gateway/modules/routes/routeDataFiles.py

1373 lines
56 KiB
Python

# Copyright (c) 2025 Patrick Motsch
# All rights reserved.
from fastapi import APIRouter, HTTPException, Depends, File, UploadFile, Form, Path, Request, status, Query, Response, Body, BackgroundTasks
from fastapi.responses import JSONResponse
from typing import List, Dict, Any, Optional
import logging
import json
import math
# Import auth module
from modules.auth import limiter, getCurrentUser, getRequestContext, RequestContext
# Import interfaces
import modules.interfaces.interfaceDbManagement as interfaceDbManagement
from modules.datamodels.datamodelFiles import FileItem, FilePreview, FileFolder
from modules.shared.attributeUtils import getModelAttributeDefinitions
from modules.datamodels.datamodelUam import User
from modules.datamodels.datamodelPagination import PaginationParams, PaginatedResponse, PaginationMetadata, normalize_pagination_dict
from modules.shared.i18nRegistry import apiRouteContext
from modules.routes.routeHelpers import enrichRowsWithFkLabels
routeApiMsg = apiRouteContext("routeDataFiles")
# Configure logger
logger = logging.getLogger(__name__)
def _resolveFileWithScope(currentUser: User, context: RequestContext, fileId: str):
"""Returns (managementInterface, fileItem) with RBAC scoped to the file's own mandate/instance.
Files generated by workflows (e.g. AI report outputs) carry their own
mandateId/featureInstanceId. Direct download links via <a href> cannot send
custom scope headers, so we resolve the scope from the FileItem itself and
re-check RBAC in that scope.
Returns (None, None) if the file does not exist or the user lacks access
in the file's actual scope.
"""
requestMandateId = str(context.mandateId) if context.mandateId else None
requestInstanceId = str(context.featureInstanceId) if context.featureInstanceId else None
mgmt = interfaceDbManagement.getInterface(
currentUser,
mandateId=requestMandateId,
featureInstanceId=requestInstanceId,
)
fileItem = mgmt.getFile(fileId)
if fileItem:
return mgmt, fileItem
metas = mgmt.db.getRecordset(FileItem, recordFilter={"id": fileId})
if not metas:
return None, None
meta = metas[0]
fileMandateId = meta.get("mandateId") or None
fileInstanceId = meta.get("featureInstanceId") or None
if not fileMandateId and not fileInstanceId:
return None, None
if fileMandateId == requestMandateId and fileInstanceId == requestInstanceId:
return None, None
scopedMgmt = interfaceDbManagement.getInterface(
currentUser,
mandateId=fileMandateId,
featureInstanceId=fileInstanceId,
)
fileItem = scopedMgmt.getFile(fileId)
if not fileItem:
return None, None
return scopedMgmt, fileItem
async def _autoIndexFile(fileId: str, fileName: str, mimeType: str, user, *, mandateId: str = None, featureInstanceId: str = None):
"""Background task: pre-scan + extraction + knowledge indexing.
Step 1: Structure Pre-Scan (AI-free) -> FileContentIndex (persisted)
Step 2: Content extraction via runExtraction -> ContentParts
Step 3: KnowledgeService.requestIngestion -> idempotent chunking + embedding -> Knowledge Store"""
userId = user.id if hasattr(user, "id") else str(user)
try:
mgmtInterface = interfaceDbManagement.getInterface(
user,
mandateId=mandateId or None,
featureInstanceId=featureInstanceId or None,
)
mgmtInterface.updateFile(fileId, {"status": "processing"})
rawBytes = mgmtInterface.getFileData(fileId)
if not rawBytes:
logger.warning(f"Auto-index: no file data for {fileId}, skipping")
mgmtInterface.updateFile(fileId, {"status": "active"})
return
file_meta = mgmtInterface.getFile(fileId)
feature_instance_id = ""
mandate_id = ""
file_scope = "personal"
if file_meta:
if isinstance(file_meta, dict):
feature_instance_id = file_meta.get("featureInstanceId") or ""
mandate_id = file_meta.get("mandateId") or ""
file_scope = file_meta.get("scope") or "personal"
else:
feature_instance_id = getattr(file_meta, "featureInstanceId", None) or ""
mandate_id = getattr(file_meta, "mandateId", None) or ""
file_scope = getattr(file_meta, "scope", None) or "personal"
logger.info(f"Auto-index starting for {fileName} ({len(rawBytes)} bytes, {mimeType})")
# Step 1: Structure Pre-Scan (AI-free)
from modules.serviceCenter.services.serviceKnowledge.subPreScan import preScanDocument
contentIndex = await preScanDocument(
fileData=rawBytes,
mimeType=mimeType,
fileId=fileId,
fileName=fileName,
userId=userId,
featureInstanceId=str(feature_instance_id) if feature_instance_id else "",
mandateId=str(mandate_id) if mandate_id else "",
scope=file_scope,
)
logger.info(
f"Pre-scan complete for {fileName}: "
f"{contentIndex.totalObjects} objects"
)
