Files
hkuds--lightrag/lightrag/api/routers/document_routes.py
T
2026-07-13 12:08:54 +08:00

3994 lines
164 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
"""
This module contains all document-related routes for the LightRAG API.
"""
import asyncio
import re
import shutil
import time
from uuid import uuid4
from lightrag.utils import (
logger,
get_pinyin_sort_key,
performance_timing_log,
validate_workspace,
)
import aiofiles
import traceback
from datetime import datetime, timezone
from pathlib import Path
from typing import Dict, List, Optional, Any, Literal
from fastapi import (
APIRouter,
BackgroundTasks,
Depends,
File,
HTTPException,
UploadFile,
)
from pydantic import BaseModel, ConfigDict, Field, field_validator, model_validator
from lightrag import LightRAG
from lightrag.base import DocProcessingStatus, DocStatus
from lightrag.constants import (
FILE_EXTRACTION_SUMMARY_PREFIX,
FULL_DOCS_FORMAT_PENDING_PARSE,
PARSED_ARTIFACT_DIR_SUFFIXES,
PARSED_DIR_NAME,
PROCESS_OPTION_CHUNK_FIXED,
PROCESS_OPTION_CHUNK_PARAGRAH,
PROCESS_OPTION_CHUNK_RECURSIVE,
PROCESS_OPTION_CHUNK_VECTOR,
)
from lightrag.parser.routing import (
FilenameParserHintError,
canonicalize_parser_hinted_basename,
chunk_strategy_key,
encode_parse_engine,
filename_parser_hint,
parse_process_options,
resolve_chunk_options,
resolve_parser_directives,
)
from lightrag.utils import (
generate_track_id,
move_file_to_parsed_dir,
)
from lightrag.api.utils_api import get_combined_auth_dependency
from ..config import global_args
# Function to format datetime to ISO format string with timezone information
def format_datetime(dt: Any) -> Optional[str]:
"""Format datetime to ISO format string with timezone information
Args:
dt: Datetime object, string, or None
Returns:
ISO format string with timezone information, or None if input is None
"""
if dt is None:
return None
if isinstance(dt, str):
return dt
# Check if datetime object has timezone information
if isinstance(dt, datetime):
# If datetime object has no timezone info (naive datetime), add UTC timezone
if dt.tzinfo is None:
dt = dt.replace(tzinfo=timezone.utc)
# Return ISO format string with timezone information
return dt.isoformat()
# NOTE: the APIRouter instance is created INSIDE `create_document_routes`
# (not at module scope). A module-level router is shared across processes,
# and re-running the factory — which the test suite does to validate
# create_app for different `--api-prefix` values — would re-decorate the
# same router each time, accumulating duplicate routes and triggering
# FastAPI's "Duplicate Operation ID" warnings.
# Temporary file prefix
temp_prefix = "__tmp__"
UNKNOWN_FILE_SOURCE = "unknown_source"
LEGACY_EMPTY_FILE_PATH_SENTINELS = {"", "no-file-path"}
ARCHIVED_FILE_SUFFIX_RE = re.compile(r"_(?:\d{3}|\d{10,})$")
def normalize_file_path(file_path: str | None) -> str:
"""Normalize missing document sources to a single non-null sentinel."""
if file_path is None:
return UNKNOWN_FILE_SOURCE
normalized = file_path.strip()
if normalized in LEGACY_EMPTY_FILE_PATH_SENTINELS:
return UNKNOWN_FILE_SOURCE
return canonicalize_parser_hinted_basename(normalized) or UNKNOWN_FILE_SOURCE
def is_valid_file_source(file_source: str | None) -> bool:
if file_source is None:
return False
return normalize_file_path(file_source) != UNKNOWN_FILE_SOURCE
def sanitize_filename(filename: str, input_dir: Path) -> str:
"""
Sanitize uploaded filename to prevent Path Traversal attacks.
Args:
filename: The original filename from the upload
input_dir: The target input directory
Returns:
str: Sanitized filename that is safe to use
Raises:
HTTPException: If the filename is unsafe or invalid
"""
# Basic validation
if not filename or not filename.strip():
raise HTTPException(status_code=400, detail="Filename cannot be empty")
# Remove path separators and traversal sequences
clean_name = filename.replace("/", "").replace("\\", "")
clean_name = clean_name.replace("..", "")
# Remove control characters and null bytes
clean_name = "".join(c for c in clean_name if ord(c) >= 32 and c != "\x7f")
# Remove leading/trailing whitespace and dots
clean_name = clean_name.strip().strip(".")
# Check if anything is left after sanitization
if not clean_name:
raise HTTPException(
status_code=400, detail="Invalid filename after sanitization"
)
# Verify the final path stays within the input directory
try:
final_path = (input_dir / clean_name).resolve()
if not final_path.is_relative_to(input_dir.resolve()):
raise HTTPException(status_code=400, detail="Unsafe filename detected")
except (OSError, ValueError):
raise HTTPException(status_code=400, detail="Invalid filename")
return clean_name
class ScanResponse(BaseModel):
"""Response model for document scanning operation
Attributes:
status: Status of the scanning operation. ``scanning_started`` when
a new background scan has been scheduled;
``scanning_skipped_pipeline_busy`` when the request was rejected
because indexing or another scan is already running.
message: Optional message with additional details
track_id: Tracking ID for monitoring scanning progress
"""
status: Literal["scanning_started", "scanning_skipped_pipeline_busy"] = Field(
description="Status of the scanning operation"
)
message: Optional[str] = Field(
default=None, description="Additional details about the scanning operation"
)
track_id: str = Field(description="Tracking ID for monitoring scanning progress")
model_config = ConfigDict(
json_schema_extra={
"example": {
"status": "scanning_started",
"message": "Scanning process has been initiated in the background",
"track_id": "scan_20250729_170612_abc123",
}
}
)
class ReprocessResponse(BaseModel):
"""Response model for reprocessing failed documents operation
Attributes:
status: Status of the reprocessing operation
message: Message describing the operation result
track_id: Always empty string. Reprocessed documents retain their original track_id.
