feat(ai): catalogue d'outils lecture & recherche (#79 phase C)

backend/tools/schemas.py

@@ -44,6 +44,100 @@ class ListTagsInput(BaseModel):
     vault: str | None = Field(None, description="Vault name or 'all' (default: all accessible)")
 
 
+class ListAllFilesInput(BaseModel):
+    """List every file of a vault (optionally under a subdirectory)."""
+
+    vault: str = Field(..., description="Vault name")
+    dir: str = Field("", description="Vault-relative directory path (empty = vault root)")
+    limit: int = Field(200, ge=1, le=2000, description="Maximum number of files")
+    recursive: bool = Field(True, description="Recurse into subdirectories")
+
+
+class ReadFileRawInput(BaseModel):
+    """Read a text file's raw content (no HTML rendering)."""
+
+    vault: str = Field(..., description="Vault name")
+    path: str = Field(..., description="Vault-relative file path")
+
+
+class GetBacklinksInput(BaseModel):
+    """List files linking to a target file via wikilinks."""
+
+    vault: str = Field(..., description="Vault name")
+    path: str = Field(..., description="Vault-relative path of the target file")
+
+
+class ListBackupsInput(BaseModel):
+    """List available backup versions of a file."""
+
+    vault: str = Field(..., description="Vault name")
+    path: str = Field(..., description="Vault-relative file path")
+
+
+class DiffBackupInput(BaseModel):
+    """Unified diff between a backup version and another version or the current file."""
+
+    vault: str = Field(..., description="Vault name")
+    path: str = Field(..., description="Vault-relative file path")
+    version: int = Field(..., description="Backup timestamp used as the old/left side")
+    compare_with: int | None = Field(
+        None,
+        description="Backup timestamp used as the new/right side (omit to compare with the current file)",
+    )
+
+
+class GetGraphInput(BaseModel):
+    """Graph data (nodes and edges) for a vault or directory."""
+
+    vault: str = Field(..., description="Vault name")
+    path: str = Field("", description="Vault-relative directory path to focus on (empty = root)")
+    depth: int = Field(1, ge=0, le=3, description="Expansion depth")
+    scope: str = Field("directory", description="'directory' for a subtree, 'full' for the whole vault")
+    tag: str = Field("", description="Only include files carrying this tag")
+
+
+class SearchAdvancedInput(BaseModel):
+    """Advanced full-text search (TF-IDF, facets, operators)."""
+
+    q: str = Field("", description="Query (supports tag:, vault:, title:, path:, ext: operators)")
+    vault: str = Field("all", description="Vault name or 'all'")
+    tag: str | None = Field(None, description="Optional comma-separated tag filter")
+    limit: int = Field(50, ge=1, le=200, description="Maximum number of results")
+    offset: int = Field(0, ge=0, description="Pagination offset")
+    sort: str = Field("relevance", description="Sort by 'relevance' or 'modified'")
+    case_sensitive: bool = Field(False, description="Match case")
+    whole_word: bool = Field(False, description="Match whole words only")
+    regex: bool = Field(False, description="Treat query as a regular expression")
+    include_paths: str | None = Field(None, description="Comma-separated glob patterns to include")
+    exclude_paths: str | None = Field(None, description="Comma-separated glob patterns to exclude")
+    created: str | None = Field(None, description="Created date filter (>date, <date, date..date)")
+    modified: str | None = Field(None, description="Modified date filter (>date, <date, date..date, <Nd)")
+    size: str | None = Field(None, description="Size filter (>size, <size, size..size, e.g. >1MB)")
+
+
+class SearchPathsInput(BaseModel):
+    """Search files and directories by path substring."""
+
+    q: str = Field(..., min_length=1, description="Path substring to search for")
+    vault: str = Field("all", description="Vault name or 'all'")
+
+
+class SuggestTagsInput(BaseModel):
+    """Suggest tags matching a prefix."""
+
+    q: str = Field(..., min_length=1, description="Tag prefix (with or without leading '#')")
+    vault: str = Field("all", description="Vault name or 'all'")
+    limit: int = Field(10, ge=1, le=50, description="Maximum number of suggestions")
+
+
+class ListRecentInput(BaseModel):
+    """List recently opened (or modified) files."""
+
+    vault: str | None = Field(None, description="Optional single-vault filter")
+    limit: int = Field(20, ge=1, le=200, description="Maximum number of files")
+    mode: str = Field("opened", description="'opened' (history) or 'modified'")
+
+
 class ToolResult(BaseModel):
     """Uniform result returned by :func:`backend.tools.registry.call_tool`."""
 

backend/tools/service.py

@@ -1,4 +1,4 @@
-"""Built-in tool services (Phase 0 — read/search).
+"""Built-in tool services (Phase 0 + Phase C — read/search catalogue).
 