# Persist FileContentIndex immediately.
# IMPORTANT: preserve `_ingestion` metadata and `status="indexed"` from any
# prior successful run — otherwise this upsert wipes the idempotency cache
# and requestIngestion cannot detect duplicates (AC4 breaks).
from modules.interfaces.interfaceDbKnowledge import getInterface as getKnowledgeInterface
knowledgeDb = getKnowledgeInterface()
try:
_existing = knowledgeDb.getFileContentIndex(fileId)
except Exception:
_existing = None
if _existing:
_existingStruct = (
_existing.get("structure") if isinstance(_existing, dict)
else getattr(_existing, "structure", {})
) or {}
_existingStatus = (
_existing.get("status") if isinstance(_existing, dict)
else getattr(_existing, "status", "")
) or ""
if "_ingestion" in _existingStruct:
contentIndex.structure = dict(contentIndex.structure or {})
contentIndex.structure["_ingestion"] = _existingStruct["_ingestion"]
if _existingStatus == "indexed":
contentIndex.status = "indexed"
knowledgeDb.upsertFileContentIndex(contentIndex)
# Step 2: Content extraction (AI-free, produces ContentParts)
from modules.serviceCenter.services.serviceExtraction.subRegistry import ExtractorRegistry, ChunkerRegistry
from modules.serviceCenter.services.serviceExtraction.subPipeline import runExtraction
from modules.datamodels.datamodelExtraction import ExtractionOptions
extractorRegistry = ExtractorRegistry()
chunkerRegistry = ChunkerRegistry()
# mergeStrategy=None: keep per-page / per-section granularity for RAG ingestion.
# The default MergeStrategy concatenates all text parts into a single blob, which
# collapses a 500-page PDF into one ContentChunk and destroys semantic retrieval.
options = ExtractionOptions(mergeStrategy=None)
extracted = runExtraction(
extractorRegistry, chunkerRegistry,
rawBytes, fileName, mimeType, options,
)
contentObjects = []
for part in extracted.parts:
contentType = "text"
if part.typeGroup == "image":
contentType = "image"
elif part.typeGroup in ("binary", "container"):
contentType = "other"
if not part.data or not part.data.strip():
continue
contentObjects.append({
"contentObjectId": part.id,
"contentType": contentType,
"data": part.data,
"contextRef": {
"containerPath": fileName,
"location": part.label or "file",
**(part.metadata or {}),
},
})
logger.info(f"Extracted {len(contentObjects)} content objects from {fileName}")
if not contentObjects:
knowledgeDb.updateFileStatus(fileId, "indexed")
mgmtInterface.updateFile(fileId, {"status": "active"})
return
# Step 3: Knowledge indexing (chunking + embedding)
from modules.serviceCenter import getService
from modules.serviceCenter.context import ServiceCenterContext
ctx = ServiceCenterContext(
user=user,
mandate_id=str(mandate_id) if mandate_id else "",
feature_instance_id=str(feature_instance_id) if feature_instance_id else "",
)
knowledgeService = getService("knowledge", ctx)
from modules.serviceCenter.services.serviceKnowledge.mainServiceKnowledge import IngestionJob
await knowledgeService.requestIngestion(
IngestionJob(
sourceKind="file",
sourceId=fileId,
fileName=fileName,
mimeType=mimeType,
userId=userId,
featureInstanceId=str(feature_instance_id) if feature_instance_id else "",
mandateId=str(mandate_id) if mandate_id else "",
contentObjects=contentObjects,
structure=contentIndex.structure,
provenance={"lane": "upload", "route": "routeDataFiles._autoIndexFile"},
)
)
# Re-acquire interface after await to avoid stale user context from the singleton
mgmtInterface = interfaceDbManagement.getInterface(user)
mgmtInterface.updateFile(fileId, {"status": "active"})
logger.info(f"Auto-index complete for file {fileId} ({fileName})")
except Exception as e:
logger.error(f"Auto-index failed for file {fileId}: {e}", exc_info=True)
try:
errMgmt = interfaceDbManagement.getInterface(user)
errMgmt.updateFile(fileId, {"status": "active"})
except Exception:
pass
# Model attributes for FileItem
fileAttributes = getModelAttributeDefinitions(FileItem)
# Create router for file endpoints
router = APIRouter(
prefix="/api/files",
tags=["Manage Files"],
responses={
404: {"description": "Not found"},
400: {"description": "Bad request"},
401: {"description": "Unauthorized"},
403: {"description": "Forbidden"},
500: {"description": "Internal server error"}
}
)
@router.get("/folders/tree")
@limiter.limit("120/minute")
def get_folder_tree(
request: Request,
owner: str = Query("me", description="'me' | 'shared'"),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
):
try:
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
o = (owner or "me").strip().lower()
if o == "me":
return managementInterface.getOwnFolderTree()
if o == "shared":
return managementInterface.getSharedFolderTree()
raise HTTPException(status_code=400, detail="owner must be 'me' or 'shared'")
except HTTPException:
raise
except Exception as e:
logger.error(f"get_folder_tree error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.post("/folders", status_code=status.HTTP_201_CREATED)
@limiter.limit("30/minute")
def create_folder(
request: Request,
body: Dict[str, Any] = Body(...),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
):
try:
name = body.get("name")