"""
status: Literal["reprocessing_started"] = Field(
description="Status of the reprocessing operation"
)
message: str = Field(description="Human-readable message describing the operation")
track_id: str = Field(
default="",
description="Always empty string. Reprocessed documents retain their original track_id from initial upload.",
)
model_config = ConfigDict(
json_schema_extra={
"example": {
"status": "reprocessing_started",
"message": "Reprocessing of failed documents has been initiated in background",
"track_id": "",
}
}
)
class CancelPipelineResponse(BaseModel):
"""Response model for pipeline cancellation operation
Attributes:
status: Status of the cancellation request
message: Message describing the operation result
"""
status: Literal["cancellation_requested", "not_busy"] = Field(
description="Status of the cancellation request"
)
message: str = Field(description="Human-readable message describing the operation")
model_config = ConfigDict(
json_schema_extra={
"example": {
"status": "cancellation_requested",
"message": "Pipeline cancellation has been requested. Documents will be marked as FAILED.",
}
}
)
TextChunkingStrategy = Literal[
"fixed_token",
"recursive_character",
"semantic_vector",
"paragraph_semantic",
]
class _StrictChunkParams(BaseModel):
"""Base for per-strategy chunking params.
``strict=True`` rejects the Pydantic-v2 lax coercions that would
otherwise let malformed requests through and fail later in the
background chunker: bool-as-int (``true`` -> 1), numeric strings
(``"5"`` -> 5), float-as-int. ``extra="forbid"`` turns unknown keys
into a 422 (replacing a hand-rolled allow-list). ``chunk_token_size``
is shared by every strategy; ``None`` means "not supplied — fall back
to ``addon_params``/env default at process time".
"""
model_config = ConfigDict(extra="forbid", strict=True)
chunk_token_size: Optional[int] = Field(default=None, ge=1)
class _OverlapChunkParams(_StrictChunkParams):
chunk_overlap_token_size: Optional[int] = Field(default=None, ge=0)
@model_validator(mode="after")
def _overlap_lt_size(self) -> "_OverlapChunkParams":
# Only enforceable when BOTH are explicit; when chunk_token_size
# is None the effective size is resolved from addon_params/env at
# process time and can't be compared against here.
if (
self.chunk_token_size is not None
and self.chunk_overlap_token_size is not None
and self.chunk_overlap_token_size >= self.chunk_token_size
):
raise ValueError("chunk_overlap_token_size must be < chunk_token_size")
return self
class FixedTokenChunkParams(_OverlapChunkParams):
split_by_character: Optional[str] = None
split_by_character_only: Optional[bool] = None
class RecursiveCharacterChunkParams(_OverlapChunkParams):
separators: Optional[list[str]] = None
class ParagraphSemanticChunkParams(_OverlapChunkParams):
# Drop the trailing reference section before chunking. ``None`` means
# "not supplied — inherit the addon_params/env default at process time".
# Detection-tuning knobs (tail window / heading prefixes) are env-only and
# read live by the chunker, so they are intentionally not exposed here.
drop_references: Optional[bool] = None
class SemanticVectorChunkParams(_StrictChunkParams):
# Enum verified against the installed langchain_experimental
# (text_splitter.py ``BreakpointThresholdType``), not from memory.
breakpoint_threshold_type: Optional[
Literal["percentile", "standard_deviation", "interquartile", "gradient"]
] = None
# A strict ``float`` field still accepts an ``int`` (e.g. JSON ``95``) and
# widens it losslessly to ``95.0`` — strict only rejects ``str`` / ``bool``
# here, which is exactly what we want. Do NOT relax strict (that would let
# numeric strings through) or switch to ``int | float`` (that would stop
# normalizing ints to float). Locked by tests in test_document_routes_chunking.
breakpoint_threshold_amount: Optional[float] = None
buffer_size: Optional[int] = Field(default=None, ge=1)
sentence_split_regex: Optional[str] = None
@field_validator("sentence_split_regex")
@classmethod
def _valid_sentence_split_regex(cls, v: Optional[str]) -> Optional[str]:
# The value is fed to LangChain's SemanticChunker and compiled during
# split_text. A malformed pattern (e.g. "(") would only blow up in the
# background, so compile it here to reject synchronously (HTTP 422).
if v is None:
return v
try:
re.compile(v)
except re.error as exc:
raise ValueError(
f"sentence_split_regex is not a valid regular expression: {exc}"
) from exc
return v
@model_validator(mode="after")
def _amount_in_range(self) -> "SemanticVectorChunkParams":
amt = self.breakpoint_threshold_amount
if amt is None:
return self
# ``> 0`` is type-independent (every threshold type wants a positive
# magnitude), so it is safe to enforce at parse time.
if amt <= 0:
raise ValueError("breakpoint_threshold_amount must be > 0")
# The ``(0, 100]`` ceiling is percentile/gradient-specific (those feed
# np.percentile, which requires q in [0, 100]). It depends on the
# threshold TYPE, so only enforce it here when the type is supplied in
# the SAME request. When the type is omitted, the effective type is
# resolved from addon_params/env later — assuming "percentile" here
# would wrongly 422 a partial override that inherits
# standard_deviation/interquartile (which allow amounts > 100). The
# ceiling against the merged type is applied by
# ``_validate_effective_semantic_amount`` in ``_resolve_text_chunking``.
if self.breakpoint_threshold_type in ("percentile", "gradient") and amt > 100:
raise ValueError(
"breakpoint_threshold_amount must be within (0, 100] "
"for percentile/gradient"
)
return self
_CHUNKING_PARAMS_MODEL: dict[str, type[_StrictChunkParams]] = {
"fixed_token": FixedTokenChunkParams,
"recursive_character": RecursiveCharacterChunkParams,
"semantic_vector": SemanticVectorChunkParams,
"paragraph_semantic": ParagraphSemanticChunkParams,
}
class TextChunkingConfig(BaseModel):
"""Chunking strategy + strategy-specific params for a text insert.
Validation is delegated to the per-strategy typed model so unknown
keys, wrong types, and out-of-range values all raise synchronously
during request parsing (HTTP 422) — never later in the background
indexing task, where the HTTP response has already been sent.
"""
model_config = ConfigDict(extra="forbid")
strategy: TextChunkingStrategy = "fixed_token"
params: Dict[str, Any] = Field(default_factory=dict)
@model_validator(mode="after")
def _validate_params(self) -> "TextChunkingConfig":
typed = _CHUNKING_PARAMS_MODEL[self.strategy].model_validate(self.params)
# Normalize down to exactly the keys the caller supplied with a real
# value (validated + coerced) so the enqueue-time merge overrides only
# what was set. ``exclude_none`` additionally drops explicit nulls:
# every param field means "inherit the addon_params/env default" when
# None, so an explicit ``"chunk_token_size": null`` must NOT be merged
# over the resolved default — otherwise the route would 200 and the
# background chunker would do ``int(None)`` and fail the document.
self.params = typed.model_dump(exclude_unset=True, exclude_none=True)
return self
class InsertTextRequest(BaseModel):
"""Request model for inserting a single text document
Attributes:
text: The text content to be inserted into the RAG system
file_source: Source of the text (optional)
chunking: Optional chunking strategy + params; omit to keep the
default fixed-token behavior and addon_params defaults.
