` | Recherche de phrase | `tag:"multi mots"` |
+
+Les opérateurs sont combinables : `tag:linux vault:IT serveur web` recherche "serveur web" dans le vault IT avec le tag linux.
+
+### Raccourcis clavier
+
+| Raccourci | Action |
+|-----------|--------|
+| `Ctrl+K` / `Cmd+K` | Focaliser la barre de recherche |
+| `/` | Focaliser la recherche (hors champ texte) |
+| `↑` / `↓` | Naviguer dans les suggestions |
+| `Enter` | Sélectionner la suggestion active ou lancer la recherche |
+| `Escape` | Fermer les suggestions / quitter la recherche |
+
+### Fonctionnalités
+
+- **TF-IDF** : Scoring basé sur la fréquence des termes pondérée par l'inverse de la fréquence documentaire
+- **Boost titre** : Les correspondances dans le titre reçoivent un score 3× supérieur
+- **Normalisation des accents** : `resume` trouve `résumé`, `elephant` trouve `éléphant`
+- **Snippets surlignés** : Les termes trouvés sont encadrés par `<mark>` dans les extraits
+- **Facettes** : Compteurs par vault et par tag dans les résultats
+- **Pagination** : Navigation par pages de 50 résultats
+- **Tri** : Par pertinence (TF-IDF) ou par date de modification
+- **Chips visuels** : Les filtres actifs sont affichés comme des chips colorés supprimables
+- **Historique** : Les 50 dernières recherches sont mémorisées en localStorage
+
+---
+
 ## 🔧 Dépannage

 ### Problèmes courants
--- a/backend/main.py
+++ b/backend/main.py
@ -25,7 +25,7 @@ from backend.indexer import (
    _extract_tags,
    SUPPORTED_EXTENSIONS,
 )
-from backend.search import search, get_all_tags
+from backend.search import search, get_all_tags, advanced_search, suggest_titles, suggest_tags
 from backend.image_processor import preprocess_images
 from backend.attachment_indexer import rescan_vault_attachments, get_attachment_stats

@ -141,6 +141,57 @@ class TreeSearchResponse(BaseModel):
    results: List[TreeSearchResult]


+class AdvancedSearchResultItem(BaseModel):
+    """A single advanced search result with highlighted snippet."""
+    vault: str
+    path: str
+    title: str
+    tags: List[str]
+    score: float
+    snippet: str
+    modified: str
+
+
+class SearchFacets(BaseModel):
+    """Faceted counts for search results."""
+    tags: Dict[str, int] = Field(default_factory=dict)
+    vaults: Dict[str, int] = Field(default_factory=dict)
+
+
+class AdvancedSearchResponse(BaseModel):
+    """Advanced search response with TF-IDF scoring, facets, and pagination."""
+    results: List[AdvancedSearchResultItem]
+    total: int
+    offset: int
+    limit: int
+    facets: SearchFacets
+
+
+class TitleSuggestion(BaseModel):
+    """A file title suggestion for autocomplete."""
+    vault: str
+    path: str
+    title: str
+
+
+class SuggestResponse(BaseModel):
+    """Autocomplete suggestions for file titles."""
+    query: str
+    suggestions: List[TitleSuggestion]
+
+
+class TagSuggestion(BaseModel):
+    """A tag suggestion for autocomplete."""
+    tag: str
+    count: int
+
+
+class TagSuggestResponse(BaseModel):
+    """Autocomplete suggestions for tags."""
+    query: str
+    suggestions: List[TagSuggestion]
+
+
 class ReloadResponse(BaseModel):
    """Index reload confirmation with per-vault stats."""
    status: str
@ -711,6 +762,84 @@ async def api_tree_search(
    return {"query": q, "vault_filter": vault, "results": results}


+@app.get("/api/search/advanced", response_model=AdvancedSearchResponse)
+async def api_advanced_search(
+    q: str = Query("", description="Advanced search query (supports tag:, vault:, title:, path: operators)"),
+    vault: str = Query("all", description="Vault filter"),
+    tag: Optional[str] = Query(None, description="Comma-separated tag filter"),
+    limit: int = Query(50, ge=1, le=200, description="Results per page"),
+    offset: int = Query(0, ge=0, description="Pagination offset"),
+    sort: str = Query("relevance", description="Sort by 'relevance' or 'modified'"),
+):
+    """Advanced full-text search with TF-IDF scoring, facets, and pagination.
+
+    Supports advanced query operators:
+    - ``tag:<name>`` or ``#<name>`` — filter by tag
+    - ``vault:<name>`` — filter by vault
+    - ``title:<text>`` — filter by title substring
+    - ``path:<text>`` — filter by path substring
+    - Remaining text is scored using TF-IDF with accent normalization.
+
+    Results include ``<mark>``-highlighted snippets and faceted tag/vault counts.
+
+    Args:
+        q: Query string with optional operators.
+        vault: Vault name or ``"all"``.
+        tag: Extra comma-separated tag names to require.
+        limit: Max results per page (1–200).
+        offset: Pagination offset.
+        sort: ``"relevance"`` (TF-IDF) or ``"modified"`` (date).
+
+    Returns:
+        ``AdvancedSearchResponse`` with scored results, facets, and pagination info.
+    """
+    return advanced_search(q, vault_filter=vault, tag_filter=tag, limit=limit, offset=offset, sort_by=sort)
+
+
+@app.get("/api/suggest", response_model=SuggestResponse)
+async def api_suggest(
+    q: str = Query("", description="Prefix to search for in file titles"),
+    vault: str = Query("all", description="Vault filter"),
+    limit: int = Query(10, ge=1, le=50, description="Max suggestions"),
+):
+    """Suggest file titles matching a prefix (accent-insensitive).
+
+    Used for autocomplete in the search input.
+
+    Args:
+        q: User-typed prefix (minimum 2 characters).
+        vault: Vault name or ``"all"``.
+        limit: Max number of suggestions.
+
+    Returns:
+        ``SuggestResponse`` with matching file title suggestions.
+    """
+    suggestions = suggest_titles(q, vault_filter=vault, limit=limit)
+    return {"query": q, "suggestions": suggestions}
+
+
+@app.get("/api/tags/suggest", response_model=TagSuggestResponse)
+async def api_tags_suggest(
+    q: str = Query("", description="Prefix to search for in tags"),
+    vault: str = Query("all", description="Vault filter"),
+    limit: int = Query(10, ge=1, le=50, description="Max suggestions"),
+):
+    """Suggest tags matching a prefix (accent-insensitive).
+
+    Used for autocomplete when typing ``tag:`` or ``#`` in the search input.
+
+    Args:
+        q: User-typed prefix (with or without ``#``, minimum 2 characters).
+        vault: Vault name or ``"all"``.
+        limit: Max number of suggestions.
+
+    Returns:
+        ``TagSuggestResponse`` with matching tag suggestions and counts.
+    """
+    suggestions = suggest_tags(q, vault_filter=vault, limit=limit)
+    return {"query": q, "suggestions": suggestions}
+
+
@app.get("/api/index/reload", response_model=ReloadResponse)
 async def api_reload():
    """Force a full re-index of all configured vaults.
--- a/backend/search.py
+++ b/backend/search.py
@ -1,14 +1,70 @@
 import logging
-from typing import List, Dict, Any, Optional
+import math
+import re
+import unicodedata
+from collections import defaultdict
+from typing import List, Dict, Any, Optional, Tuple

 from backend.indexer import index

 logger = logging.getLogger("obsigate.search")

-# Default maximum number of search results returned
+# ---------------------------------------------------------------------------
+# Constants
+# ---------------------------------------------------------------------------
 DEFAULT_SEARCH_LIMIT = 200
+ADVANCED_SEARCH_DEFAULT_LIMIT = 50
+SNIPPET_CONTEXT_CHARS = 120
+MAX_SNIPPET_HIGHLIGHTS = 5
+TITLE_BOOST = 3.0          # TF-IDF multiplier for title matches
+PATH_BOOST = 1.5           # TF-IDF multiplier for path matches
+TAG_BOOST = 2.0            # TF-IDF multiplier for tag matches
+MIN_PREFIX_LENGTH = 2      # Minimum chars for prefix matching
+SUGGEST_LIMIT = 10         # Default max suggestions returned
+
+# Regex to tokenize text into alphanumeric words (Unicode-aware)
+_WORD_RE = re.compile(r"[\w]+", re.UNICODE)


+# ---------------------------------------------------------------------------
+# Accent / Unicode normalization helpers
+# ---------------------------------------------------------------------------
+def normalize_text(text: str) -> str:
+    """Normalize text for accent-insensitive comparison.
+
+    Decomposes Unicode characters (NFD), strips combining diacritical marks,
+    then lowercases the result.  For example ``"Éléphant"`` → ``"elephant"``.
+
+    Args:
+        text: Raw input string.
+
+    Returns:
+        Lowercased, accent-stripped string.
+    """
+    if not text:
+        return ""
+    # NFD decomposition splits base char + combining mark
+    nfkd = unicodedata.normalize("NFKD", text)
+    # Strip combining marks (category "Mn" = Mark, Nonspacing)
+    stripped = "".join(ch for ch in nfkd if unicodedata.category(ch) != "Mn")
+    return stripped.lower()
+
+
+def tokenize(text: str) -> List[str]:
+    """Split text into normalized tokens (accent-stripped, lowercased words).
+
+    Args:
+        text: Raw text to tokenize.
+
+    Returns:
+        List of normalized word tokens.
+    """
+    return _WORD_RE.findall(normalize_text(text))
+
+
+# ---------------------------------------------------------------------------
+# Tag filter helper (unchanged for backward compat)
+# ---------------------------------------------------------------------------
 def _normalize_tag_filter(tag_filter: Optional[str]) -> List[str]:
    """Parse a comma-separated tag filter string into a clean list.