if not name or not str(name).strip():
raise HTTPException(status_code=400, detail="name is required")
parentId = body.get("parentId") or None
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
return managementInterface.createFolder(str(name).strip(), parentId)
except PermissionError as e:
raise HTTPException(status_code=403, detail=str(e))
except interfaceDbManagement.FileNotFoundError as e:
raise HTTPException(status_code=404, detail=str(e))
except HTTPException:
raise
except Exception as e:
logger.error(f"create_folder error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.patch("/folders/{folderId}")
@limiter.limit("30/minute")
def rename_folder(
request: Request,
folderId: str = Path(...),
body: Dict[str, Any] = Body(...),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
):
try:
name = body.get("name")
if not name or not str(name).strip():
raise HTTPException(status_code=400, detail="name is required")
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
return managementInterface.renameFolder(folderId, str(name).strip())
except PermissionError as e:
raise HTTPException(status_code=403, detail=str(e))
except interfaceDbManagement.FileNotFoundError as e:
raise HTTPException(status_code=404, detail=str(e))
except HTTPException:
raise
except Exception as e:
logger.error(f"rename_folder error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.post("/folders/{folderId}/move")
@limiter.limit("30/minute")
def move_folder(
request: Request,
folderId: str = Path(...),
body: Dict[str, Any] = Body(...),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
):
try:
newParentId = body.get("parentId")
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
return managementInterface.moveFolder(folderId, newParentId or None)
except ValueError as e:
raise HTTPException(status_code=400, detail=str(e))
except PermissionError as e:
raise HTTPException(status_code=403, detail=str(e))
except interfaceDbManagement.FileNotFoundError as e:
raise HTTPException(status_code=404, detail=str(e))
except HTTPException:
raise
except Exception as e:
logger.error(f"move_folder error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.delete("/folders/{folderId}")
@limiter.limit("30/minute")
def delete_folder(
request: Request,
folderId: str = Path(...),
cascade: bool = Query(True, description="Cascade delete sub-folders and files"),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
):
try:
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
return managementInterface.deleteFolderCascade(folderId)
except PermissionError as e:
raise HTTPException(status_code=403, detail=str(e))
except interfaceDbManagement.FileNotFoundError as e:
raise HTTPException(status_code=404, detail=str(e))
except HTTPException:
raise
except Exception as e:
logger.error(f"delete_folder error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.patch("/folders/{folderId}/scope")
@limiter.limit("30/minute")
def patch_folder_scope(
request: Request,
folderId: str = Path(...),
body: Dict[str, Any] = Body(...),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
):
try:
scope = body.get("scope")
if not scope:
raise HTTPException(status_code=400, detail="scope is required")
cascadeToFiles = body.get("cascadeToFiles", False)
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
return managementInterface.patchFolderScope(folderId, scope, cascadeToFiles)
except ValueError as e:
raise HTTPException(status_code=400, detail=str(e))
except PermissionError as e:
raise HTTPException(status_code=403, detail=str(e))
except interfaceDbManagement.FileNotFoundError as e:
raise HTTPException(status_code=404, detail=str(e))
except HTTPException:
raise
except Exception as e:
logger.error(f"patch_folder_scope error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.patch("/folders/{folderId}/neutralize")
@limiter.limit("30/minute")
def patch_folder_neutralize(
request: Request,
folderId: str = Path(...),
body: Dict[str, Any] = Body(...),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
):
try:
neutralize = body.get("neutralize")
if neutralize is None:
raise HTTPException(status_code=400, detail="neutralize is required")
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
return managementInterface.patchFolderNeutralize(folderId, bool(neutralize))
except PermissionError as e:
raise HTTPException(status_code=403, detail=str(e))
except interfaceDbManagement.FileNotFoundError as e:
raise HTTPException(status_code=404, detail=str(e))
except HTTPException:
raise
except Exception as e:
logger.error(f"patch_folder_neutralize error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.get("/list")
@limiter.limit("120/minute")
def get_files(
request: Request,
pagination: Optional[str] = Query(None, description="JSON-encoded PaginationParams object"),
mode: Optional[str] = Query(None, description="'filterValues' for distinct column values, 'ids' for all filtered IDs"),
column: Optional[str] = Query(None, description="Column key (required when mode=filterValues)"),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext)
):
"""
Get files with optional pagination, sorting, and filtering.