"""
text: str = Field(
min_length=1,
description="The text to insert",
)
file_source: Optional[str] = Field(
default=None, min_length=0, description="File Source"
)
chunking: Optional[TextChunkingConfig] = Field(
default=None,
description="Chunking strategy and params; omit for default fixed-token chunking",
)
@field_validator("text", mode="after")
@classmethod
def strip_text_after(cls, text: str) -> str:
return text.strip()
@field_validator("file_source", mode="before")
@classmethod
def normalize_source_before(cls, file_source: Optional[str]) -> str:
return normalize_file_path(file_source)
model_config = ConfigDict(
json_schema_extra={
"example": {
"text": "This is a sample text to be inserted into the RAG system.",
"file_source": "Source of the text (optional)",
"chunking": {
"strategy": "fixed_token",
"params": {
"chunk_token_size": 1200,
"chunk_overlap_token_size": 100,
"split_by_character": "\n\n",
"split_by_character_only": True,
},
},
}
}
)
class InsertTextsRequest(BaseModel):
"""Request model for inserting multiple text documents
Attributes:
texts: List of text contents to be inserted into the RAG system
file_sources: Sources of the texts (optional)
"""
texts: list[str] = Field(
min_length=1,
description="The texts to insert",
)
file_sources: Optional[list[str]] = Field(
default=None, min_length=0, description="Sources of the texts"
)
chunking: Optional[TextChunkingConfig] = Field(
default=None,
description="Shared chunking strategy and params for all texts; omit for default fixed-token chunking",
)
@field_validator("texts", mode="after")
@classmethod
def strip_texts_after(cls, texts: list[str]) -> list[str]:
return [text.strip() for text in texts]
@field_validator("file_sources", mode="before")
@classmethod
def normalize_sources_before(
cls, file_sources: Optional[list[str]]
) -> Optional[list[str]]:
if file_sources is None:
return None
return [normalize_file_path(file_source) for file_source in file_sources]
model_config = ConfigDict(
json_schema_extra={
"example": {
"texts": [
"This is the first text to be inserted.",
"This is the second text to be inserted.",
],
"file_sources": [
"First file source (optional)",
],
"chunking": {
"strategy": "recursive_character",
"params": {"chunk_token_size": 1000},
},
}
}
)
class InsertResponse(BaseModel):
"""Response model for document insertion operations
Attributes:
status: Status of the operation (success, partial_success, failure).
Same-name conflicts are rejected with HTTP 409 rather than being
reported as a "duplicated" 200 response, so this field never
takes that value any more.
message: Detailed message describing the operation result
track_id: Tracking ID for monitoring processing status
"""
status: Literal["success", "partial_success", "failure"] = Field(
description="Status of the operation"
)
message: str = Field(description="Message describing the operation result")
track_id: str = Field(description="Tracking ID for monitoring processing status")
model_config = ConfigDict(
json_schema_extra={
"example": {
"status": "success",
"message": "File 'document.pdf' uploaded successfully. Processing will continue in background.",
"track_id": "upload_20250729_170612_abc123",
}
}
)
class ClearDocumentsResponse(BaseModel):
"""Response model for document clearing operation
Attributes:
status: Status of the clear operation
message: Detailed message describing the operation result
"""
status: Literal["success", "partial_success", "busy", "fail"] = Field(
description="Status of the clear operation"
)
message: str = Field(description="Message describing the operation result")
model_config = ConfigDict(
json_schema_extra={
"example": {
"status": "success",
"message": "All documents cleared successfully. Deleted 15 files.",
}
}
)
class ClearCacheRequest(BaseModel):
"""Request model for clearing cache
This model is kept for API compatibility but no longer accepts any parameters.
All cache will be cleared regardless of the request content.
"""
model_config = ConfigDict(json_schema_extra={"example": {}})
class ClearCacheResponse(BaseModel):
"""Response model for cache clearing operation
Attributes:
status: Status of the clear operation
message: Detailed message describing the operation result
"""
status: Literal["success", "fail"] = Field(
description="Status of the clear operation"
)
message: str = Field(description="Message describing the operation result")
model_config = ConfigDict(
json_schema_extra={
"example": {
"status": "success",
"message": "Successfully cleared cache for modes: ['default', 'naive']",
}
}
)
"""Response model for document status
Attributes:
id: Document identifier
content_summary: Summary of document content
content_length: Length of document content
status: Current processing status
created_at: Creation timestamp (ISO format string)
updated_at: Last update timestamp (ISO format string)
chunks_count: Number of chunks (optional)
error: Error message if any (optional)
metadata: Additional metadata (optional)
file_path: Path to the document file
"""
class DeleteDocRequest(BaseModel):
doc_ids: List[str] = Field(..., description="The IDs of the documents to delete.")