@ -25,7 +81,10 @@ def _normalize_tag_filter(tag_filter: Optional[str]) -> List[str]:
    return [tag.strip().lstrip("#") for tag in tag_filter.split(",") if tag.strip()]


-def _extract_snippet(content: str, query: str, context_chars: int = 120) -> str:
+# ---------------------------------------------------------------------------
+# Snippet extraction helpers
+# ---------------------------------------------------------------------------
+def _extract_snippet(content: str, query: str, context_chars: int = SNIPPET_CONTEXT_CHARS) -> str:
    """Extract a text snippet around the first occurrence of *query*.

    Returns up to ``context_chars`` characters before and after the match.
@ -57,6 +116,263 @@ def _extract_snippet(content: str, query: str, context_chars: int = 120) -> str:
    return snippet


+def _extract_highlighted_snippet(
+    content: str,
+    query_terms: List[str],
+    context_chars: int = SNIPPET_CONTEXT_CHARS,
+    max_highlights: int = MAX_SNIPPET_HIGHLIGHTS,
+) -> str:
+    """Extract a snippet and wrap matching terms in ``<mark>`` tags.
+
+    Performs accent-normalized matching so ``"resume"`` highlights ``"résumé"``.
+    Returns at most *max_highlights* highlighted regions to keep snippets concise.
+
+    Args:
+        content: Full text to search within.
+        query_terms: Normalized search terms.
+        context_chars: Number of context characters on each side.
+        max_highlights: Maximum highlighted regions.
+
+    Returns:
+        HTML snippet string with ``<mark>`` highlights.
+    """
+    if not content or not query_terms:
+        return content[:200].strip() if content else ""
+
+    norm_content = normalize_text(content)
+
+    # Find best position — first occurrence of any query term
+    best_pos = len(content)
+    for term in query_terms:
+        pos = norm_content.find(term)
+        if pos != -1 and pos < best_pos:
+            best_pos = pos
+
+    if best_pos == len(content):
+        # No match found — return beginning of content
+        return _escape_html(content[:200].strip())
+
+    start = max(0, best_pos - context_chars)
+    end = min(len(content), best_pos + context_chars + 40)
+    raw_snippet = content[start:end].strip()
+
+    prefix = "..." if start > 0 else ""
+    suffix = "..." if end < len(content) else ""
+
+    # Highlight all term occurrences in the snippet
+    highlighted = _highlight_terms(raw_snippet, query_terms, max_highlights)
+    return prefix + highlighted + suffix
+
+
+def _highlight_terms(text: str, terms: List[str], max_highlights: int) -> str:
+    """Wrap occurrences of *terms* in *text* with ``<mark>`` tags.
+
+    Uses accent-normalized comparison so diacritical variants are matched.
+    Escapes HTML in non-highlighted portions to prevent XSS.
+
+    Args:
+        text: Raw text snippet.
+        terms: Normalized search terms.
+        max_highlights: Cap on highlighted regions.
+
+    Returns:
+        HTML-safe string with ``<mark>`` wrapped matches.
+    """
+    if not terms or not text:
+        return _escape_html(text)
+
+    norm = normalize_text(text)
+    # Collect (start, end) spans for all term matches
+    spans: List[Tuple[int, int]] = []
+    for term in terms:
+        idx = 0
+        while idx < len(norm):
+            pos = norm.find(term, idx)
+            if pos == -1:
+                break
+            spans.append((pos, pos + len(term)))
+            idx = pos + 1
+
+    if not spans:
+        return _escape_html(text)
+
+    # Merge overlapping spans and limit count
+    spans.sort()
+    merged: List[Tuple[int, int]] = [spans[0]]
+    for s, e in spans[1:]:
+        if s <= merged[-1][1]:
+            merged[-1] = (merged[-1][0], max(merged[-1][1], e))
+        else:
+            merged.append((s, e))
+    merged = merged[:max_highlights]
+
+    # Build result with highlights
+    parts: List[str] = []
+    prev = 0
+    for s, e in merged:
+        if s > prev:
+            parts.append(_escape_html(text[prev:s]))
+        parts.append(f"<mark>{_escape_html(text[s:e])}</mark>")
+        prev = e
+    if prev < len(text):
+        parts.append(_escape_html(text[prev:]))
+
+    return "".join(parts)
+
+
+def _escape_html(text: str) -> str:
+    """Escape HTML special characters."""
+    return (
+        text.replace("&", "&amp;")
+        .replace("<", "&lt;")
+        .replace(">", "&gt;")
+        .replace('"', "&quot;")
+    )
+
+
+# ---------------------------------------------------------------------------
+# Inverted Index for TF-IDF
+# ---------------------------------------------------------------------------
+class InvertedIndex:
+    """In-memory inverted index supporting TF-IDF scoring.
+
+    Built lazily from the global ``index`` dict whenever a search or
+    suggestion request detects that the underlying vault index has changed.
+    The class is designed to be a singleton — use ``get_inverted_index()``.
+
+    Attributes:
+        word_index: ``{token: {doc_key: term_frequency}}``
+        title_index: ``{token: [doc_key, ...]}``
+        tag_norm_map: ``{normalized_tag: original_tag}``
+        tag_prefix_index: ``{prefix: [original_tag, ...]}``
+        doc_count: Total number of indexed documents.
+        _source_id: Fingerprint of the source index to detect staleness.
+    """
+
+    def __init__(self) -> None:
+        self.word_index: Dict[str, Dict[str, int]] = defaultdict(dict)
+        self.title_index: Dict[str, List[str]] = defaultdict(list)
+        self.tag_norm_map: Dict[str, str] = {}
+        self.tag_prefix_index: Dict[str, List[str]] = defaultdict(list)
+        self.title_norm_map: Dict[str, List[Dict[str, str]]] = defaultdict(list)
+        self.doc_count: int = 0
+        self._source_id: Optional[int] = None
+
+    def is_stale(self) -> bool:
+        """Check if the inverted index needs rebuilding."""
+        current_id = id(index)
+        return current_id != self._source_id
+
+    def rebuild(self) -> None:
+        """Rebuild inverted index from the global ``index`` dict.
+
+        Tokenizes titles and content of every file, computes term frequencies,
+        and builds auxiliary indexes for tag and title prefix suggestions.
+        """
+        logger.info("Rebuilding inverted index...")
+        self.word_index = defaultdict(dict)
+        self.title_index = defaultdict(list)
+        self.tag_norm_map = {}
+        self.tag_prefix_index = defaultdict(list)
+        self.title_norm_map = defaultdict(list)
+        self.doc_count = 0
+
+        for vault_name, vault_data in index.items():
+            for file_info in vault_data.get("files", []):
+                doc_key = f"{vault_name}::{file_info['path']}"
+                self.doc_count += 1
+
+                # --- Title tokens ---
+                title_tokens = tokenize(file_info.get("title", ""))
+                for token in set(title_tokens):
+                    self.title_index[token].append(doc_key)
+
+                # --- Normalized title for prefix suggestions ---
+                norm_title = normalize_text(file_info.get("title", ""))
+                if norm_title:
+                    self.title_norm_map[norm_title].append({
+                        "vault": vault_name,
+                        "path": file_info["path"],
+                        "title": file_info["title"],
+                    })
+
+                # --- Content tokens (including title for combined scoring) ---
+                content = file_info.get("content", "")
+                full_text = (file_info.get("title", "") + " " + content)
+                tokens = tokenize(full_text)
+                tf: Dict[str, int] = defaultdict(int)
+                for token in tokens:
+                    tf[token] += 1
+                for token, freq in tf.items():
+                    self.word_index[token][doc_key] = freq
+
+            # --- Tag indexes ---
+            for tag in vault_data.get("tags", {}):
+                norm_tag = normalize_text(tag)
+                self.tag_norm_map[norm_tag] = tag
+                # Build prefix entries for each prefix length ≥ MIN_PREFIX_LENGTH
+                for plen in range(MIN_PREFIX_LENGTH, len(norm_tag) + 1):
+                    prefix = norm_tag[:plen]
+                    if tag not in self.tag_prefix_index[prefix]:
+                        self.tag_prefix_index[prefix].append(tag)
+
+        self._source_id = id(index)
+        logger.info(
+            "Inverted index built: %d documents, %d unique tokens, %d tags",
+            self.doc_count,
+            len(self.word_index),
+            len(self.tag_norm_map),
+        )
+
+    def idf(self, term: str) -> float:
+        """Inverse Document Frequency for a term.
+
+        ``idf(t) = log(N / (1 + df(t)))`` where *df(t)* is the number
+        of documents containing term *t*.
+
+        Args:
+            term: Normalized term.
+
+        Returns:
+            IDF score (≥ 0).
+        """
+        df = len(self.word_index.get(term, {}))
+        if df == 0:
+            return 0.0
+        return math.log((self.doc_count + 1) / (1 + df))
+
+    def tf_idf(self, term: str, doc_key: str) -> float:
+        """TF-IDF score for a term in a document.
+
+        Uses raw term frequency (no log normalization) × IDF.
+
+        Args:
+            term: Normalized term.
+            doc_key: ``"vault::path"`` document key.
+
+        Returns:
+            TF-IDF score.
+        """
+        tf = self.word_index.get(term, {}).get(doc_key, 0)
+        if tf == 0:
+            return 0.0
+        return tf * self.idf(term)
+
+
+# Singleton inverted index
+_inverted_index = InvertedIndex()
+
+
+def get_inverted_index() -> InvertedIndex:
+    """Return the singleton inverted index, rebuilding if stale."""
+    if _inverted_index.is_stale():
+        _inverted_index.rebuild()
+    return _inverted_index
+
+
+# ---------------------------------------------------------------------------
+# Backward-compatible search (unchanged API)
+# ---------------------------------------------------------------------------
 def search(
    query: str,
    vault_filter: str = "all",
@ -155,6 +471,342 @@ def search(
    return results[:limit]