Query Parameters:
- pagination: JSON-encoded PaginationParams object, or None for no pagination
Examples:
- GET /api/files/list (no pagination - returns all items)
- GET /api/files/list?pagination={"page":1,"pageSize":10,"sort":[]}
- GET /api/files/list?pagination={"page":2,"pageSize":20,"sort":[{"field":"fileName","direction":"asc"}]}
"""
try:
# Parse pagination parameter
paginationParams = None
if pagination:
try:
paginationDict = json.loads(pagination)
if paginationDict:
paginationDict = normalize_pagination_dict(paginationDict)
paginationParams = PaginationParams(**paginationDict)
except (json.JSONDecodeError, ValueError) as e:
raise HTTPException(
status_code=400,
detail=f"Invalid pagination parameter: {str(e)}"
)
from modules.routes.routeHelpers import (
handleIdsMode,
handleFilterValuesInMemory,
resolveView, applyViewToParams, buildGroupLayout, effective_group_by_levels,
)
import modules.interfaces.interfaceDbApp as _appIface
from modules.datamodels.datamodelPagination import AppliedViewMeta
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None
)
appInterface = _appIface.getInterface(currentUser)
# Resolve view and merge config into params
viewKey = paginationParams.viewKey if paginationParams else None
viewConfig, viewDisplayName = resolveView(appInterface, "files/list", viewKey)
viewMeta = AppliedViewMeta(viewKey=viewKey, displayName=viewDisplayName) if viewKey else None
paginationParams = applyViewToParams(paginationParams, viewConfig)
groupByLevels = effective_group_by_levels(paginationParams, viewConfig)
def _filesToDicts(fileItems):
return [f.model_dump() if hasattr(f, "model_dump") else (dict(f) if not isinstance(f, dict) else f) for f in fileItems]
if mode == "groupSummary":
if not pagination:
raise HTTPException(status_code=400, detail="pagination required for groupSummary")
from modules.routes.routeHelpers import (
apply_strategy_b_filters_and_sort,
build_group_summary_groups,
)
if not groupByLevels or not groupByLevels[0].get("field"):
raise HTTPException(
status_code=400,
detail="groupByLevels[0].field required for groupSummary",
)
field = groupByLevels[0]["field"]
null_label = str(groupByLevels[0].get("nullLabel") or "")
allFiles = managementInterface.getAllFiles()
allItems = enrichRowsWithFkLabels(
_filesToDicts(allFiles if isinstance(allFiles, list) else (allFiles.items if hasattr(allFiles, "items") else [])),
FileItem,
)
filtered = apply_strategy_b_filters_and_sort(allItems, paginationParams, currentUser)
groups_out = build_group_summary_groups(filtered, field, null_label)
return JSONResponse(content={"groups": groups_out})
if mode == "filterValues":
if not column:
raise HTTPException(status_code=400, detail="column parameter required for mode=filterValues")
allFiles = managementInterface.getAllFiles()
items = allFiles if isinstance(allFiles, list) else (allFiles.items if hasattr(allFiles, "items") else [])
itemDicts = _filesToDicts(items)
enrichRowsWithFkLabels(itemDicts, FileItem)
return handleFilterValuesInMemory(itemDicts, column, pagination)
if mode == "ids":
recordFilter = {"sysCreatedBy": managementInterface.userId}
return handleIdsMode(managementInterface.db, FileItem, pagination, recordFilter)
if not groupByLevels:
# No grouping: let DB handle pagination directly (fastest path)
result = managementInterface.getAllFiles(pagination=paginationParams)
if paginationParams and hasattr(result, 'items'):
enriched = enrichRowsWithFkLabels(_filesToDicts(result.items), FileItem)
resp: dict = {
"items": enriched,
"pagination": PaginationMetadata(
currentPage=paginationParams.page,
pageSize=paginationParams.pageSize,
totalItems=result.totalItems,
totalPages=result.totalPages,
sort=paginationParams.sort,
filters=paginationParams.filters
).model_dump(),
}
else:
items = result if isinstance(result, list) else (result.items if hasattr(result, "items") else [result])
resp = {"items": enrichRowsWithFkLabels(_filesToDicts(items), FileItem), "pagination": None}
if viewMeta:
resp["appliedView"] = viewMeta.model_dump()
return resp
# Strategy B grouping: load full list, group, then slice
allFiles = managementInterface.getAllFiles()
allItems = enrichRowsWithFkLabels(
_filesToDicts(allFiles if isinstance(allFiles, list) else (allFiles.items if hasattr(allFiles, "items") else [])),
FileItem,
)
from modules.routes.routeHelpers import apply_strategy_b_filters_and_sort
if paginationParams.filters or paginationParams.sort:
allItems = apply_strategy_b_filters_and_sort(allItems, paginationParams, currentUser)
if not paginationParams:
resp = {"items": allItems, "pagination": None}
if viewMeta:
resp["appliedView"] = viewMeta.model_dump()
return resp
totalItems = len(allItems)
totalPages = math.ceil(totalItems / paginationParams.pageSize) if totalItems > 0 else 0
page_items, groupLayout = buildGroupLayout(allItems, groupByLevels, paginationParams.page, paginationParams.pageSize)
resp = {
"items": page_items,
"pagination": PaginationMetadata(
currentPage=paginationParams.page,
pageSize=paginationParams.pageSize,
totalItems=totalItems,
totalPages=totalPages,
sort=paginationParams.sort,
filters=paginationParams.filters
).model_dump(),
}
if groupLayout:
resp["groupLayout"] = groupLayout.model_dump()
if viewMeta:
resp["appliedView"] = viewMeta.model_dump()
return resp
except HTTPException:
raise
except Exception as e:
logger.error(f"Error getting files: {str(e)}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Failed to get files: {str(e)}"
)
def _LEGACY_addFileToGroup_REMOVED():
"""Removed — file-group tree no longer exists. Use multi-select bulk operations."""
pass
@router.post("/upload", status_code=status.HTTP_201_CREATED)
@limiter.limit("10/minute")
async def upload_file(
request: Request,
file: UploadFile = File(...),
workflowId: Optional[str] = Form(None),
featureInstanceId: Optional[str] = Form(None),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
) -> JSONResponse:
# Add fileName property to UploadFile for consistency with backend model
file.fileName = file.filename
"""Upload a file"""
try:
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
# Read file
fileContent = await file.read()
# Check size limits
maxSize = int(interfaceDbManagement.APP_CONFIG.get("File_Management_MAX_UPLOAD_SIZE_MB")) * 1024 * 1024 # in bytes
if len(fileContent) > maxSize:
raise HTTPException(
status_code=status.HTTP_413_REQUEST_ENTITY_TOO_LARGE,
detail=f"File too large. Maximum size: {interfaceDbManagement.APP_CONFIG.get('File_Management_MAX_UPLOAD_SIZE_MB')}MB"
)
# Save file via LucyDOM interface in the database
fileItem, duplicateType = managementInterface.saveUploadedFile(
fileContent, file.filename
)
if featureInstanceId and not fileItem.featureInstanceId:
managementInterface.updateFile(fileItem.id, {"featureInstanceId": featureInstanceId})
fileItem.featureInstanceId = featureInstanceId
# Determine response message based on duplicate type
if duplicateType == "exact_duplicate":
message = f"File '{file.filename}' already exists with identical content. Reusing existing file."