delete_file: bool = Field(
default=False,
description="Whether to delete the corresponding file in the upload directory.",
)
delete_llm_cache: bool = Field(
default=False,
description="Whether to delete cached LLM extraction results for the documents.",
)
@field_validator("doc_ids", mode="after")
@classmethod
def validate_doc_ids(cls, doc_ids: List[str]) -> List[str]:
if not doc_ids:
raise ValueError("Document IDs list cannot be empty")
validated_ids = []
for doc_id in doc_ids:
if not doc_id or not doc_id.strip():
raise ValueError("Document ID cannot be empty")
validated_ids.append(doc_id.strip())
# Check for duplicates
if len(validated_ids) != len(set(validated_ids)):
raise ValueError("Document IDs must be unique")
return validated_ids
class DocStatusResponse(BaseModel):
id: str = Field(description="Document identifier")
content_summary: str = Field(description="Summary of document content")
content_length: int = Field(description="Length of document content in characters")
status: DocStatus = Field(description="Current processing status")
created_at: str = Field(description="Creation timestamp (ISO format string)")
updated_at: str = Field(description="Last update timestamp (ISO format string)")
track_id: Optional[str] = Field(
default=None, description="Tracking ID for monitoring progress"
)
chunks_count: Optional[int] = Field(
default=None, description="Number of chunks the document was split into"
)
error_msg: Optional[str] = Field(
default=None, description="Error message if processing failed"
)
metadata: Optional[dict[str, Any]] = Field(
default=None, description="Additional metadata about the document"
)
file_path: str = Field(description="Path to the document file")
model_config = ConfigDict(
json_schema_extra={
"example": {
"id": "doc_123456",
"content_summary": "Research paper on machine learning",
"content_length": 15240,
"status": "processed",
"created_at": "2025-03-31T12:34:56",
"updated_at": "2025-03-31T12:35:30",
"track_id": "upload_20250729_170612_abc123",
"chunks_count": 12,
"error": None,
"metadata": {"author": "John Doe", "year": 2025},
"file_path": "research_paper.pdf",
}
}
)
class DocsStatusesResponse(BaseModel):
"""Response model for document statuses
Attributes:
statuses: Dictionary mapping document status to lists of document status responses
"""
statuses: Dict[DocStatus, List[DocStatusResponse]] = Field(
default_factory=dict,
description="Dictionary mapping document status to lists of document status responses",
)
model_config = ConfigDict(
json_schema_extra={
"example": {
"statuses": {
"PENDING": [
{
"id": "doc_123",
"content_summary": "Pending document",
"content_length": 5000,
"status": "pending",
"created_at": "2025-03-31T10:00:00",
"updated_at": "2025-03-31T10:00:00",
"track_id": "upload_20250331_100000_abc123",
"chunks_count": None,
"error": None,
"metadata": None,
"file_path": "pending_doc.pdf",
}
],
"PREPROCESSED": [
{
"id": "doc_789",
"content_summary": "Document pending final indexing",
"content_length": 7200,
"status": "preprocessed",
"created_at": "2025-03-31T09:30:00",
"updated_at": "2025-03-31T09:35:00",
"track_id": "upload_20250331_093000_xyz789",
"chunks_count": 10,
"error": None,
"metadata": None,
"file_path": "preprocessed_doc.pdf",
}
],
"PROCESSED": [
{
"id": "doc_456",
"content_summary": "Processed document",
"content_length": 8000,
"status": "processed",
"created_at": "2025-03-31T09:00:00",
"updated_at": "2025-03-31T09:05:00",
"track_id": "insert_20250331_090000_def456",
"chunks_count": 8,
"error": None,
"metadata": {"author": "John Doe"},
"file_path": "processed_doc.pdf",
}
],
}
}
}
)
class TrackStatusResponse(BaseModel):
"""Response model for tracking document processing status by track_id
Attributes:
track_id: The tracking ID
documents: List of documents associated with this track_id
total_count: Total number of documents for this track_id
status_summary: Count of documents by status
"""
track_id: str = Field(description="The tracking ID")
documents: List[DocStatusResponse] = Field(
description="List of documents associated with this track_id"
)
total_count: int = Field(description="Total number of documents for this track_id")
status_summary: Dict[str, int] = Field(description="Count of documents by status")
model_config = ConfigDict(
json_schema_extra={
"example": {
"track_id": "upload_20250729_170612_abc123",
"documents": [
{
"id": "doc_123456",
"content_summary": "Research paper on machine learning",
"content_length": 15240,
"status": "PROCESSED",
"created_at": "2025-03-31T12:34:56",
"updated_at": "2025-03-31T12:35:30",
"track_id": "upload_20250729_170612_abc123",
"chunks_count": 12,
"error": None,
"metadata": {"author": "John Doe", "year": 2025},
"file_path": "research_paper.pdf",
}
],
"total_count": 1,
"status_summary": {"PROCESSED": 1},
}
}
)
class DocumentsRequest(BaseModel):
"""Request model for paginated document queries
Attributes:
status_filter: Legacy single-status filter, ignored when status_filters is set
status_filters: Filter by multiple document statuses, None for all statuses
page: Page number (1-based)
page_size: Number of documents per page (10-200)
sort_field: Field to sort by ('created_at', 'updated_at', 'id', 'file_path')
sort_direction: Sort direction ('asc' or 'desc')
"""
status_filter: Optional[DocStatus] = Field(
default=None,
description="Legacy single-status filter, ignored when status_filters is set",
)
status_filters: Optional[List[DocStatus]] = Field(
default=None, description="Filter by multiple document statuses"
)
page: int = Field(default=1, ge=1, description="Page number (1-based)")
page_size: int = Field(
default=50, ge=10, le=200, description="Number of documents per page (10-200)"
)
sort_field: Literal["created_at", "updated_at", "id", "file_path"] = Field(
default="updated_at", description="Field to sort by"
)
sort_direction: Literal["asc", "desc"] = Field(
default="desc", description="Sort direction"
)
model_config = ConfigDict(
json_schema_extra={
"example": {
"status_filters": ["PREPROCESSED", "PARSING", "ANALYZING"],
"page": 1,
"page_size": 50,
"sort_field": "updated_at",
"sort_direction": "desc",
}
}
)
class PaginationInfo(BaseModel):
"""Pagination information
Attributes:
page: Current page number
page_size: Number of items per page
total_count: Total number of items
total_pages: Total number of pages
has_next: Whether there is a next page
has_prev: Whether there is a previous page
"""
page: int = Field(description="Current page number")
page_size: int = Field(description="Number of items per page")
total_count: int = Field(description="Total number of items")
total_pages: int = Field(description="Total number of pages")
has_next: bool = Field(description="Whether there is a next page")
has_prev: bool = Field(description="Whether there is a previous page")
model_config = ConfigDict(
json_schema_extra={
"example": {
"page": 1,
"page_size": 50,
"total_count": 150,
"total_pages": 3,
"has_next": True,
"has_prev": False,
}
}
)
class PaginatedDocsResponse(BaseModel):
"""Response model for paginated document queries
Attributes:
documents: List of documents for the current page
pagination: Pagination information