+# ---------------------------------------------------------------------------
+# Advanced search with TF-IDF scoring
+# ---------------------------------------------------------------------------
+def _parse_advanced_query(raw_query: str) -> Dict[str, Any]:
+    """Parse an advanced query string into structured filters and free text.
+
+    Supported operators:
+    - ``tag:<name>`` or ``#<name>`` — tag filter
+    - ``vault:<name>`` — vault filter
+    - ``title:<text>`` — title filter
+    - ``path:<text>`` — path filter
+    - Remaining tokens are treated as free-text search terms.
+
+    Args:
+        raw_query: Raw query string from the user.
+
+    Returns:
+        Dict with keys ``tags``, ``vault``, ``title``, ``path``, ``terms``.
+    """
+    parsed: Dict[str, Any] = {
+        "tags": [],
+        "vault": None,
+        "title": None,
+        "path": None,
+        "terms": [],
+    }
+    if not raw_query:
+        return parsed
+
+    # Use shlex-like tokenizing but handle quotes manually
+    tokens = _split_query_tokens(raw_query)
+    for token in tokens:
+        lower = token.lower()
+        if lower.startswith("tag:"):
+            tag_val = token[4:].strip().lstrip("#")
+            if tag_val:
+                parsed["tags"].append(tag_val)
+        elif lower.startswith("#") and len(token) > 1:
+            parsed["tags"].append(token[1:])
+        elif lower.startswith("vault:"):
+            parsed["vault"] = token[6:].strip()
+        elif lower.startswith("title:"):
+            parsed["title"] = token[6:].strip()
+        elif lower.startswith("path:"):
+            parsed["path"] = token[5:].strip()
+        else:
+            parsed["terms"].append(token)
+
+    return parsed
+
+
+def _split_query_tokens(raw: str) -> List[str]:
+    """Split a query string respecting quoted phrases.
+
+    ``tag:"my tag" hello world`` → ``['tag:my tag', 'hello', 'world']``
+
+    Args:
+        raw: Raw query string.
+
+    Returns:
+        List of token strings.
+    """
+    tokens: List[str] = []
+    i = 0
+    n = len(raw)
+    while i < n:
+        # Skip whitespace
+        while i < n and raw[i] == " ":
+            i += 1
+        if i >= n:
+            break
+
+        # Check for operator with quoted value, e.g., tag:"foo bar"
+        if i < n and raw[i] != '"':
+            # Read until space or quote
+            j = i
+            while j < n and raw[j] != " ":
+                if raw[j] == '"':
+                    # Read quoted portion
+                    j += 1
+                    while j < n and raw[j] != '"':
+                        j += 1
+                    if j < n:
+                        j += 1  # skip closing quote
+                else:
+                    j += 1
+            token = raw[i:j].replace('"', "")
+            tokens.append(token)
+            i = j
+        else:
+            # Quoted token
+            i += 1  # skip opening quote
+            j = i
+            while j < n and raw[j] != '"':
+                j += 1
+            tokens.append(raw[i:j])
+            i = j + 1  # skip closing quote
+
+    return tokens
+
+
+def advanced_search(
+    query: str,
+    vault_filter: str = "all",
+    tag_filter: Optional[str] = None,
+    limit: int = ADVANCED_SEARCH_DEFAULT_LIMIT,
+    offset: int = 0,
+    sort_by: str = "relevance",
+) -> Dict[str, Any]:
+    """Advanced full-text search with TF-IDF scoring, facets, and pagination.
+
+    Parses the query for operators (``tag:``, ``vault:``, ``title:``,
+    ``path:``), falls back remaining tokens to TF-IDF scored free-text
+    search using the inverted index.  Results include highlighted snippets
+    with ``<mark>`` tags and faceted counts for tags and vaults.
+
+    Args:
+        query: Raw query string (may include operators).
+        vault_filter: Vault name or ``"all"`` (overridden by ``vault:`` op).
+        tag_filter: Comma-separated tag names (merged with ``tag:`` ops).
+        limit: Max results per page.
+        offset: Pagination offset.
+        sort_by: ``"relevance"`` or ``"modified"``.
+
+    Returns:
+        Dict with ``results``, ``total``, ``offset``, ``limit``, ``facets``.
+    """
+    query = query.strip() if query else ""
+    parsed = _parse_advanced_query(query)
+
+    # Merge explicit tag_filter with parsed tag: operators
+    all_tags = list(parsed["tags"])
+    extra_tags = _normalize_tag_filter(tag_filter)
+    for t in extra_tags:
+        if t not in all_tags:
+            all_tags.append(t)
+
+    # Vault filter — parsed vault: overrides parameter
+    effective_vault = parsed["vault"] or vault_filter
+
+    # Normalize free-text terms
+    query_terms = [normalize_text(t) for t in parsed["terms"] if t.strip()]
+    has_terms = len(query_terms) > 0
+
+    if not has_terms and not all_tags and not parsed["title"] and not parsed["path"]:
+        return {"results": [], "total": 0, "offset": offset, "limit": limit, "facets": {"tags": {}, "vaults": {}}}
+
+    inv = get_inverted_index()
+    scored_results: List[Tuple[float, Dict[str, Any]]] = []
+    facet_tags: Dict[str, int] = defaultdict(int)
+    facet_vaults: Dict[str, int] = defaultdict(int)
+
+    for vault_name, vault_data in index.items():
+        if effective_vault != "all" and vault_name != effective_vault:
+            continue
+
+        for file_info in vault_data.get("files", []):
+            doc_key = f"{vault_name}::{file_info['path']}"
+
+            # --- Tag filter ---
+            if all_tags:
+                file_tags_lower = [t.lower() for t in file_info.get("tags", [])]
+                if not all(t.lower() in file_tags_lower for t in all_tags):
+                    continue
+
+            # --- Title filter ---
+            if parsed["title"]:
+                norm_title_filter = normalize_text(parsed["title"])
+                norm_file_title = normalize_text(file_info.get("title", ""))
+                if norm_title_filter not in norm_file_title:
+                    continue
+
+            # --- Path filter ---
+            if parsed["path"]:
+                norm_path_filter = normalize_text(parsed["path"])
+                norm_file_path = normalize_text(file_info.get("path", ""))
+                if norm_path_filter not in norm_file_path:
+                    continue
+
+            # --- Scoring ---
+            score = 0.0
+            if has_terms:
+                # TF-IDF scoring for each term
+                for term in query_terms:
+                    tfidf = inv.tf_idf(term, doc_key)
+                    score += tfidf
+
+                    # Title boost — check if term appears in title tokens
+                    norm_title = normalize_text(file_info.get("title", ""))
+                    if term in norm_title:
+                        score += tfidf * TITLE_BOOST
+
+                    # Path boost
+                    norm_path = normalize_text(file_info.get("path", ""))
+                    if term in norm_path:
+                        score += tfidf * PATH_BOOST
+
+                    # Tag boost
+                    for tag in file_info.get("tags", []):
+                        if term in normalize_text(tag):
+                            score += tfidf * TAG_BOOST
+                            break
+
+                # Also add prefix matching bonus for partial words
+                for term in query_terms:
+                    if len(term) >= MIN_PREFIX_LENGTH:
+                        for indexed_term, docs in inv.word_index.items():
+                            if indexed_term.startswith(term) and indexed_term != term:
+                                if doc_key in docs:
+                                    score += inv.tf_idf(indexed_term, doc_key) * 0.5
+            else:
+                # Filter-only search (tag/title/path): score = 1
+                score = 1.0
+
+            if score > 0:
+                # Build highlighted snippet
+                content = file_info.get("content", "")
+                if has_terms:
+                    snippet = _extract_highlighted_snippet(content, query_terms)
+                else:
+                    snippet = _escape_html(content[:200].strip()) if content else ""
+
+                result = {
+                    "vault": vault_name,
+                    "path": file_info["path"],
+                    "title": file_info["title"],
+                    "tags": file_info.get("tags", []),
+                    "score": round(score, 4),
+                    "snippet": snippet,
+                    "modified": file_info.get("modified", ""),
+                }
+                scored_results.append((score, result))
+
+                # Facets
+                facet_vaults[vault_name] = facet_vaults.get(vault_name, 0) + 1
+                for tag in file_info.get("tags", []):
+                    facet_tags[tag] = facet_tags.get(tag, 0) + 1
+
+    # Sort
+    if sort_by == "modified":
+        scored_results.sort(key=lambda x: x[1].get("modified", ""), reverse=True)
+    else:
+        scored_results.sort(key=lambda x: -x[0])
+
+    total = len(scored_results)
+    page = scored_results[offset: offset + limit]
+
+    return {
+        "results": [r for _, r in page],
+        "total": total,
+        "offset": offset,
+        "limit": limit,
+        "facets": {
+            "tags": dict(sorted(facet_tags.items(), key=lambda x: -x[1])[:20]),
+            "vaults": dict(sorted(facet_vaults.items(), key=lambda x: -x[1])),
+        },
+    }
+
+
+# ---------------------------------------------------------------------------
+# Suggestion helpers
+# ---------------------------------------------------------------------------
+def suggest_titles(
+    prefix: str,
+    vault_filter: str = "all",
+    limit: int = SUGGEST_LIMIT,
+) -> List[Dict[str, str]]:
+    """Suggest file titles matching a prefix (accent-insensitive).
+
+    Args:
+        prefix: User-typed prefix string.
+        vault_filter: Vault name or ``"all"``.
+        limit: Maximum suggestions.
+
+    Returns:
+        List of ``{"vault", "path", "title"}`` dicts.
+    """
+    if not prefix or len(prefix) < MIN_PREFIX_LENGTH:
+        return []
+
+    inv = get_inverted_index()
+    norm_prefix = normalize_text(prefix)
+    results: List[Dict[str, str]] = []
+    seen: set = set()
+
+    for norm_title, entries in inv.title_norm_map.items():
+        if norm_prefix in norm_title:
+            for entry in entries:
+                if vault_filter != "all" and entry["vault"] != vault_filter:
+                    continue
+                key = f"{entry['vault']}::{entry['path']}"
+                if key not in seen:
+                    seen.add(key)
+                    results.append(entry)
+                    if len(results) >= limit:
+                        return results
+
+    return results
+
+
+def suggest_tags(
+    prefix: str,
+    vault_filter: str = "all",
+    limit: int = SUGGEST_LIMIT,
+) -> List[Dict[str, Any]]:
+    """Suggest tags matching a prefix (accent-insensitive).
+
+    Args:
+        prefix: User-typed prefix (with or without leading ``#``).
+        vault_filter: Vault name or ``"all"``.
+        limit: Maximum suggestions.
+
+    Returns:
+        List of ``{"tag", "count"}`` dicts sorted by descending count.
+    """
+    prefix = prefix.lstrip("#").strip()
+    if not prefix or len(prefix) < MIN_PREFIX_LENGTH:
+        return []
+
+    norm_prefix = normalize_text(prefix)
+    all_tag_counts = get_all_tags(vault_filter)
+
+    matches: List[Dict[str, Any]] = []
+    for tag, count in all_tag_counts.items():
+        norm_tag = normalize_text(tag)
+        if norm_prefix in norm_tag:
+            matches.append({"tag": tag, "count": count})
+            if len(matches) >= limit:
+                break
+
+    return matches
+
+
+# ---------------------------------------------------------------------------
+# Backward-compatible tag aggregation (unchanged API)
+# ---------------------------------------------------------------------------
 def get_all_tags(vault_filter: Optional[str] = None) -> Dict[str, int]:
    """Aggregate tag counts across vaults, sorted by descending count.