elif duplicateType == "name_conflict":
message = f"File '{file.filename}' already exists with different content. Uploaded as '{fileItem.fileName}'."
else: # new_file
message = "File uploaded successfully"
# Convert FileItem to dictionary for JSON response
fileMeta = fileItem.model_dump()
# If workflowId is provided, include it in the response (not stored in FileItem model)
if workflowId:
fileMeta["workflowId"] = workflowId
# Trigger background auto-index pipeline (non-blocking)
# Also runs for duplicates in case the original was never successfully indexed
shouldIndex = duplicateType == "new_file"
if not shouldIndex:
try:
from modules.interfaces.interfaceDbKnowledge import getInterface as _getKnowledgeInterface
_kDb = _getKnowledgeInterface()
_existingIndex = _kDb.getFileContentIndex(fileItem.id)
if not _existingIndex:
shouldIndex = True
logger.info(f"Re-triggering auto-index for duplicate {fileItem.id} (not yet indexed)")
except Exception:
shouldIndex = True
if shouldIndex:
try:
import asyncio
asyncio.ensure_future(_autoIndexFile(
fileId=fileItem.id,
fileName=fileItem.fileName,
mimeType=fileItem.mimeType,
user=currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
))
except Exception as indexErr:
logger.warning(f"Auto-index trigger failed (non-blocking): {indexErr}")
# Response with duplicate information
return JSONResponse({
"message": message,
"file": fileMeta,
"duplicateType": duplicateType,
"originalFileName": file.filename,
"storedFileName": fileItem.fileName,
"isDuplicate": duplicateType != "new_file"
})
except interfaceDbManagement.FileStorageError as e:
logger.error(f"Error during file upload (storage): {str(e)}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=str(e)
)
except Exception as e:
logger.error(f"Error during file upload: {str(e)}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Error during file upload: {str(e)}"
)
@router.post("/batch-delete")
@limiter.limit("10/minute")
def batch_delete_items(
request: Request,
body: Dict[str, Any] = Body(...),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext)
) -> Dict[str, Any]:
"""Batch delete files."""
fileIds = body.get("fileIds") or []
if not isinstance(fileIds, list):
raise HTTPException(status_code=400, detail=routeApiMsg("fileIds must be an array"))
try:
mgmt = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
result = {"deletedFiles": 0}
if fileIds:
fileResult = mgmt.deleteFilesBatch(fileIds)
result["deletedFiles"] += fileResult.get("deletedFiles", 0)
return result
except ValueError as e:
raise HTTPException(status_code=400, detail=str(e))
except Exception as e:
logger.error(f"Error in batch delete: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.post("/batch-download")
@limiter.limit("10/minute")
def batchDownload(
request: Request,
body: Dict[str, Any] = Body(...),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
):
"""Download multiple files and/or folders as a single ZIP archive,
preserving the folder hierarchy as ZIP paths."""
import io, zipfile
fileIds = body.get("fileIds") or []
folderIds = body.get("folderIds") or []
if not fileIds and not folderIds:
raise HTTPException(status_code=400, detail="fileIds or folderIds required")
try:
mgmt = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
folderCache: dict[str, dict] = {}
def _getFolder(fid: str):
if fid not in folderCache:
f = mgmt.getFolder(fid)
folderCache[fid] = f if f else {}
return folderCache[fid]
def _folderPath(fid: str) -> str:
"""Build the full path for a folder by walking up parentId."""