status_counts: Count of documents by status for all documents
"""
documents: List[DocStatusResponse] = Field(
description="List of documents for the current page"
)
pagination: PaginationInfo = Field(description="Pagination information")
status_counts: Dict[str, int] = Field(
description="Count of documents by status for all documents"
)
model_config = ConfigDict(
json_schema_extra={
"example": {
"documents": [
{
"id": "doc_123456",
"content_summary": "Research paper on machine learning",
"content_length": 15240,
"status": "PROCESSED",
"created_at": "2025-03-31T12:34:56",
"updated_at": "2025-03-31T12:35:30",
"track_id": "upload_20250729_170612_abc123",
"chunks_count": 12,
"error_msg": None,
"metadata": {"author": "John Doe", "year": 2025},
"file_path": "research_paper.pdf",
}
],
"pagination": {
"page": 1,
"page_size": 50,
"total_count": 150,
"total_pages": 3,
"has_next": True,
"has_prev": False,
},
"status_counts": {
"PENDING": 10,
"PROCESSING": 5,
"PREPROCESSED": 5,
"PROCESSED": 130,
"FAILED": 5,
},
}
}
)
class StatusCountsResponse(BaseModel):
"""Response model for document status counts
Attributes:
status_counts: Count of documents by status
"""
status_counts: Dict[str, int] = Field(description="Count of documents by status")
model_config = ConfigDict(
json_schema_extra={
"example": {
"status_counts": {
"PENDING": 10,
"PROCESSING": 5,
"PREPROCESSED": 5,
"PROCESSED": 130,
"FAILED": 5,
}
}
}
)
class PipelineStatusResponse(BaseModel):
"""Response model for pipeline status
Attributes:
autoscanned: Whether auto-scan has started
busy: Whether the pipeline is currently busy
job_name: Current job name (e.g., indexing files/indexing texts)
job_start: Job start time as ISO format string with timezone (optional)
docs: Total number of documents to be indexed
batchs: Number of batches for processing documents
cur_batch: Current processing batch
request_pending: Flag for pending request for processing
latest_message: Latest message from pipeline processing
history_messages: List of history messages
update_status: Status of update flags for all namespaces
"""
autoscanned: bool = False
busy: bool = False
job_name: str = "Default Job"
job_start: Optional[str] = None
docs: int = 0
batchs: int = 0
cur_batch: int = 0
request_pending: bool = False
latest_message: str = ""
history_messages: Optional[List[str]] = None
update_status: Optional[dict] = None
@field_validator("job_start", mode="before")
@classmethod
def parse_job_start(cls, value):
"""Process datetime and return as ISO format string with timezone"""
return format_datetime(value)
model_config = ConfigDict(extra="allow")
class DocumentManager:
def __init__(
self,
input_dir: str,
workspace: str = "", # New parameter for workspace isolation
):
# Reject path traversal before using workspace in the upload path
validate_workspace(workspace)
# Store the base input directory and workspace
self.base_input_dir = Path(input_dir)
self.workspace = workspace
self.indexed_files = set()
# Create workspace-specific input directory
# If workspace is provided, create a subdirectory for data isolation
if workspace:
self.input_dir = self.base_input_dir / workspace
else:
self.input_dir = self.base_input_dir
# Create input directory if it doesn't exist
self.input_dir.mkdir(parents=True, exist_ok=True)
@property
def supported_extensions(self) -> tuple:
"""Suffixes accepted for an unhinted filename, derived live.
A suffix is advertised only when it is *routable without extra
directives*: the engine that ``resolve_file_parser_engine`` picks for
a bare ``x.<suffix>`` (filename hint absent; ``LIGHTRAG_PARSER``
rules + default apply) must itself support the suffix. This keeps
"uploadable" aligned with "will actually parse": e.g. mineru's
``png`` joins only when its endpoint is configured AND a routing
rule (or per-file hint, see ``is_supported_file``) sends pngs to it
— otherwise the default ``legacy`` engine would fail the suffix gate
at the parse stage. A default deployment equals the local engines'
(legacy native) types; no hardcoded list to keep in sync.
"""
from lightrag.parser.registry import available_engine_suffixes
from lightrag.parser.routing import (
parser_engine_supports_suffix,
resolve_file_parser_engine,
)
out = []
for s in sorted(available_engine_suffixes()):
engine = resolve_file_parser_engine(f"x.{s}")
if parser_engine_supports_suffix(engine, s):
out.append(f".{s}")
return tuple(out)
def scan_directory_for_new_files(self) -> List[Path]:
"""Scan input directory for new, routable files.
Globs over every *available* engine suffix (capability surface, so a
hint-carrying file like ``img.[mineru].png`` is discoverable even
when bare ``.png`` is not advertised), then keeps only files whose
resolved engine actually supports them (``is_supported_file``).
"""
from lightrag.parser.registry import available_engine_suffixes
from lightrag.parser.routing import FilenameParserHintError
new_files = []
for s in sorted(available_engine_suffixes()):
ext = f".{s}"
logger.debug(f"Scanning for {ext} files in {self.input_dir}")
for file_path in self.input_dir.glob(f"*{ext}"):
if file_path in self.indexed_files:
continue
try:
if not self.is_supported_file(file_path.name):
continue
except FilenameParserHintError:
# Malformed hint: pass the file through — the enqueue
# path reports a detailed error document, instead of the
# scan silently ignoring the user's file.
pass
new_files.append(file_path)
return new_files
def mark_as_indexed(self, file_path: Path):
self.indexed_files.add(file_path)
def is_supported_file(self, filename: str) -> bool:
"""True when THIS filename routes to an engine that can parse it.
Resolves the engine for the concrete name — so a per-file hint
(``img.[mineru].png``) is honoured — and checks the resolved engine
supports the suffix. A bare suffix that would fall through to the
default ``legacy`` engine is rejected here instead of failing later
at the parse worker's suffix gate.
Raises :class:`FilenameParserHintError` for a malformed hint —
callers surface it (upload → HTTP 400 with the detailed message;
scan passes the file through so enqueue emits an error document).
"""
from lightrag.parser.routing import (
parser_engine_supports_suffix,
parser_suffix,
resolve_file_parser_engine,
)
engine = resolve_file_parser_engine(filename)
return parser_engine_supports_suffix(engine, parser_suffix(filename))
def validate_file_path_security(file_path_str: str, base_dir: Path) -> Optional[Path]:
"""
Validate file path security to prevent Path Traversal attacks.
Args:
file_path_str: The file path string to validate
base_dir: The base directory that the file must be within
Returns:
Path: Safe file path if valid, None if unsafe or invalid
"""
if not file_path_str or not file_path_str.strip():
return None
try:
# Clean the file path string
clean_path_str = file_path_str.strip()
# Check for obvious path traversal patterns before processing
# This catches both Unix (..) and Windows (..\) style traversals
if ".." in clean_path_str:
# Additional check for Windows-style backslash traversal
if (
"\\..\\" in clean_path_str
or clean_path_str.startswith("..\\")
or clean_path_str.endswith("\\..")