--- a/frontend/app.js
+++ b/frontend/app.js
@ -25,6 +25,21 @@
  let activeSidebarTab = "vaults";
  let filterDebounce = null;

+  // Advanced search state
+  let advancedSearchOffset = 0;
+  let advancedSearchTotal = 0;
+  let advancedSearchSort = "relevance";
+  let advancedSearchLastQuery = "";
+  let suggestAbortController = null;
+  let dropdownActiveIndex = -1;
+  let dropdownItems = [];
+
+  // Advanced search constants
+  const SEARCH_HISTORY_KEY = "obsigate_search_history";
+  const MAX_HISTORY_ENTRIES = 50;
+  const SUGGEST_DEBOUNCE_MS = 150;
+  const ADVANCED_SEARCH_LIMIT = 50;
+
  // ---------------------------------------------------------------------------
  // File extension → Lucide icon mapping
  // ---------------------------------------------------------------------------
@ -75,6 +90,414 @@
    return EXT_ICONS[ext] || "file";
  }

+  // ---------------------------------------------------------------------------
+  // Search History Service (localStorage, LIFO, max 50, dedup)
+  // ---------------------------------------------------------------------------
+  const SearchHistory = {
+    _load() {
+      try {
+        const raw = localStorage.getItem(SEARCH_HISTORY_KEY);
+        return raw ? JSON.parse(raw) : [];
+      } catch { return []; }
+    },
+    _save(entries) {
+      try { localStorage.setItem(SEARCH_HISTORY_KEY, JSON.stringify(entries)); } catch {}
+    },
+    getAll() { return this._load(); },
+    add(query) {
+      if (!query || !query.trim()) return;
+      const q = query.trim();
+      let entries = this._load().filter(e => e !== q);
+      entries.unshift(q);
+      if (entries.length > MAX_HISTORY_ENTRIES) entries = entries.slice(0, MAX_HISTORY_ENTRIES);
+      this._save(entries);
+    },
+    remove(query) {
+      const entries = this._load().filter(e => e !== query);
+      this._save(entries);
+    },
+    clear() { this._save([]); },
+    filter(prefix) {
+      if (!prefix) return this.getAll().slice(0, 8);
+      const lp = prefix.toLowerCase();
+      return this._load().filter(e => e.toLowerCase().includes(lp)).slice(0, 8);
+    },
+  };
+
+  // ---------------------------------------------------------------------------
+  // Query Parser — extracts operators (tag:, #, vault:, title:, path:)
+  // ---------------------------------------------------------------------------
+  const QueryParser = {
+    parse(raw) {
+      const result = { tags: [], vault: null, title: null, path: null, freeText: "" };
+      if (!raw) return result;
+      const tokens = this._tokenize(raw);
+      const freeTokens = [];
+      for (const tok of tokens) {
+        const lower = tok.toLowerCase();
+        if (lower.startsWith("tag:")) {
+          const v = tok.slice(4).replace(/"/g, "").trim().replace(/^#/, "");
+          if (v) result.tags.push(v);
+        } else if (lower.startsWith("#") && tok.length > 1) {
+          result.tags.push(tok.slice(1));
+        } else if (lower.startsWith("vault:")) {
+          result.vault = tok.slice(6).replace(/"/g, "").trim();
+        } else if (lower.startsWith("title:")) {
+          result.title = tok.slice(6).replace(/"/g, "").trim();
+        } else if (lower.startsWith("path:")) {
+          result.path = tok.slice(5).replace(/"/g, "").trim();
+        } else {
+          freeTokens.push(tok);
+        }
+      }
+      result.freeText = freeTokens.join(" ");
+      return result;
+    },
+    _tokenize(raw) {
+      const tokens = [];
+      let i = 0;
+      const n = raw.length;
+      while (i < n) {
+        while (i < n && raw[i] === " ") i++;
+        if (i >= n) break;
+        if (raw[i] !== '"') {
+          let j = i;
+          while (j < n && raw[j] !== " ") {
+            if (raw[j] === '"') { j++; while (j < n && raw[j] !== '"') j++; if (j < n) j++; }
+            else j++;
+          }
+          tokens.push(raw.slice(i, j).replace(/"/g, ""));
+          i = j;
+        } else {
+          i++;
+          let j = i;
+          while (j < n && raw[j] !== '"') j++;
+          tokens.push(raw.slice(i, j));
+          i = j + 1;
+        }
+      }
+      return tokens;
+    },
+    /** Detect the current operator context at cursor for autocomplete */
+    getContext(raw, cursorPos) {
+      const before = raw.slice(0, cursorPos);
+      // Check if we're typing a tag: or # value
+      const tagMatch = before.match(/(?:tag:|#)([\w-]*)$/i);
+      if (tagMatch) return { type: "tag", prefix: tagMatch[1] };
+      // Check if typing title:
+      const titleMatch = before.match(/title:([\w-]*)$/i);
+      if (titleMatch) return { type: "title", prefix: titleMatch[1] };
+      // Default: free text
+      const words = before.trim().split(/\s+/);
+      const lastWord = words[words.length - 1] || "";
+      return { type: "text", prefix: lastWord };
+    },
+  };
+
+  // ---------------------------------------------------------------------------
+  // Autocomplete Dropdown Controller
+  // ---------------------------------------------------------------------------
+  const AutocompleteDropdown = {
+    _dropdown: null,
+    _historySection: null,
+    _titlesSection: null,
+    _tagsSection: null,
+    _historyList: null,
+    _titlesList: null,
+    _tagsList: null,
+    _emptyEl: null,
+    _suggestTimer: null,
+
+    init() {
+      this._dropdown = document.getElementById("search-dropdown");
+      this._historySection = document.getElementById("search-dropdown-history");
+      this._titlesSection = document.getElementById("search-dropdown-titles");
+      this._tagsSection = document.getElementById("search-dropdown-tags");
+      this._historyList = document.getElementById("search-dropdown-history-list");
+      this._titlesList = document.getElementById("search-dropdown-titles-list");
+      this._tagsList = document.getElementById("search-dropdown-tags-list");
+      this._emptyEl = document.getElementById("search-dropdown-empty");
+
+      // Clear history button
+      const clearBtn = document.getElementById("search-dropdown-clear-history");
+      if (clearBtn) {
+        clearBtn.addEventListener("click", (e) => {
+          e.stopPropagation();
+          SearchHistory.clear();
+          this.hide();
+        });
+      }
+
+      // Close dropdown on outside click
+      document.addEventListener("click", (e) => {
+        if (this._dropdown && !this._dropdown.contains(e.target) &&
+            e.target.id !== "search-input") {
+          this.hide();
+        }
+      });
+    },
+
+    show() {
+      if (this._dropdown) this._dropdown.hidden = false;
+    },
+
+    hide() {
+      if (this._dropdown) this._dropdown.hidden = true;
+      dropdownActiveIndex = -1;
+      dropdownItems = [];
+    },
+
+    isVisible() {
+      return this._dropdown && !this._dropdown.hidden;
+    },
+
+    /** Populate and show the dropdown with history, title suggestions, and tag suggestions */
+    async populate(inputValue, cursorPos) {
+      // Cancel previous suggestion request
+      if (suggestAbortController) { suggestAbortController.abort(); suggestAbortController = null; }
+
+      const ctx = QueryParser.getContext(inputValue, cursorPos);
+      const vault = document.getElementById("vault-filter").value;
+
+      // History — always show filtered history
+      const historyItems = SearchHistory.filter(inputValue);
+      this._renderHistory(historyItems, inputValue);
+
+      // Title and tag suggestions from API (debounced)
+      clearTimeout(this._suggestTimer);
+      if (ctx.prefix && ctx.prefix.length >= 2) {
+        this._suggestTimer = setTimeout(() => this._fetchSuggestions(ctx, vault, inputValue), SUGGEST_DEBOUNCE_MS);
+      } else {
+        this._renderTitles([], "");
+        this._renderTags([], "");
+      }
+
+      // Show/hide sections
+      const hasContent = historyItems.length > 0;
+      this._historySection.hidden = historyItems.length === 0;
+      this._emptyEl.hidden = hasContent;
+
+      if (hasContent || (ctx.prefix && ctx.prefix.length >= 2)) {
+        this.show();
+      } else if (!hasContent) {
+        this.hide();
+      }
+
+      this._collectItems();
+    },
+
+    async _fetchSuggestions(ctx, vault, inputValue) {
+      suggestAbortController = new AbortController();
+      try {
+        const [titlesRes, tagsRes] = await Promise.all([
+          ctx.type !== "tag" ? api(`/api/suggest?q=${encodeURIComponent(ctx.prefix)}&vault=${encodeURIComponent(vault)}&limit=8`, { signal: suggestAbortController.signal }) : Promise.resolve({ suggestions: [] }),
+          (ctx.type === "tag" || ctx.type === "text") ? api(`/api/tags/suggest?q=${encodeURIComponent(ctx.prefix)}&vault=${encodeURIComponent(vault)}&limit=6`, { signal: suggestAbortController.signal }) : Promise.resolve({ suggestions: [] }),
+        ]);
+
+        this._renderTitles(titlesRes.suggestions || [], ctx.prefix);
+        this._renderTags(tagsRes.suggestions || [], ctx.prefix);
+
+        // Update visibility
+        const hasTitles = (titlesRes.suggestions || []).length > 0;
+        const hasTags = (tagsRes.suggestions || []).length > 0;
+        this._titlesSection.hidden = !hasTitles;
+        this._tagsSection.hidden = !hasTags;
+
+        const historyVisible = !this._historySection.hidden;
+        const hasAny = historyVisible || hasTitles || hasTags;
+        this._emptyEl.hidden = hasAny;
+        if (hasAny) this.show(); else if (!historyVisible) this.hide();