parts: list[str] = []
current = fid
visited: set[str] = set()
while current and current not in visited:
visited.add(current)
folder = _getFolder(current)
if not folder:
break
parts.append(folder.get("name", current))
current = folder.get("parentId")
parts.reverse()
return "/".join(parts)
# Collect files from requested folders (recursive)
fileEntries: list[tuple[str, str]] = []
seenFileIds: set[str] = set()
for fid in folderIds:
childFolderIds = mgmt._collectChildFolderIds(fid)
for cfid in childFolderIds:
prefix = _folderPath(cfid)
items = mgmt.db.getRecordset(FileItem, recordFilter={"folderId": cfid})
for item in items:
itemId = item.get("id") if isinstance(item, dict) else getattr(item, "id", None)
if itemId and itemId not in seenFileIds:
seenFileIds.add(itemId)
fileEntries.append((itemId, prefix))
# Loose files (not via folder selection)
for fid in fileIds:
if fid in seenFileIds:
continue
seenFileIds.add(fid)
fileMeta = mgmt.getFile(fid)
if not fileMeta:
continue
fileFolderId = fileMeta.get("folderId") if isinstance(fileMeta, dict) else getattr(fileMeta, "folderId", None)
prefix = _folderPath(fileFolderId) if fileFolderId else ""
fileEntries.append((fid, prefix))
if not fileEntries:
raise HTTPException(status_code=404, detail="No downloadable files found")
buf = io.BytesIO()
with zipfile.ZipFile(buf, "w", zipfile.ZIP_DEFLATED) as zf:
for fid, prefix in fileEntries:
try:
fileMeta = mgmt.getFile(fid)
fileData = mgmt.getFileData(fid)
if fileMeta and fileData:
name = (fileMeta.get("fileName") if isinstance(fileMeta, dict) else getattr(fileMeta, "fileName", fid)) or fid
zipPath = f"{prefix}/{name}" if prefix else name
zf.writestr(zipPath, fileData)
except Exception as fe:
logger.warning(f"batch_download: skipping file {fid}: {fe}")
buf.seek(0)
from fastapi.responses import StreamingResponse
return StreamingResponse(
buf,
media_type="application/zip",
headers={"Content-Disposition": 'attachment; filename="download.zip"'},
)
except HTTPException:
raise
except Exception as e:
logger.error(f"batch_download error: {e}")
raise HTTPException(status_code=500, detail=str(e))
# ── Bulk file operations (replace former group-based bulk routes) ─────────────
@router.post("/bulk/scope")
@limiter.limit("30/minute")
def bulk_set_scope(
request: Request,
body: dict = Body(...),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
):
"""Set scope for a list of files by their IDs."""
fileIds: list = body.get("fileIds") or []
scope: str = body.get("scope") or ""
if not fileIds:
raise HTTPException(status_code=400, detail="fileIds is required")
validScopes = {"personal", "featureInstance", "mandate", "global"}
if scope not in validScopes:
raise HTTPException(status_code=400, detail=f"Invalid scope. Must be one of {validScopes}")
if scope == "global" and not context.isSysAdmin:
raise HTTPException(status_code=403, detail="Only sysadmins can set global scope")
try:
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
updated = 0
for fid in fileIds:
try:
managementInterface.updateFile(fid, {"scope": scope})
updated += 1
except Exception as e:
logger.error(f"bulk_set_scope: failed for file {fid}: {e}")
return {"scope": scope, "filesUpdated": updated}
except HTTPException:
raise
except Exception as e:
logger.error(f"bulk_set_scope error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.post("/bulk/neutralize")
@limiter.limit("30/minute")
def bulk_set_neutralize(
request: Request,
body: dict = Body(...),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
):
"""Set neutralize flag for a list of files by their IDs (incl. knowledge purge/reindex)."""
fileIds: list = body.get("fileIds") or []
neutralize = body.get("neutralize")
if not fileIds:
raise HTTPException(status_code=400, detail="fileIds is required")
if neutralize is None:
raise HTTPException(status_code=400, detail="neutralize is required")
try:
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
updated = 0
for fid in fileIds:
try:
managementInterface.updateFile(fid, {"neutralize": neutralize})
if not neutralize:
try:
from modules.interfaces import interfaceDbKnowledge
kIface = interfaceDbKnowledge.getInterface(currentUser)
kIface.purgeFileKnowledge(fid)
except Exception as ke:
logger.warning(f"bulk_set_neutralize: knowledge purge failed for {fid}: {ke}")
updated += 1
except Exception as e:
logger.error(f"bulk_set_neutralize: failed for file {fid}: {e}")
return {"neutralize": neutralize, "filesUpdated": updated}
except HTTPException:
raise
except Exception as e:
logger.error(f"bulk_set_neutralize error: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.post("/bulk/download-zip")
@limiter.limit("10/minute")
async def bulk_download_zip(
request: Request,
body: dict = Body(...),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext),
):
"""Download a list of files as a ZIP archive."""
import io, zipfile
fileIds: list = body.get("fileIds") or []
if not fileIds:
raise HTTPException(status_code=400, detail="fileIds is required")
try:
managementInterface = interfaceDbManagement.getInterface(
currentUser,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
buf = io.BytesIO()
with zipfile.ZipFile(buf, "w", zipfile.ZIP_DEFLATED) as zf:
for fid in fileIds:
try:
fileMeta = managementInterface.getFile(fid)
fileData = managementInterface.getFileData(fid)
if fileMeta and fileData:
name = (getattr(fileMeta, "fileName", None) or fid)
zf.writestr(name, fileData)
except Exception as fe:
logger.warning(f"bulk_download_zip: skipping file {fid}: {fe}")
buf.seek(0)
from fastapi.responses import StreamingResponse
return StreamingResponse(
buf,
media_type="application/zip",
headers={"Content-Disposition": 'attachment; filename="files.zip"'},
)
except HTTPException:
raise
except Exception as e:
logger.error(f"bulk_download_zip error: {e}")
raise HTTPException(status_code=500, detail=str(e))
# ── Scope & neutralize tagging endpoints (before /{fileId} catch-all) ─────────
@router.patch("/{fileId}/scope")
@limiter.limit("30/minute")
def updateFileScope(
request: Request,
background_tasks: BackgroundTasks,
fileId: str = Path(..., description="ID of the file"),
scope: str = Body(..., embed=True),
context: RequestContext = Depends(getRequestContext),
) -> Dict[str, Any]:
"""Update the scope of a file. Global scope requires sysAdmin."""