):
# logger.warning(
# f"Security violation: Windows path traversal attempt detected - {file_path_str}"
# )
return None
# Normalize path separators (convert backslashes to forward slashes)
# This helps handle Windows-style paths on Unix systems
normalized_path = clean_path_str.replace("\\", "/")
# Create path object and resolve it (handles symlinks and relative paths)
candidate_path = (base_dir / normalized_path).resolve()
base_dir_resolved = base_dir.resolve()
# Check if the resolved path is within the base directory
if not candidate_path.is_relative_to(base_dir_resolved):
# logger.warning(
# f"Security violation: Path traversal attempt detected - {file_path_str}"
# )
return None
return candidate_path
except (OSError, ValueError, Exception) as e:
logger.warning(f"Invalid file path detected: {file_path_str} - {str(e)}")
return None
def get_doc_status_value(doc_status: Any) -> str:
"""Read status from dict or DocProcessingStatus-like objects."""
status = (
doc_status.get("status")
if isinstance(doc_status, dict)
else getattr(doc_status, "status", None)
)
if isinstance(status, DocStatus):
return status.value
return str(status or "")
def get_doc_track_id(doc_status: Any) -> str:
"""Read track_id from dict or DocProcessingStatus-like objects."""
track_id = (
doc_status.get("track_id")
if isinstance(doc_status, dict)
else getattr(doc_status, "track_id", None)
)
return str(track_id or "")
async def get_existing_doc_by_file_path_candidates(
doc_status: Any, file_path: Path | str
) -> dict[str, Any] | None:
"""Find an existing document by canonical basename."""
basename = normalize_file_path(str(file_path))
if basename == UNKNOWN_FILE_SOURCE:
return None
match = await doc_status.get_doc_by_file_basename(basename)
if not match:
return None
_, existing_doc_data = match
return existing_doc_data
async def _reserve_enqueue_slot(rag: LightRAG) -> bool:
"""Atomically check exclusive-writer state and reserve a
pending-enqueue slot.
Concurrent enqueues are permitted while the processing loop is
running — the loop is notified via ``request_pending`` and picks up
newly-enqueued docs after its current batch. This includes the
scan task's processing phase: once classification is done, the
scan transitions to driving the processing pipeline like any
other enqueuer, and uploads can land alongside it.
Two states block new uploads/inserts:
- ``scanning_exclusive``: scan task is in its CLASSIFICATION
phase — reading doc_status to classify files (PROCESSED →
archive, FAILED-without-full_docs → retry-as-new, etc.) and
possibly deleting stale stubs. Concurrent enqueue would race
against scan's reads / stub deletions. ``scanning`` alone
(the processing phase) does NOT block uploads.
- ``destructive_busy``: a /documents/clear or per-doc delete is in
flight. These DROP storages and remove input files; an enqueue
accepted in this window would write to a storage that is being
torn down and silently lose the document after the client saw
success.
``pending_enqueues`` is incremented so the scan endpoint can refuse
while bg tasks are mid-enqueue. The counter does NOT gate
``apipeline_process_enqueue_documents`` — concurrent processing is
explicitly allowed and is what makes "upload while pipeline is
busy" possible.
A workspace whose ``pipeline_status`` has never been initialised
(mocked test rigs) is treated as idle; no slot is reserved.
Returns:
True when a slot was reserved (caller MUST pair with
``_release_enqueue_slot``); False when pipeline_status is not
bootstrapped.
Raises:
HTTPException(409): when
``pipeline_status['scanning_exclusive']`` or
``pipeline_status['destructive_busy']`` is set.
"""
from lightrag.exceptions import PipelineNotInitializedError
from lightrag.kg.shared_storage import get_namespace_data, get_namespace_lock
try:
pipeline_status = await get_namespace_data(
"pipeline_status", workspace=rag.workspace
)
except PipelineNotInitializedError:
return False
pipeline_status_lock = get_namespace_lock(
"pipeline_status", workspace=rag.workspace
)
async with pipeline_status_lock:
if pipeline_status.get("scanning_exclusive"):
raise HTTPException(
status_code=409,
detail=(
"Document scan is classifying files. "
"Wait for the classification phase to finish before "
"submitting new work."
),
)
if pipeline_status.get("destructive_busy"):
raise HTTPException(
status_code=409,
detail=(
"Pipeline is clearing or deleting documents. "
"Wait for the running job to finish before submitting "
"new work."
),
)
pipeline_status["pending_enqueues"] = (
pipeline_status.get("pending_enqueues", 0) + 1
)
return True
async def check_pipeline_busy_or_raise(rag: LightRAG) -> None:
"""Refuse the request with HTTP 409 when the document pipeline is busy.
Intended for short, fine-grained graph mutations (entity/relation
edit/create/delete/merge). Reads ``pipeline_status['busy']`` under
the namespace lock and raises immediately on contention -- it does
NOT set any flag, so it cannot block the pipeline itself.
``busy`` is set by the processing loop and by destructive jobs
(``/documents/clear`` / per-doc delete). Both paths concurrently
write the same graph storages that these endpoints mutate, so a
409 here mirrors the existing UI guard and tells clients to wait.
A narrow race remains between this check and the underlying graph
write: if the pipeline transitions to busy in that window, the
per-edge/-node locks inside the storage layer are the last line of
defense. That trade-off is deliberate -- holding ``busy`` here
would serialise every UI edit against document ingestion, which is
a worse user-visible failure mode than tolerating the race.
No-op (returns silently) when ``pipeline_status`` was never
bootstrapped, matching the behaviour of ``_acquire_destructive_busy``
so test rigs without a real shared-storage Manager keep working.
"""
from lightrag.exceptions import PipelineNotInitializedError
from lightrag.kg.shared_storage import get_namespace_data, get_namespace_lock
try:
pipeline_status = await get_namespace_data(
"pipeline_status", workspace=rag.workspace
)
except PipelineNotInitializedError:
return
pipeline_status_lock = get_namespace_lock(
"pipeline_status", workspace=rag.workspace
)
async with pipeline_status_lock:
if pipeline_status.get("busy"):
raise HTTPException(
status_code=409,
detail=(
"Pipeline is busy with another operation. "
"Wait for the running job to finish before editing "
"the knowledge graph."
),
)
async def _acquire_destructive_busy(rag: LightRAG) -> tuple[bool, str | None]:
"""Atomically reserve the destructive busy slot for ``/documents/clear``
or ``/documents/delete_document``.
Both jobs DROP storages and (for clear) remove input files. They
must serialise against:
- any other ``busy`` work (processing loop, another destructive job),
- an in-flight ``scanning`` task that reads/writes doc_status and
INPUT/, and
- any ``pending_enqueues`` reservation whose bg task has not yet
written to doc_status — accepting the destructive job in that
window would drop storages while the enqueue is mid-write,
losing a document the client already saw success for.
All three checks happen inside a single ``pipeline_status_lock``
critical section together with the flag write, so a concurrent
enqueue/scan reservation cannot squeeze past us.
Caller is responsible for clearing both flags in its finally block.
Returns:
(acquired, reason). ``acquired=True`` and ``reason=None`` on
success. ``acquired=False`` with a human-readable ``reason``
when another writer has the lock; the caller surfaces this to
the client (HTTP 200 with status="busy" for these endpoints).
For test rigs where ``pipeline_status`` was never bootstrapped,
returns (True, None) — there is nothing to coordinate against.