+
+        this._collectItems();
+      } catch (err) {
+        if (err.name !== "AbortError") console.error("Suggestion fetch error:", err);
+      }
+    },
+
+    _renderHistory(items, query) {
+      this._historyList.innerHTML = "";
+      items.forEach((entry) => {
+        const li = el("li", { class: "search-dropdown__item search-dropdown__item--history", role: "option" });
+        const iconEl = el("span", { class: "search-dropdown__icon" });
+        iconEl.innerHTML = '<i data-lucide="clock" style="width:14px;height:14px"></i>';
+        const textEl = el("span", { class: "search-dropdown__text" });
+        textEl.textContent = entry;
+        li.appendChild(iconEl);
+        li.appendChild(textEl);
+        li.addEventListener("click", () => {
+          const input = document.getElementById("search-input");
+          input.value = entry;
+          this.hide();
+          _triggerAdvancedSearch(entry);
+        });
+        this._historyList.appendChild(li);
+      });
+    },
+
+    _renderTitles(items, prefix) {
+      this._titlesList.innerHTML = "";
+      items.forEach((item) => {
+        const li = el("li", { class: "search-dropdown__item search-dropdown__item--title", role: "option" });
+        const iconEl = el("span", { class: "search-dropdown__icon" });
+        iconEl.innerHTML = '<i data-lucide="file-text" style="width:14px;height:14px"></i>';
+        const textEl = el("span", { class: "search-dropdown__text" });
+        if (prefix) {
+          this._highlightText(textEl, item.title, prefix);
+        } else {
+          textEl.textContent = item.title;
+        }
+        const metaEl = el("span", { class: "search-dropdown__meta" });
+        metaEl.textContent = item.vault;
+        li.appendChild(iconEl);
+        li.appendChild(textEl);
+        li.appendChild(metaEl);
+        li.addEventListener("click", () => {
+          this.hide();
+          openFile(item.vault, item.path);
+        });
+        this._titlesList.appendChild(li);
+      });
+    },
+
+    _renderTags(items, prefix) {
+      this._tagsList.innerHTML = "";
+      items.forEach((item) => {
+        const li = el("li", { class: "search-dropdown__item search-dropdown__item--tag", role: "option" });
+        const iconEl = el("span", { class: "search-dropdown__icon" });
+        iconEl.innerHTML = '<i data-lucide="hash" style="width:14px;height:14px"></i>';
+        const textEl = el("span", { class: "search-dropdown__text" });
+        if (prefix) {
+          this._highlightText(textEl, item.tag, prefix);
+        } else {
+          textEl.textContent = item.tag;
+        }
+        const badge = el("span", { class: "search-dropdown__badge" });
+        badge.textContent = item.count;
+        li.appendChild(iconEl);
+        li.appendChild(textEl);
+        li.appendChild(badge);
+        li.addEventListener("click", () => {
+          const input = document.getElementById("search-input");
+          // Append tag: operator if not already typing one
+          const current = input.value;
+          const ctx = QueryParser.getContext(current, input.selectionStart);
+          if (ctx.type === "tag") {
+            // Replace the partial tag prefix
+            const before = current.slice(0, input.selectionStart - ctx.prefix.length);
+            input.value = before + item.tag + " ";
+          } else {
+            input.value = (current ? current + " " : "") + "tag:" + item.tag + " ";
+          }
+          this.hide();
+          input.focus();
+          _triggerAdvancedSearch(input.value);
+        });
+        this._tagsList.appendChild(li);
+      });
+    },
+
+    _highlightText(container, text, query) {
+      const lower = text.toLowerCase();
+      const needle = query.toLowerCase();
+      const pos = lower.indexOf(needle);
+      if (pos === -1) { container.textContent = text; return; }
+      container.appendChild(document.createTextNode(text.slice(0, pos)));
+      const markEl = el("mark", {}, [document.createTextNode(text.slice(pos, pos + query.length))]);
+      container.appendChild(markEl);
+      container.appendChild(document.createTextNode(text.slice(pos + query.length)));
+    },
+
+    _collectItems() {
+      dropdownItems = Array.from(this._dropdown.querySelectorAll(".search-dropdown__item"));
+      dropdownActiveIndex = -1;
+      dropdownItems.forEach(item => item.classList.remove("active"));
+    },
+
+    navigateDown() {
+      if (!this.isVisible() || dropdownItems.length === 0) return;
+      if (dropdownActiveIndex >= 0) dropdownItems[dropdownActiveIndex].classList.remove("active");
+      dropdownActiveIndex = (dropdownActiveIndex + 1) % dropdownItems.length;
+      dropdownItems[dropdownActiveIndex].classList.add("active");
+      dropdownItems[dropdownActiveIndex].scrollIntoView({ block: "nearest" });
+    },
+
+    navigateUp() {
+      if (!this.isVisible() || dropdownItems.length === 0) return;
+      if (dropdownActiveIndex >= 0) dropdownItems[dropdownActiveIndex].classList.remove("active");
+      dropdownActiveIndex = dropdownActiveIndex <= 0 ? dropdownItems.length - 1 : dropdownActiveIndex - 1;
+      dropdownItems[dropdownActiveIndex].classList.add("active");
+      dropdownItems[dropdownActiveIndex].scrollIntoView({ block: "nearest" });
+    },
+
+    selectActive() {
+      if (dropdownActiveIndex >= 0 && dropdownActiveIndex < dropdownItems.length) {
+        dropdownItems[dropdownActiveIndex].click();
+        return true;
+      }
+      return false;
+    },
+  };
+
+  // ---------------------------------------------------------------------------
+  // Search Chips Controller — renders active filter chips from parsed query
+  // ---------------------------------------------------------------------------
+  const SearchChips = {
+    _container: null,
+    init() { this._container = document.getElementById("search-chips"); },
+    update(parsed) {
+      if (!this._container) return;
+      this._container.innerHTML = "";
+      let hasChips = false;
+      parsed.tags.forEach(tag => { this._addChip("tag", `tag:${tag}`, tag); hasChips = true; });
+      if (parsed.vault) { this._addChip("vault", `vault:${parsed.vault}`, parsed.vault); hasChips = true; }
+      if (parsed.title) { this._addChip("title", `title:${parsed.title}`, parsed.title); hasChips = true; }
+      if (parsed.path) { this._addChip("path", `path:${parsed.path}`, parsed.path); hasChips = true; }
+      this._container.hidden = !hasChips;
+    },
+    clear() {
+      if (!this._container) return;
+      this._container.innerHTML = "";
+      this._container.hidden = true;
+    },
+    _addChip(type, fullOperator, displayText) {
+      const chip = el("span", { class: `search-chip search-chip--${type}` });
+      const label = el("span", { class: "search-chip__label" });
+      label.textContent = fullOperator;
+      const removeBtn = el("button", { class: "search-chip__remove", title: "Retirer ce filtre", type: "button" });
+      removeBtn.innerHTML = '<i data-lucide="x" style="width:10px;height:10px"></i>';
+      removeBtn.addEventListener("click", () => {
+        // Remove this operator from the input
+        const input = document.getElementById("search-input");
+        const raw = input.value;
+        // Remove the operator text from the query
+        const escaped = fullOperator.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
+        input.value = raw.replace(new RegExp("\\s*" + escaped + "\\s*", "i"), " ").trim();
+        _triggerAdvancedSearch(input.value);
+      });
+      chip.appendChild(label);
+      chip.appendChild(removeBtn);
+      this._container.appendChild(chip);
+      safeCreateIcons();
+    },
+  };
+
+  // ---------------------------------------------------------------------------
+  // Helper: trigger advanced search from input value
+  // ---------------------------------------------------------------------------
+  function _triggerAdvancedSearch(rawQuery) {
+    const q = (rawQuery || "").trim();
+    const vault = document.getElementById("vault-filter").value;
+    const tagFilter = selectedTags.length > 0 ? selectedTags.join(",") : null;
+    advancedSearchOffset = 0;
+    if (q.length > 0 || tagFilter) {
+      SearchHistory.add(q);
+      performAdvancedSearch(q, vault, tagFilter);
+    } else {
+      SearchChips.clear();
+      showWelcome();
+    }
+  }
+
  // ---------------------------------------------------------------------------
  // Safe CDN helpers
  // ---------------------------------------------------------------------------
@ -1076,7 +1499,7 @@
    } else {
      const input = document.getElementById("search-input");
      if (input.value.trim()) {
-        performSearch(input.value.trim(), document.getElementById("vault-filter").value, null);
+        performAdvancedSearch(input.value.trim(), document.getElementById("vault-filter").value, null);
      } else {
        showWelcome();
      }
@ -1087,7 +1510,7 @@
    const input = document.getElementById("search-input");
    const query = input.value.trim();
    const vault = document.getElementById("vault-filter").value;
-    performSearch(query, vault, selectedTags.length > 0 ? selectedTags.join(",") : null);
+    performAdvancedSearch(query, vault, selectedTags.length > 0 ? selectedTags.join(",") : null);
  }

  function buildSearchResultsHeader(data, query, tagFilter) {
@ -1535,77 +1958,348 @@
  }