try:
validScopes = {"personal", "featureInstance", "mandate", "global"}
if scope not in validScopes:
raise HTTPException(status_code=400, detail=f"Invalid scope: {scope}. Must be one of {validScopes}")
if scope == "global" and not context.isSysAdmin:
raise HTTPException(status_code=403, detail=routeApiMsg("Only sysadmins can set global scope"))
managementInterface = interfaceDbManagement.getInterface(
context.user,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
managementInterface.updateFile(fileId, {"scope": scope})
try:
from modules.interfaces.interfaceDbKnowledge import getInterface as getKnowledgeInterface
from modules.datamodels.datamodelKnowledge import FileContentIndex
knowledgeDb = getKnowledgeInterface()
indices = knowledgeDb.db.getRecordset(FileContentIndex, recordFilter={"id": fileId})
for idx in indices:
idxId = idx.get("id") if isinstance(idx, dict) else getattr(idx, "id", None)
if idxId:
knowledgeDb.db.recordModify(FileContentIndex, idxId, {"scope": scope})
except Exception as e:
logger.warning(f"Failed to update FileContentIndex scope for file {fileId}: {e}")
fileMeta = managementInterface.getFile(fileId)
if fileMeta:
fn = fileMeta.fileName if hasattr(fileMeta, "fileName") else fileMeta.get("fileName", "")
mt = fileMeta.mimeType if hasattr(fileMeta, "mimeType") else fileMeta.get("mimeType", "")
async def _runReindexAfterScopeChange():
try:
await _autoIndexFile(
fileId=fileId, fileName=fn, mimeType=mt, user=context.user,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
except Exception as ex:
logger.warning("Re-index after scope change failed for %s: %s", fileId, ex)
background_tasks.add_task(_runReindexAfterScopeChange)
return {"fileId": fileId, "scope": scope, "updated": True}
except HTTPException:
raise
except Exception as e:
logger.error(f"Error updating file scope: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.patch("/{fileId}/neutralize")
@limiter.limit("30/minute")
def updateFileNeutralize(
request: Request,
background_tasks: BackgroundTasks,
fileId: str = Path(..., description="ID of the file"),
neutralize: bool = Body(..., embed=True),
context: RequestContext = Depends(getRequestContext),
) -> Dict[str, Any]:
"""Toggle neutralization flag on a file.
FAILSAFE: When turning neutralize ON, the existing Knowledge Store index
and all content chunks are deleted SYNCHRONOUSLY before the response is
returned. The re-index happens in a background task. If re-indexing
fails the file simply has no index — no un-neutralized data can leak.
"""
try:
managementInterface = interfaceDbManagement.getInterface(
context.user,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
managementInterface.updateFile(fileId, {"neutralize": neutralize})
from modules.interfaces.interfaceDbKnowledge import getInterface as getKnowledgeInterface
knowledgeDb = getKnowledgeInterface()
if neutralize:
# ── CRITICAL: purge existing (potentially un-neutralized) index
# This MUST succeed before the response is sent so that no stale
# raw-text chunks remain searchable while re-indexing runs.
try:
knowledgeDb.deleteFileContentIndex(fileId)
logger.info("Neutralize toggle ON: deleted index + chunks for file %s", fileId)
except Exception as e:
logger.error("Neutralize toggle ON: FAILED to delete index for file %s: %s", fileId, e)
raise HTTPException(
status_code=500,
detail=f"Could not purge existing index for neutralization — aborting toggle. Error: {e}",
)
else:
# Turning neutralize OFF: update metadata only; re-index will overwrite
try:
from modules.datamodels.datamodelKnowledge import FileContentIndex
indices = knowledgeDb.db.getRecordset(FileContentIndex, recordFilter={"id": fileId})
for idx in indices:
idxId = idx.get("id") if isinstance(idx, dict) else getattr(idx, "id", None)
if idxId:
knowledgeDb.db.recordModify(FileContentIndex, idxId, {
"neutralizationStatus": "original",
"isNeutralized": False,
})
except Exception as e:
logger.warning("Failed to update FileContentIndex after neutralize-OFF for %s: %s", fileId, e)
# Background re-index (safe: if it fails, there is simply no index)
fileMeta = managementInterface.getFile(fileId)
if fileMeta:
fn = fileMeta.fileName if hasattr(fileMeta, "fileName") else fileMeta.get("fileName", "")
mt = fileMeta.mimeType if hasattr(fileMeta, "mimeType") else fileMeta.get("mimeType", "")
async def _runReindexAfterNeutralizeToggle():
try:
await _autoIndexFile(
fileId=fileId, fileName=fn, mimeType=mt, user=context.user,
mandateId=str(context.mandateId) if context.mandateId else None,
featureInstanceId=str(context.featureInstanceId) if context.featureInstanceId else None,
)
except Exception as ex:
logger.error("Re-index after neutralize toggle failed for %s: %s (file has NO index until next re-index)", fileId, ex)
background_tasks.add_task(_runReindexAfterNeutralizeToggle)
return {"fileId": fileId, "neutralize": neutralize, "updated": True}
except HTTPException:
raise
except Exception as e:
logger.error(f"Error updating file neutralize flag: {e}")
raise HTTPException(status_code=500, detail=str(e))
# ── File endpoints with path parameters (catch-all /{fileId}) ─────────────────
@router.get("/{fileId}")
@limiter.limit("30/minute")
def get_file(
request: Request,
fileId: str = Path(..., description="ID of the file"),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext)
):
"""Get a file. Resolves the file's mandate/instance scope automatically."""