 These functions are the single source of truth consumed by both the in-app
 assistant and the MCP server. They delegate to the shared business-logic
@@ -11,18 +11,35 @@ from __future__ import annotations
 import logging
 from typing import Any
 
+from backend.indexer import get_backlinks as _get_backlinks
+from backend.indexer import get_vault_names
+from backend.services.backups import diff_backup as _diff_backup
+from backend.services.backups import list_backup_files as _list_backup_files
 from backend.services.files import read_file_text
+from backend.services.graph import get_graph as _get_graph
+from backend.services.recent import list_recent as _list_recent
+from backend.services.search import advanced_search_vaults, search_vaults
 from backend.services.search import list_tags as _list_tags
-from backend.services.search import search_vaults
-from backend.services.vaults import browse_directory, list_accessible_vaults
+from backend.services.search import search_paths as _search_paths
+from backend.services.vaults import browse_directory, list_accessible_vaults, list_all_files
 from backend.tools.context import ToolContext, ToolRisk
 from backend.tools.registry import tool
 from backend.tools.schemas import (
+    DiffBackupInput,
+    GetBacklinksInput,
+    GetGraphInput,
+    ListAllFilesInput,
+    ListBackupsInput,
     ListDirectoryInput,
+    ListRecentInput,
     ListTagsInput,
     ListVaultsInput,
     ReadFileInput,
+    ReadFileRawInput,
+    SearchAdvancedInput,
     SearchFulltextInput,
+    SearchPathsInput,
+    SuggestTagsInput,
 )
 
 logger = logging.getLogger("obsigate.tools.service")
@@ -31,6 +48,9 @@ logger = logging.getLogger("obsigate.tools.service")
 TOOL_MAX_READ_BYTES = 200_000
 
 
+# ── C1. Vaults / navigation ────────────────────────────────────────────────
+
+
 @tool(
     name="list_vaults",
     description="List the vaults the current user is allowed to access.",
@@ -61,6 +81,21 @@ def list_directory(ctx: ToolContext, params: ListDirectoryInput) -> list[dict[st
     ]
 
 
+@tool(
+    name="list_all_files",
+    description="List every file of a vault (optionally under a subdirectory), newest first.",
+    input_model=ListAllFilesInput,
+    risk=ToolRisk.READ,
+    requires_vault=True,
+)
+def list_all_files_tool(ctx: ToolContext, params: ListAllFilesInput) -> dict[str, Any]:
+    """Return a flat list of files with metadata."""
+    return list_all_files(params.vault, dir=params.dir, limit=params.limit, recursive=params.recursive)
+
+
+# ── C2. Content reading ────────────────────────────────────────────────────
+
+
 @tool(
     name="read_file",
     description="Read the text content of a file inside a vault (secrets redacted).",
@@ -78,6 +113,81 @@ def read_file(ctx: ToolContext, params: ReadFileInput) -> dict[str, Any]:
     )
 
 
+@tool(
+    name="read_file_raw",
+    description="Read a file's raw text content, without the size cap (secrets redacted).",
+    input_model=ReadFileRawInput,
+    risk=ToolRisk.READ,
+    requires_vault=True,
+)
+def read_file_raw(ctx: ToolContext, params: ReadFileRawInput) -> dict[str, Any]:
+    """Return the full (redacted) raw text of a vault file."""
+    data = read_file_text(params.vault, params.path, redact=True, max_bytes=None)
+    return {"vault": data["vault"], "path": data["path"], "raw": data["content"]}
+
+
+@tool(
+    name="get_backlinks",
+    description="List files that link to a target file via [[wikilinks]].",
+    input_model=GetBacklinksInput,
+    risk=ToolRisk.READ,
+    requires_vault=True,
+)
+def get_backlinks(ctx: ToolContext, params: GetBacklinksInput) -> list[dict[str, Any]]:
+    """Return backlinks, filtered to accessible vaults."""
+    backlinks = _get_backlinks(params.vault, params.path)
+    return [b for b in backlinks if ctx.has_vault_access(b["vault"])]
+
+
+@tool(
+    name="list_backups",
+    description="List the available backup versions of a file (newest first).",
+    input_model=ListBackupsInput,
+    risk=ToolRisk.READ,
+    requires_vault=True,
+)
+def list_backups(ctx: ToolContext, params: ListBackupsInput) -> dict[str, Any]:
+    """Return the backup versions of a vault file."""
+    return {
+        "vault": params.vault,
+        "path": params.path,
+        "backups": _list_backup_files(params.vault, params.path),
+    }
+
+
+@tool(
+    name="diff_backup",
+    description="Show a unified diff between a backup version and another version or the current file.",
+    input_model=DiffBackupInput,
+    risk=ToolRisk.READ,
+    requires_vault=True,
+)
+def diff_backup(ctx: ToolContext, params: DiffBackupInput) -> dict[str, Any]:
+    """Return the unified diff for a backup version."""
+    return _diff_backup(params.vault, params.path, params.version, params.compare_with)
+
+
+@tool(
+    name="get_graph",
+    description="Return the graph (nodes and edges: parent/child + wikilinks) of a vault or directory.",
+    input_model=GetGraphInput,
+    risk=ToolRisk.READ,
+    requires_vault=True,
+)
+def get_graph(ctx: ToolContext, params: GetGraphInput) -> dict[str, Any]:
+    """Return graph data for a vault subtree or the whole vault."""
+    return _get_graph(
+        params.vault,
+        path=params.path,
+        depth=params.depth,
+        scope=params.scope,
+        tag=params.tag,
+    )
+
+
+# ── C3. Search ─────────────────────────────────────────────────────────────
+
+
 @tool(
     name="search_fulltext",
     description="Full-text search across one vault or all accessible vaults.",
@@ -90,6 +200,45 @@ def search_fulltext(ctx: ToolContext, params: SearchFulltextInput) -> list[dict[
     return [r for r in payload["results"] if ctx.has_vault_access(r["vault"])]
 