  // ---------------------------------------------------------------------------
-  // Search
+  // Search (enhanced with autocomplete, keyboard nav, global shortcuts)
  // ---------------------------------------------------------------------------
  function initSearch() {
    const input = document.getElementById("search-input");
    const caseBtn = document.getElementById("search-case-btn");
    const clearBtn = document.getElementById("search-clear-btn");
-    
+
+    // Initialize sub-controllers
+    AutocompleteDropdown.init();
+    SearchChips.init();
+
    // Initially hide clear button
    clearBtn.style.display = "none";
-    
+
+    // --- Input handler: debounced search + autocomplete dropdown ---
    input.addEventListener("input", () => {
      const hasText = input.value.length > 0;
      clearBtn.style.display = hasText ? "flex" : "none";
-      
+
+      // Show autocomplete dropdown while typing
+      AutocompleteDropdown.populate(input.value, input.selectionStart);
+
+      // Debounced search execution
      clearTimeout(searchTimeout);
      searchTimeout = setTimeout(() => {
        const q = input.value.trim();
        const vault = document.getElementById("vault-filter").value;
        const tagFilter = selectedTags.length > 0 ? selectedTags.join(",") : null;
+        advancedSearchOffset = 0;
        if (q.length > 0 || tagFilter) {
-          performSearch(q, vault, tagFilter);
+          performAdvancedSearch(q, vault, tagFilter);
        } else {
+          SearchChips.clear();
          showWelcome();
        }
      }, 300);
    });
-    
+
+    // --- Focus handler: show history dropdown ---
+    input.addEventListener("focus", () => {
+      if (input.value.length === 0) {
+        const historyItems = SearchHistory.filter("");
+        if (historyItems.length > 0) {
+          AutocompleteDropdown.populate("", 0);
+        }
+      }
+    });
+
+    // --- Keyboard navigation in dropdown ---
+    input.addEventListener("keydown", (e) => {
+      if (AutocompleteDropdown.isVisible()) {
+        if (e.key === "ArrowDown") {
+          e.preventDefault();
+          AutocompleteDropdown.navigateDown();
+        } else if (e.key === "ArrowUp") {
+          e.preventDefault();
+          AutocompleteDropdown.navigateUp();
+        } else if (e.key === "Enter") {
+          if (AutocompleteDropdown.selectActive()) {
+            e.preventDefault();
+            return;
+          }
+          // No active item — execute search normally
+          AutocompleteDropdown.hide();
+          const q = input.value.trim();
+          if (q) {
+            SearchHistory.add(q);
+            clearTimeout(searchTimeout);
+            advancedSearchOffset = 0;
+            const vault = document.getElementById("vault-filter").value;
+            const tagFilter = selectedTags.length > 0 ? selectedTags.join(",") : null;
+            performAdvancedSearch(q, vault, tagFilter);
+          }
+          e.preventDefault();
+        } else if (e.key === "Escape") {
+          AutocompleteDropdown.hide();
+          e.stopPropagation();
+        }
+      } else if (e.key === "Enter") {
+        const q = input.value.trim();
+        if (q) {
+          SearchHistory.add(q);
+          clearTimeout(searchTimeout);
+          advancedSearchOffset = 0;
+          const vault = document.getElementById("vault-filter").value;
+          const tagFilter = selectedTags.length > 0 ? selectedTags.join(",") : null;
+          performAdvancedSearch(q, vault, tagFilter);
+        }
+        e.preventDefault();
+      }
+    });
+
    caseBtn.addEventListener("click", () => {
      searchCaseSensitive = !searchCaseSensitive;
      caseBtn.classList.toggle("active");
    });
-    
+
    clearBtn.addEventListener("click", () => {
      input.value = "";
      clearBtn.style.display = "none";
      searchCaseSensitive = false;
      caseBtn.classList.remove("active");
+      SearchChips.clear();
+      AutocompleteDropdown.hide();
      showWelcome();
    });
+
+    // --- Global keyboard shortcuts ---
+    document.addEventListener("keydown", (e) => {
+      // Ctrl+K or Cmd+K: focus search
+      if ((e.ctrlKey || e.metaKey) && e.key === "k") {
+        e.preventDefault();
+        input.focus();
+        input.select();
+      }
+      // "/" key: focus search (when not in an input/textarea)
+      if (e.key === "/" && !_isInputFocused()) {
+        e.preventDefault();
+        input.focus();
+      }
+      // Escape: blur search input and close dropdown
+      if (e.key === "Escape" && document.activeElement === input) {
+        AutocompleteDropdown.hide();
+        input.blur();
+      }
+    });
  }

+  /** Check if user is focused on an input/textarea/contenteditable */
+  function _isInputFocused() {
+    const tag = document.activeElement?.tagName;
+    if (tag === "INPUT" || tag === "TEXTAREA" || tag === "SELECT") return true;
+    return document.activeElement?.isContentEditable === true;
+  }
+
+  // --- Backward-compatible search (existing /api/search endpoint) ---
  async function performSearch(query, vaultFilter, tagFilter) {
-    // Cancel any in-flight search request
-    if (searchAbortController) {
-      searchAbortController.abort();
-    }
+    if (searchAbortController) searchAbortController.abort();
    searchAbortController = new AbortController();
-
    showLoading();
-
    let url = `/api/search?q=${encodeURIComponent(query)}&vault=${encodeURIComponent(vaultFilter)}`;
    if (tagFilter) url += `&tag=${encodeURIComponent(tagFilter)}`;
-
    try {
      const data = await api(url, { signal: searchAbortController.signal });
      renderSearchResults(data, query, tagFilter);
    } catch (err) {
-      if (err.name === "AbortError") return; // superseded by newer request
+      if (err.name === "AbortError") return;
      showWelcome();
    } finally {
      searchAbortController = null;
    }
  }

+  // --- Advanced search with TF-IDF, facets, pagination ---
+  async function performAdvancedSearch(query, vaultFilter, tagFilter, offset, sort) {
+    if (searchAbortController) searchAbortController.abort();
+    searchAbortController = new AbortController();
+    showLoading();
+
+    const ofs = offset !== undefined ? offset : advancedSearchOffset;
+    const sortBy = sort || advancedSearchSort;
+    advancedSearchLastQuery = query;
+
+    // Update chips from parsed query
+    const parsed = QueryParser.parse(query);
+    SearchChips.update(parsed);
+
+    let url = `/api/search/advanced?q=${encodeURIComponent(query)}&vault=${encodeURIComponent(vaultFilter)}&limit=${ADVANCED_SEARCH_LIMIT}&offset=${ofs}&sort=${sortBy}`;
+    if (tagFilter) url += `&tag=${encodeURIComponent(tagFilter)}`;
+
+    try {
+      const data = await api(url, { signal: searchAbortController.signal });
+      advancedSearchTotal = data.total;
+      advancedSearchOffset = ofs;
+      renderAdvancedSearchResults(data, query, tagFilter);
+    } catch (err) {
+      if (err.name === "AbortError") return;
+      showWelcome();
+    } finally {
+      searchAbortController = null;
+    }
+  }
+
+  // --- Legacy search results renderer (kept for backward compat) ---
  function renderSearchResults(data, query, tagFilter) {
    const area = document.getElementById("content-area");
    area.innerHTML = "";
-
    const header = buildSearchResultsHeader(data, query, tagFilter);
    area.appendChild(header);
+    if (data.results.length === 0) {
+      area.appendChild(el("p", { style: "color:var(--text-muted);margin-top:20px" }, [
+        document.createTextNode("Aucun résultat trouvé."),
+      ]));
+      return;
+    }
+    const container = el("div", { class: "search-results" });
+    data.results.forEach((r) => {
+      const titleDiv = el("div", { class: "search-result-title" });
+      if (query && query.trim()) {
+        highlightSearchText(titleDiv, r.title, query, searchCaseSensitive);
+      } else {
+        titleDiv.textContent = r.title;
+      }
+      const snippetDiv = el("div", { class: "search-result-snippet" });
+      if (query && query.trim() && r.snippet) {
+        highlightSearchText(snippetDiv, r.snippet, query, searchCaseSensitive);
+      } else {
+        snippetDiv.textContent = r.snippet || "";
+      }
+      const item = el("div", { class: "search-result-item" }, [
+        titleDiv,
+        el("div", { class: "search-result-vault" }, [document.createTextNode(r.vault + " / " + r.path)]),
+        snippetDiv,
+      ]);
+      if (r.tags && r.tags.length > 0) {
+        const tagsDiv = el("div", { class: "search-result-tags" });
+        r.tags.forEach((tag) => {
+          if (!TagFilterService.isTagFiltered(tag)) {
+            const tagEl = el("span", { class: "file-tag" }, [document.createTextNode(`#${tag}`)]);
+            tagEl.addEventListener("click", (e) => { e.stopPropagation(); addTagFilter(tag); });
+            tagsDiv.appendChild(tagEl);
+          }
+        });
+        if (tagsDiv.children.length > 0) item.appendChild(tagsDiv);
+      }
+      item.addEventListener("click", () => openFile(r.vault, r.path));
+      container.appendChild(item);
+    });
+    area.appendChild(container);
+  }