try:
_mgmt, fileData = _resolveFileWithScope(currentUser, context, fileId)
if not fileData:
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=f"File with ID {fileId} not found"
)
fileDict = fileData.model_dump() if hasattr(fileData, "model_dump") else dict(fileData)
enriched = enrichRowsWithFkLabels([fileDict], FileItem)
return enriched[0]
except interfaceDbManagement.FileNotFoundError as e:
logger.warning(f"File not found: {str(e)}")
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=str(e)
)
except interfaceDbManagement.FilePermissionError as e:
logger.warning(f"No permission for file: {str(e)}")
raise HTTPException(
status_code=status.HTTP_403_FORBIDDEN,
detail=str(e)
)
except interfaceDbManagement.FileError as e:
logger.error(f"Error retrieving file: {str(e)}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=str(e)
)
except Exception as e:
logger.error(f"Unexpected error retrieving file: {str(e)}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Error retrieving file: {str(e)}"
)
@router.put("/{fileId}", response_model=FileItem)
@limiter.limit("10/minute")
def update_file(
request: Request,
fileId: str = Path(..., description="ID of the file to update"),
file_info: Dict[str, Any] = Body(...),
currentUser: User = Depends(getCurrentUser)
) -> FileItem:
"""Update file info"""
try:
_EDITABLE_FIELDS = {"fileName", "folderId", "scope", "tags", "description", "neutralize"}
safeData = {k: v for k, v in file_info.items() if k in _EDITABLE_FIELDS}
if not safeData:
raise HTTPException(status_code=400, detail=routeApiMsg("No editable fields provided"))
managementInterface = interfaceDbManagement.getInterface(currentUser)
file = managementInterface.getFile(fileId)
if not file:
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=f"File with ID {fileId} not found"
)
if safeData.get("scope") == "global" and not getattr(currentUser, "isSysAdmin", False):
raise HTTPException(
status_code=status.HTTP_403_FORBIDDEN,
detail=routeApiMsg("Only sysadmins can set global scope"),
)
if not managementInterface.checkRbacPermission(FileItem, "update", fileId):
raise HTTPException(
status_code=status.HTTP_403_FORBIDDEN,
detail=routeApiMsg("Not authorized to update this file")
)
result = managementInterface.updateFile(fileId, safeData)
if not result:
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=routeApiMsg("Failed to update file")
)
# Get updated file
updatedFile = managementInterface.getFile(fileId)
return updatedFile
except HTTPException as he:
raise he
except Exception as e:
logger.error(f"Error updating file: {str(e)}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=str(e)
)
@router.delete("/{fileId}", response_model=Dict[str, Any])
@limiter.limit("10/minute")
def delete_file(
request: Request,
fileId: str = Path(..., description="ID of the file to delete"),
currentUser: User = Depends(getCurrentUser)
) -> Dict[str, Any]:
"""Delete a file"""
managementInterface = interfaceDbManagement.getInterface(currentUser)
# Check if the file exists
existingFile = managementInterface.getFile(fileId)
if not existingFile:
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=f"File with ID {fileId} not found"
)
success = managementInterface.deleteFile(fileId)
if not success:
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=routeApiMsg("Error deleting the file")
)
return {"message": f"File with ID {fileId} successfully deleted"}
@router.get("/{fileId}/download")
@limiter.limit("30/minute")
def download_file(
request: Request,
fileId: str = Path(..., description="ID of the file to download"),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext)
) -> Response:
"""Download a file. Resolves the file's mandate/instance scope automatically,
so direct <a href> links work even when X-Mandate-Id / X-Instance-Id headers
are not sent by the browser."""
try:
managementInterface, fileData = _resolveFileWithScope(currentUser, context, fileId)
if not fileData:
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=f"File with ID {fileId} not found"
)
fileContent = managementInterface.getFileData(fileId)
if not fileContent:
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=f"File content not found for ID {fileId}"
)
# Return file as response
# Properly encode filename for Content-Disposition header to handle Unicode characters
import urllib.parse
encoded_filename = urllib.parse.quote(fileData.fileName)
return Response(
content=fileContent,
media_type=fileData.mimeType,
headers={
"Content-Disposition": f"attachment; filename*=UTF-8''{encoded_filename}"
}
)
except HTTPException:
raise
except Exception as e:
logger.error(f"Error downloading file: {str(e)}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Error downloading file: {str(e)}"
)
@router.get("/{fileId}/preview", response_model=FilePreview)
@limiter.limit("30/minute")
def preview_file(
request: Request,
fileId: str = Path(..., description="ID of the file to preview"),
currentUser: User = Depends(getCurrentUser),
context: RequestContext = Depends(getRequestContext)
) -> FilePreview:
"""Preview a file's content. Resolves the file's mandate/instance scope automatically."""
try:
managementInterface, fileMeta = _resolveFileWithScope(currentUser, context, fileId)
if not fileMeta:
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=f"File with ID {fileId} not found"
)
preview = managementInterface.getFileContent(fileId)
if not preview:
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=f"File with ID {fileId} not found or no content available"
)
return preview
except HTTPException:
raise
except Exception as e:
logger.error(f"Error previewing file: {str(e)}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Error previewing file: {str(e)}"
)