 
+@tool(
+    name="search_advanced",
+    description="Advanced full-text search with operators, filters and facets.",
+    input_model=SearchAdvancedInput,
+    risk=ToolRisk.READ,
+)
+def search_advanced(ctx: ToolContext, params: SearchAdvancedInput) -> list[dict[str, Any]]:
+    """Return advanced search results, filtered to accessible vaults."""
+    payload = advanced_search_vaults(
+        params.q,
+        vault=params.vault,
+        tag=params.tag,
+        limit=params.limit,
+        offset=params.offset,
+        sort=params.sort,
+        case_sensitive=params.case_sensitive,
+        whole_word=params.whole_word,
+        regex=params.regex,
+        include_paths=params.include_paths,
+        exclude_paths=params.exclude_paths,
+        created=params.created,
+        modified=params.modified,
+        size=params.size,
+    )
+    return [r for r in payload["results"] if ctx.has_vault_access(r["vault"])]
+
+
+@tool(
+    name="search_paths",
+    description="Search files and directories by path substring.",
+    input_model=SearchPathsInput,
+    risk=ToolRisk.READ,
+)
+def search_paths(ctx: ToolContext, params: SearchPathsInput) -> list[dict[str, Any]]:
+    """Return path matches, filtered to accessible vaults."""
+    payload = _search_paths(params.q, vault=params.vault)
+    return [r for r in payload["results"] if ctx.has_vault_access(r["vault"])]
+
+
 @tool(
     name="list_tags",
     description="List tags with their occurrence counts for a vault or all accessible vaults.",
@@ -102,8 +251,6 @@ def list_tags(ctx: ToolContext, params: ListTagsInput) -> list[dict[str, Any]]:
         ctx.require_vault_access(params.vault)
         return [{"tag": tag, "count": count} for tag, count in _list_tags(params.vault).items()]
 
-    from backend.indexer import get_vault_names
-
     merged: dict[str, int] = {}
     for name in get_vault_names():
         if not ctx.has_vault_access(name):
@@ -111,3 +258,45 @@ def list_tags(ctx: ToolContext, params: ListTagsInput) -> list[dict[str, Any]]:
         for tag, count in _list_tags(name).items():
             merged[tag] = merged.get(tag, 0) + count
     return [{"tag": tag, "count": count} for tag, count in sorted(merged.items(), key=lambda x: -x[1])]
+
+
+@tool(
+    name="suggest_tags",
+    description="Suggest tags matching a prefix, across accessible vaults.",
+    input_model=SuggestTagsInput,
+    risk=ToolRisk.READ,
+)
+def suggest_tags(ctx: ToolContext, params: SuggestTagsInput) -> list[dict[str, Any]]:
+    """Return tag suggestions, restricted to accessible vaults."""
+    from backend.search import suggest_tags as _suggest
+
+    if params.vault and params.vault != "all":
+        ctx.require_vault_access(params.vault)
+        return _suggest(params.q, vault_filter=params.vault, limit=params.limit)
+
+    merged: dict[str, int] = {}
+    for name in get_vault_names():
+        if not ctx.has_vault_access(name):
+            continue
+        for item in _suggest(params.q, vault_filter=name, limit=params.limit):
+            merged[item["tag"]] = merged.get(item["tag"], 0) + item["count"]
+    ordered = sorted(merged.items(), key=lambda x: -x[1])
+    return [{"tag": tag, "count": count} for tag, count in ordered[: params.limit]]
+
+
+@tool(
+    name="list_recent",
+    description="List the current user's recently opened (or modified) files.",
+    input_model=ListRecentInput,
+    risk=ToolRisk.READ,
+)
+def list_recent(ctx: ToolContext, params: ListRecentInput) -> dict[str, Any]:
+    """Return recent files for the calling user."""
+    user_vaults = ctx.user.get("_token_vaults") or ctx.user.get("vaults", [])
+    return _list_recent(
+        ctx.username,
+        user_vaults,
+        vault=params.vault,
+        limit=params.limit,
+        mode=params.mode,
+    )