+  // --- Advanced search results renderer (facets, highlighted snippets, pagination, sort) ---
+  function renderAdvancedSearchResults(data, query, tagFilter) {
+    const area = document.getElementById("content-area");
+    area.innerHTML = "";
+
+    // Header with result count and sort controls
+    const header = el("div", { class: "search-results-header" });
+    const summaryText = el("span", { class: "search-results-summary-text" });
+    const parsed = QueryParser.parse(query);
+    const freeText = parsed.freeText;
+
+    if (freeText && tagFilter) {
+      summaryText.textContent = `${data.total} résultat(s) pour "${freeText}" avec filtres`;
+    } else if (freeText) {
+      summaryText.textContent = `${data.total} résultat(s) pour "${freeText}"`;
+    } else if (parsed.tags.length > 0 || tagFilter) {
+      summaryText.textContent = `${data.total} fichier(s) avec filtres`;
+    } else {
+      summaryText.textContent = `${data.total} résultat(s)`;
+    }
+    header.appendChild(summaryText);
+
+    // Sort controls
+    const sortDiv = el("div", { class: "search-sort" });
+    const btnRelevance = el("button", { class: "search-sort__btn" + (advancedSearchSort === "relevance" ? " active" : ""), type: "button" });
+    btnRelevance.textContent = "Pertinence";
+    btnRelevance.addEventListener("click", () => {
+      advancedSearchSort = "relevance";
+      advancedSearchOffset = 0;
+      const vault = document.getElementById("vault-filter").value;
+      performAdvancedSearch(query, vault, tagFilter, 0, "relevance");
+    });
+    const btnDate = el("button", { class: "search-sort__btn" + (advancedSearchSort === "modified" ? " active" : ""), type: "button" });
+    btnDate.textContent = "Date";
+    btnDate.addEventListener("click", () => {
+      advancedSearchSort = "modified";
+      advancedSearchOffset = 0;
+      const vault = document.getElementById("vault-filter").value;
+      performAdvancedSearch(query, vault, tagFilter, 0, "modified");
+    });
+    sortDiv.appendChild(btnRelevance);
+    sortDiv.appendChild(btnDate);
+    header.appendChild(sortDiv);
+    area.appendChild(header);
+
+    // Active sidebar tag chips
+    if (selectedTags.length > 0) {
+      const activeTags = el("div", { class: "search-results-active-tags" });
+      selectedTags.forEach((tag) => {
+        const removeBtn = el("button", {
+          class: "search-results-active-tag-remove",
+          title: `Retirer ${tag} du filtre`,
+        }, [document.createTextNode("×")]);
+        removeBtn.addEventListener("click", (e) => { e.stopPropagation(); removeTagFilter(tag); });
+        const chip = el("span", { class: "search-results-active-tag" }, [
+          document.createTextNode(`#${tag}`), removeBtn,
+        ]);
+        activeTags.appendChild(chip);
+      });
+      area.appendChild(activeTags);
+    }
+
+    // Facets panel
+    if (data.facets && (Object.keys(data.facets.tags || {}).length > 0 || Object.keys(data.facets.vaults || {}).length > 0)) {
+      const facetsDiv = el("div", { class: "search-facets" });
+
+      // Vault facets
+      const vaultFacets = data.facets.vaults || {};
+      if (Object.keys(vaultFacets).length > 1) {
+        const group = el("div", { class: "search-facets__group" });
+        const label = el("span", { class: "search-facets__label" });
+        label.textContent = "Vaults";
+        group.appendChild(label);
+        for (const [vaultName, count] of Object.entries(vaultFacets)) {
+          const item = el("span", { class: "search-facets__item" });
+          item.innerHTML = `${vaultName} <span class="facet-count">${count}</span>`;
+          item.addEventListener("click", () => {
+            const input = document.getElementById("search-input");
+            // Add vault: operator
+            const current = input.value.replace(/vault:\S+\s*/gi, "").trim();
+            input.value = current + " vault:" + vaultName;
+            _triggerAdvancedSearch(input.value);
+          });
+          group.appendChild(item);
+        }
+        facetsDiv.appendChild(group);
+      }
+
+      // Tag facets
+      const tagFacets = data.facets.tags || {};
+      if (Object.keys(tagFacets).length > 0) {
+        const group = el("div", { class: "search-facets__group" });
+        const label = el("span", { class: "search-facets__label" });
+        label.textContent = "Tags";
+        group.appendChild(label);
+        const entries = Object.entries(tagFacets).slice(0, 12);
+        for (const [tagName, count] of entries) {
+          const item = el("span", { class: "search-facets__item" });
+          item.innerHTML = `#${tagName} <span class="facet-count">${count}</span>`;
+          item.addEventListener("click", () => {
+            addTagFilter(tagName);
+          });
+          group.appendChild(item);
+        }
+        facetsDiv.appendChild(group);
+      }
+
+      area.appendChild(facetsDiv);
+    }
+
+    // Empty state
    if (data.results.length === 0) {
      area.appendChild(el("p", { style: "color:var(--text-muted);margin-top:20px" }, [
        document.createTextNode("Aucun résultat trouvé."),
@ -1613,52 +2307,89 @@
      return;
    }

+    // Results list
    const container = el("div", { class: "search-results" });
    data.results.forEach((r) => {
-      // Create title with highlighting
      const titleDiv = el("div", { class: "search-result-title" });
-      if (query && query.trim()) {
-        highlightSearchText(titleDiv, r.title, query, searchCaseSensitive);
+      if (freeText) {
+        highlightSearchText(titleDiv, r.title, freeText, searchCaseSensitive);
      } else {
        titleDiv.textContent = r.title;
      }
-      
-      // Create snippet with highlighting
-      const snippetDiv = el("div", { class: "search-result-snippet" });
-      if (query && query.trim() && r.snippet) {
-        highlightSearchText(snippetDiv, r.snippet, query, searchCaseSensitive);
+
+      // Snippet — use HTML from backend (already has <mark> tags)
+      const snippetDiv = el("div", { class: "search-result-snippet search-result__snippet" });
+      if (r.snippet && r.snippet.includes("<mark>")) {
+        snippetDiv.innerHTML = r.snippet;
+      } else if (freeText && r.snippet) {
+        highlightSearchText(snippetDiv, r.snippet, freeText, searchCaseSensitive);
      } else {
        snippetDiv.textContent = r.snippet || "";
      }
-      
-      const item = el("div", { class: "search-result-item" }, [
-        titleDiv,
-        el("div", { class: "search-result-vault" }, [document.createTextNode(r.vault + " / " + r.path)]),
-        snippetDiv,
+
+      // Score badge
+      const scoreEl = el("span", { class: "search-result-score", style: "font-size:0.7rem;color:var(--text-muted);margin-left:8px" });
+      scoreEl.textContent = `score: ${r.score}`;
+
+      const vaultPath = el("div", { class: "search-result-vault" }, [
+        document.createTextNode(r.vault + " / " + r.path),
+        scoreEl,
      ]);

+      const item = el("div", { class: "search-result-item" }, [titleDiv, vaultPath, snippetDiv]);
+
      if (r.tags && r.tags.length > 0) {
        const tagsDiv = el("div", { class: "search-result-tags" });
        r.tags.forEach((tag) => {
          if (!TagFilterService.isTagFiltered(tag)) {
            const tagEl = el("span", { class: "file-tag" }, [document.createTextNode(`#${tag}`)]);
-            tagEl.addEventListener("click", (e) => {
-              e.stopPropagation();
-              addTagFilter(tag);
-            });
+            tagEl.addEventListener("click", (e) => { e.stopPropagation(); addTagFilter(tag); });
            tagsDiv.appendChild(tagEl);
          }
        });
-        if (tagsDiv.children.length > 0) {
-          item.appendChild(tagsDiv);
-        }
+        if (tagsDiv.children.length > 0) item.appendChild(tagsDiv);
      }

      item.addEventListener("click", () => openFile(r.vault, r.path));
      container.appendChild(item);
    });
-
    area.appendChild(container);
+
+    // Pagination
+    if (data.total > ADVANCED_SEARCH_LIMIT) {
+      const paginationDiv = el("div", { class: "search-pagination" });
+      const prevBtn = el("button", { class: "search-pagination__btn", type: "button" });
+      prevBtn.textContent = "← Précédent";
+      prevBtn.disabled = advancedSearchOffset === 0;
+      prevBtn.addEventListener("click", () => {
+        advancedSearchOffset = Math.max(0, advancedSearchOffset - ADVANCED_SEARCH_LIMIT);
+        const vault = document.getElementById("vault-filter").value;
+        performAdvancedSearch(query, vault, tagFilter, advancedSearchOffset);
+        document.getElementById("content-area").scrollTop = 0;
+      });
+
+      const info = el("span", { class: "search-pagination__info" });
+      const from = advancedSearchOffset + 1;
+      const to = Math.min(advancedSearchOffset + ADVANCED_SEARCH_LIMIT, data.total);
+      info.textContent = `${from}–${to} sur ${data.total}`;
+
+      const nextBtn = el("button", { class: "search-pagination__btn", type: "button" });
+      nextBtn.textContent = "Suivant →";
+      nextBtn.disabled = advancedSearchOffset + ADVANCED_SEARCH_LIMIT >= data.total;
+      nextBtn.addEventListener("click", () => {
+        advancedSearchOffset += ADVANCED_SEARCH_LIMIT;
+        const vault = document.getElementById("vault-filter").value;
+        performAdvancedSearch(query, vault, tagFilter, advancedSearchOffset);
+        document.getElementById("content-area").scrollTop = 0;
+      });
+
+      paginationDiv.appendChild(prevBtn);
+      paginationDiv.appendChild(info);
+      paginationDiv.appendChild(nextBtn);
+      area.appendChild(paginationDiv);
+    }
+
+    safeCreateIcons();
  }

  // ---------------------------------------------------------------------------
--- a/frontend/index.html
+++ b/frontend/index.html
@ -103,6 +103,31 @@
              </button>
            </div>
          </div>
+          <!-- Advanced search autocomplete dropdown -->
+          <div class="search-dropdown" id="search-dropdown" role="listbox" aria-label="Suggestions de recherche" hidden>
+            <div class="search-dropdown__section search-dropdown__section--history" id="search-dropdown-history">
+              <div class="search-dropdown__section-header">
+                <span>Historique</span>
+                <button class="search-dropdown__clear-btn" id="search-dropdown-clear-history" type="button" title="Effacer l'historique">
+                  <i data-lucide="trash-2" style="width:12px;height:12px"></i>
+                </button>
+              </div>
+              <ul class="search-dropdown__list" id="search-dropdown-history-list"></ul>
+            </div>
+            <div class="search-dropdown__section search-dropdown__section--titles" id="search-dropdown-titles">
+              <div class="search-dropdown__section-header"><span>Fichiers</span></div>
+              <ul class="search-dropdown__list" id="search-dropdown-titles-list"></ul>
+            </div>
+            <div class="search-dropdown__section search-dropdown__section--tags" id="search-dropdown-tags">
+              <div class="search-dropdown__section-header"><span>Tags</span></div>
+              <ul class="search-dropdown__list" id="search-dropdown-tags-list"></ul>
+            </div>
+            <div class="search-dropdown__empty" id="search-dropdown-empty" hidden>
+              Aucune suggestion
+            </div>
+          </div>
+          <!-- Active search filter chips -->
+          <div class="search-chips" id="search-chips" hidden></div>
        </div>
      </div>

--- a/frontend/style.css
+++ b/frontend/style.css
@ -2162,3 +2162,331 @@ body.resizing-v {
  justify-content: center;
  padding: 8px 16px;
 }
+
+/* ---------------------------------------------------------------------------
+   Advanced Search — Autocomplete Dropdown
+   --------------------------------------------------------------------------- */
+.search-dropdown {
+  position: absolute;
+  top: 100%;
+  left: 0;
+  right: 0;
+  margin-top: 4px;
+  background: var(--bg-secondary);
+  border: 1px solid var(--border);
+  border-radius: 8px;
+  box-shadow: 0 8px 24px rgba(0,0,0,0.25);
+  z-index: 200;
+  max-height: 380px;
+  overflow-y: auto;
+  overflow-x: hidden;
+}
+.search-dropdown[hidden] {
+  display: none;
+}
+.search-dropdown__section {
+  border-bottom: 1px solid var(--border);
+}
+.search-dropdown__section:last-child {
+  border-bottom: none;
+}
+.search-dropdown__section[hidden] {
+  display: none;
+}
+.search-dropdown__section-header {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  padding: 6px 12px;
+  font-size: 0.7rem;
+  font-weight: 600;
+  text-transform: uppercase;
+  letter-spacing: 0.05em;
+  color: var(--text-muted);
+  background: var(--bg-primary);
+  position: sticky;
+  top: 0;
+  z-index: 1;
+}
+.search-dropdown__clear-btn {
+  background: none;
+  border: none;
+  color: var(--text-muted);
+  cursor: pointer;
+  padding: 2px 4px;
+  border-radius: 4px;
+  display: flex;
+  align-items: center;
+}
+.search-dropdown__clear-btn:hover {
+  color: var(--danger);
+  background: var(--danger-bg, rgba(255,0,0,0.08));
+}
+.search-dropdown__list {
+  list-style: none;
+  margin: 0;
+  padding: 0;
+}
+.search-dropdown__item {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  padding: 7px 12px;
+  cursor: pointer;
+  font-size: 0.82rem;
+  color: var(--text-primary);
+  transition: background 120ms ease;
+}
+.search-dropdown__item:hover,
+.search-dropdown__item.active {
+  background: var(--bg-hover);
+}
+.search-dropdown__item--history {
+  color: var(--text-secondary);
+}
+.search-dropdown__item--history .search-dropdown__icon {
+  color: var(--text-muted);
+}
+.search-dropdown__item--title .search-dropdown__meta {
+  font-size: 0.72rem;
+  color: var(--text-muted);
+  margin-left: auto;
+  white-space: nowrap;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  max-width: 180px;
+}
+.search-dropdown__item--tag .search-dropdown__badge {
+  margin-left: auto;
+  font-size: 0.7rem;
+  color: var(--text-muted);
+  background: var(--bg-primary);
+  padding: 1px 6px;
+  border-radius: 10px;
+}
+.search-dropdown__icon {
+  flex-shrink: 0;
+  width: 14px;
+  height: 14px;
+  color: var(--text-muted);
+}
+.search-dropdown__text {
+  flex: 1;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  white-space: nowrap;
+}
+.search-dropdown__text mark {
+  background: var(--accent);
+  color: var(--bg-primary);
+  border-radius: 2px;
+  padding: 0 1px;
+}
+.search-dropdown__empty {
+  padding: 16px 12px;
+  text-align: center;
+  font-size: 0.82rem;
+  color: var(--text-muted);
+}
+.search-dropdown__empty[hidden] {
+  display: none;
+}
+
+/* ---------------------------------------------------------------------------
+   Advanced Search — Filter Chips
+   --------------------------------------------------------------------------- */
+.search-chips {
+  display: flex;
+  flex-wrap: wrap;
+  gap: 6px;
+  padding: 6px 0 0;
+}
+.search-chips[hidden] {
+  display: none;
+}
+.search-chip {
+  display: inline-flex;
+  align-items: center;
+  gap: 4px;
+  padding: 3px 8px;
+  font-size: 0.75rem;
+  font-family: 'JetBrains Mono', monospace;
+  border-radius: 12px;
+  background: var(--accent);
+  color: var(--bg-primary);
+  cursor: default;
+  max-width: 200px;
+  white-space: nowrap;
+}
+.search-chip__label {
+  overflow: hidden;
+  text-overflow: ellipsis;
+}
+.search-chip__remove {
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  background: none;
+  border: none;
+  color: inherit;
+  cursor: pointer;
+  padding: 0;
+  opacity: 0.7;
+  transition: opacity 120ms;
+}
+.search-chip__remove:hover {
+  opacity: 1;
+}
+.search-chip--tag {
+  background: var(--accent);
+  color: var(--bg-primary);
+}
+.search-chip--vault {
+  background: var(--success, #22c55e);
+  color: #fff;
+}
+.search-chip--title {
+  background: var(--warning, #f59e0b);
+  color: #1a1a1a;
+}
+.search-chip--path {
+  background: var(--text-muted);
+  color: var(--bg-primary);
+}
+
+/* ---------------------------------------------------------------------------
+   Advanced Search — Snippet Highlights
+   --------------------------------------------------------------------------- */
+.search-result__snippet mark {
+  background: var(--accent);
+  color: var(--bg-primary);
+  border-radius: 2px;
+  padding: 0 2px;
+  font-weight: 600;
+}
+
+/* ---------------------------------------------------------------------------
+   Advanced Search — Facets Panel
+   --------------------------------------------------------------------------- */
+.search-facets {
+  display: flex;
+  flex-wrap: wrap;
+  gap: 12px;
+  margin-bottom: 12px;
+  padding: 10px 12px;
+  background: var(--bg-primary);
+  border: 1px solid var(--border);
+  border-radius: 8px;
+}
+.search-facets[hidden] {
+  display: none;
+}
+.search-facets__group {
+  display: flex;
+  flex-wrap: wrap;
+  gap: 6px;
+  align-items: center;
+}
+.search-facets__label {
+  font-size: 0.72rem;
+  font-weight: 600;
+  text-transform: uppercase;
+  letter-spacing: 0.05em;
+  color: var(--text-muted);
+  margin-right: 4px;
+}
+.search-facets__item {
+  display: inline-flex;
+  align-items: center;
+  gap: 4px;
+  padding: 2px 8px;
+  font-size: 0.75rem;
+  border: 1px solid var(--border);
+  border-radius: 12px;
+  color: var(--text-secondary);
+  background: var(--bg-secondary);
+  cursor: pointer;
+  transition: border-color 120ms, background 120ms;
+}
+.search-facets__item:hover {
+  border-color: var(--accent);
+  background: var(--bg-hover);
+}
+.search-facets__item .facet-count {
+  font-size: 0.68rem;
+  color: var(--text-muted);
+}
+
+/* ---------------------------------------------------------------------------
+   Advanced Search — Pagination
+   --------------------------------------------------------------------------- */
+.search-pagination {
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  gap: 8px;
+  padding: 12px 0;
+}
+.search-pagination__btn {
+  padding: 6px 14px;
+  font-size: 0.82rem;
+  border: 1px solid var(--border);
+  border-radius: 6px;
+  background: var(--bg-secondary);
+  color: var(--text-primary);
+  cursor: pointer;
+  transition: border-color 120ms, background 120ms;
+}
+.search-pagination__btn:hover:not(:disabled) {
+  border-color: var(--accent);
+  background: var(--bg-hover);
+}
+.search-pagination__btn:disabled {
+  opacity: 0.4;
+  cursor: not-allowed;
+}
+.search-pagination__info {
+  font-size: 0.78rem;
+  color: var(--text-muted);
+}
+
+/* ---------------------------------------------------------------------------
+   Advanced Search — Sort Controls
+   --------------------------------------------------------------------------- */
+.search-sort {
+  display: flex;
+  align-items: center;
+  gap: 6px;
+  margin-left: auto;
+}
+.search-sort__btn {
+  padding: 3px 10px;
+  font-size: 0.75rem;
+  border: 1px solid var(--border);
+  border-radius: 6px;
+  background: var(--bg-secondary);
+  color: var(--text-secondary);
+  cursor: pointer;
+  transition: all 120ms;
+}
+.search-sort__btn.active {
+  border-color: var(--accent);
+  color: var(--accent);
+  background: var(--bg-hover);
+}
+
+/* ---------------------------------------------------------------------------
+   Advanced Search — Responsive
+   --------------------------------------------------------------------------- */
+@media (max-width: 768px) {
+  .search-dropdown {
+    max-height: 280px;
+  }
+  .search-facets {
+    flex-direction: column;
+    gap: 8px;
+  }
+  .search-chip {
+    max-width: 150px;
+  }
+}