"""
from lightrag.exceptions import PipelineNotInitializedError
from lightrag.kg.shared_storage import get_namespace_data, get_namespace_lock
try:
pipeline_status = await get_namespace_data(
"pipeline_status", workspace=rag.workspace
)
except PipelineNotInitializedError:
return True, None
pipeline_status_lock = get_namespace_lock(
"pipeline_status", workspace=rag.workspace
)
async with pipeline_status_lock:
if pipeline_status.get("busy"):
return False, "Pipeline is busy with another operation."
if pipeline_status.get("scanning"):
return False, (
"Document scan is in progress. "
"Wait for the scan to complete before clearing or deleting."
)
if pipeline_status.get("pending_enqueues", 0) > 0:
return False, (
"Document upload/insert is being enqueued. "
"Wait for in-flight work to complete before clearing or "
"deleting."
)
pipeline_status["busy"] = True
pipeline_status["destructive_busy"] = True
return True, None
async def _release_destructive_busy(rag: LightRAG) -> None:
"""Release the destructive busy slot acquired by
``_acquire_destructive_busy``. Never raises.
Distinct from ``_release_enqueue_slot``: that helper clears
``pending_enqueues`` (the upload/insert reservation), this one
clears ``busy + destructive_busy`` (the clear/delete reservation).
"""
from lightrag.exceptions import PipelineNotInitializedError
from lightrag.kg.shared_storage import get_namespace_data, get_namespace_lock
try:
pipeline_status = await get_namespace_data(
"pipeline_status", workspace=rag.workspace
)
except PipelineNotInitializedError:
return
pipeline_status_lock = get_namespace_lock(
"pipeline_status", workspace=rag.workspace
)
async with pipeline_status_lock:
pipeline_status["busy"] = False
pipeline_status["destructive_busy"] = False
async def _release_enqueue_slot(rag: LightRAG) -> None:
"""Release a slot reserved by ``_reserve_enqueue_slot``.
Pure decrement; the bg task itself drives processing by calling
``apipeline_process_enqueue_documents`` after enqueue (the call is
a cheap no-op when the loop is already busy — it just sets
``request_pending``). Drain coordination across sibling bg tasks
is unnecessary in the new contract: each task triggers processing
independently and the loop's request_pending mechanism collapses
duplicate triggers safely.
Decrement is clamped at 0 so a stray release (e.g. from a workspace
whose reservation returned False but whose bg task wrapper still
calls release) is harmless. Never raises.
"""
from lightrag.exceptions import PipelineNotInitializedError
from lightrag.kg.shared_storage import get_namespace_data, get_namespace_lock
try:
pipeline_status = await get_namespace_data(
"pipeline_status", workspace=rag.workspace
)
except PipelineNotInitializedError:
return
pipeline_status_lock = get_namespace_lock(
"pipeline_status", workspace=rag.workspace
)
async with pipeline_status_lock:
current = pipeline_status.get("pending_enqueues", 0)
if current > 0:
pipeline_status["pending_enqueues"] = current - 1
def find_existing_file_by_file_path(input_dir: Path, file_path: str) -> Path | None:
"""Find an input-dir file whose canonical basename matches ``file_path``.
Callers pass the stored canonical ``file_path`` (already hint-stripped);
on-disk filenames are normalized before comparison so a hint-bearing
variant on disk still matches a canonical stored ``file_path``.
"""
if not file_path or file_path == UNKNOWN_FILE_SOURCE:
return None
try:
for candidate in input_dir.iterdir():
if not candidate.is_file():
continue
if normalize_file_path(candidate.name) == file_path:
return candidate
except FileNotFoundError:
return None
return None
def canonicalize_archived_file_variant_basename(
file_path: Path | str, *, strip_archive_suffix: bool = False
) -> str:
"""Canonical basename for original files and numbered archive variants."""
name = Path(file_path).name
path = Path(name)
stem = (
ARCHIVED_FILE_SUFFIX_RE.sub("", path.stem)
if strip_archive_suffix
else path.stem
)
return normalize_file_path(f"{stem}{path.suffix}")
def _file_path_for_parsed_artifact_dir(dir_name: str) -> str | None:
"""Return the canonical source basename for a parser artifact dir.
Recognized layouts (suffix list in
:data:`lightrag.constants.PARSED_ARTIFACT_DIR_SUFFIXES`):
- ``<basename>.parsed[_NNN]/`` — sidecar output (every engine)
- ``<basename>.mineru_raw[_NNN]/`` — MinerU preserved raw bundle
- ``<basename>.docling_raw[_NNN]/`` — Docling preserved raw bundle
Raw bundles are preserved across re-parses for cache reuse and on-demand
diagnostics; they are cleaned only when the user deletes the document
with ``delete_file=True`` so the raw artifacts and source file go away
together.
"""
stripped = ARCHIVED_FILE_SUFFIX_RE.sub("", dir_name)
for suffix in PARSED_ARTIFACT_DIR_SUFFIXES:
if stripped.endswith(suffix):
basename = stripped[: -len(suffix)]
if basename:
return normalize_file_path(basename)
return None
def delete_file_variants_by_file_path(
input_dir: Path,
file_path: str | None,
) -> tuple[list[str], list[str]]:
"""Delete input/__parsed__ source files matching a canonical ``file_path``."""
if not file_path:
return [], []
canonical = normalize_file_path(file_path)
if canonical == UNKNOWN_FILE_SOURCE:
return [], []
canonical_names = {canonical}
deleted_files: list[str] = []
errors: list[str] = []
candidate_dirs = [input_dir, input_dir / PARSED_DIR_NAME]
input_dir_resolved = input_dir.resolve()
for candidate_dir in candidate_dirs:
try:
candidates = list(candidate_dir.iterdir())
except FileNotFoundError:
continue
except Exception as e:
errors.append(f"Failed to scan {candidate_dir}: {e}")
continue
in_parsed_dir = candidate_dir.name == PARSED_DIR_NAME
for candidate in candidates:
if candidate.is_file():
if (
canonicalize_archived_file_variant_basename(
candidate.name,
strip_archive_suffix=in_parsed_dir,
)
not in canonical_names
):
continue
safe_candidate = validate_file_path_security(
candidate.name, candidate_dir
)
if safe_candidate is None:
errors.append(f"Unsafe file path skipped: {candidate.name}")
continue
try:
safe_candidate.unlink()
deleted_files.append(
str(safe_candidate.relative_to(input_dir_resolved))
)
except Exception as e:
errors.append(f"Failed to delete {candidate.name}: {e}")
continue
if in_parsed_dir and candidate.is_dir():
canonical_for_dir = _file_path_for_parsed_artifact_dir(candidate.name)
if (
canonical_for_dir is None
or canonical_for_dir not in canonical_names
):
continue
safe_candidate = validate_file_path_security(
candidate.name, candidate_dir
)
if safe_candidate is None:
errors.append(f"Unsafe artifact dir skipped: {candidate.name}")
continue
try:
shutil.rmtree(safe_candidate)
deleted_files.append(
str(safe_candidate.relative_to(input_dir_resolved))
)
except Exception as e:
errors.append(
f"Failed to delete artifact dir {candidate.name}: {e}"
)
return deleted_files, errors
async def record_scan_warning(rag: LightRAG, message: str) -> None:
logger.warning(message)
try:
from lightrag.kg import shared_storage
if not getattr(shared_storage, "_initialized", False):
return
workspace = getattr(rag, "workspace", "")
pipeline_status = await shared_storage.get_namespace_data(
"pipeline_status", workspace=workspace
)
pipeline_status_lock = shared_storage.get_namespace_lock(
"pipeline_status", workspace=workspace
)
async with pipeline_status_lock:
pipeline_status["latest_message"] = message
pipeline_status["history_messages"].append(message)
except Exception:
pass
# Legacy text extractors moved to lightrag.parser.legacy.extractors; the
# legacy engine now extracts at the worker stage (LegacyParser), not here.
async def pipeline_enqueue_file(
rag: LightRAG,
file_path: Path,
track_id: str = None,
from_scan: bool = False,
) -> tuple[bool, str]:
"""Add a file to the queue for processing
Args:
rag: LightRAG instance
file_path: Path to the saved file
track_id: Optional tracking ID, if not provided will be generated
from_scan: True only when invoked by the scan-owned background task,
which already holds ``pipeline_status["scanning"]``. Forwarded to
``apipeline_enqueue_documents`` so the scan can enqueue the files
it just discovered without tripping the scanning guard there.
Returns:
tuple: (success: bool, track_id: str)
"""
# Generate track_id if not provided
if track_id is None:
track_id = generate_track_id("unknown")
try:
file_size = 0
# Get file size for error reporting
try:
stat = await asyncio.to_thread(file_path.stat)
file_size = stat.st_size
except Exception:
file_size = 0
try:
directives = resolve_parser_directives(file_path)
except FilenameParserHintError as e:
error_files = [
{
"file_path": str(file_path.name),
"error_description": FILE_EXTRACTION_SUMMARY_PREFIX
+ "Filename hint error",
"original_error": str(e),
"file_size": file_size,
}
]
await rag.apipeline_enqueue_error_documents(error_files, track_id)
logger.error(
f"[File Extraction]Invalid filename hint in {file_path.name}: {e}"
)
return False, track_id
extraction_engine = directives.engine
process_options = directives.process_options
api_process_options = process_options or PROCESS_OPTION_CHUNK_FIXED
# Overlay any per-file chunk parameters (from the filename hint or a
# LIGHTRAG_PARSER rule) onto the active strategy's chunk_options so the
# parse worker chunks this document with them. Absent params keep the
# legacy path (chunk_options built at enqueue time from addon_params).
hint_chunk_options = None
active_strategy = parse_process_options(api_process_options).chunking
hint_chunk_params = directives.chunk_params.get(active_strategy)
if hint_chunk_params:
try:
strategy_key = chunk_strategy_key(api_process_options)
hint_chunk_options = resolve_chunk_options(
rag.addon_params, process_options=api_process_options
)
hint_chunk_options[strategy_key].update(hint_chunk_params)
_validate_effective_chunk_overlap(
hint_chunk_options, strategy_key, strategy_key
)
except ValueError as e:
error_files = [
{
"file_path": str(file_path.name),
"error_description": FILE_EXTRACTION_SUMMARY_PREFIX
+ "Chunk parameter error",
"original_error": str(e),
"file_size": file_size,
}
]
await rag.apipeline_enqueue_error_documents(error_files, track_id)
logger.error(
f"[File Extraction]Invalid chunk parameters in "
f"{file_path.name}: {e}"
)
return False, track_id
# All engines defer parsing to the worker stage: the file is already
# saved on disk, so we enqueue PENDING_PARSE with the chosen engine.
# Legacy now extracts at the worker (LegacyParser) instead of eagerly
# here, so every engine shares one ingestion path.
# Encode any per-file engine params into the parse_engine field
# (e.g. "mineru(page_range=1-3,language=en)") so they ride the existing
# persisted column to the parse worker. Bare engine when there are none.
parse_engine_field = encode_parse_engine(
extraction_engine, directives.engine_params
)
try:
enqueue_kwargs = {
"file_paths": str(file_path),
"track_id": track_id,
"docs_format": FULL_DOCS_FORMAT_PENDING_PARSE,
"parse_engine": parse_engine_field,
"process_options": api_process_options,
"from_scan": from_scan,
}
if hint_chunk_options is not None:
enqueue_kwargs["chunk_options"] = hint_chunk_options
enqueue_result = await rag.apipeline_enqueue_documents("", **enqueue_kwargs)
if enqueue_result is None:
try:
await move_file_to_parsed_dir(file_path)
except Exception as move_error:
logger.error(
f"Failed to move duplicate file {file_path.name} to {PARSED_DIR_NAME} directory: {move_error}"
)
return False, track_id
logger.info(
f"[File Extraction]Deferred {file_path.name} to {extraction_engine} parser"
)
return True, track_id
except Exception as e:
error_files = [
{
"file_path": str(file_path.name),
"error_description": FILE_EXTRACTION_SUMMARY_PREFIX
+ "Parser enqueue error",
"original_error": f"Failed to enqueue file for parser: {str(e)}",
"file_size": file_size,
}
]
await rag.apipeline_enqueue_error_documents(error_files, track_id)
logger.error(
f"[File Extraction]Error enqueuing {file_path.name} for {extraction_engine}: {str(e)}"
)
return False, track_id
except Exception as e:
# Catch-all for any unexpected errors
try:
file_size = file_path.stat().st_size if file_path.exists() else 0
except Exception:
file_size = 0
error_files = [
{
"file_path": str(file_path.name),
"error_description": "Unexpected processing error",
"original_error": f"Unexpected error: {str(e)}",
"file_size": file_size,
}
]
await rag.apipeline_enqueue_error_documents(error_files, track_id)
logger.error(f"Enqueuing file {file_path.name} error: {str(e)}")
logger.error(traceback.format_exc())
return False, track_id
finally:
if file_path.name.startswith(temp_prefix):
try:
file_path.unlink()
except Exception as e:
logger.error(f"Error deleting file {file_path}: {str(e)}")
async def pipeline_index_file(rag: LightRAG, file_path: Path, track_id: str = None):
"""Index a file with track_id
Args:
rag: LightRAG instance
file_path: Path to the saved file
track_id: Optional tracking ID
"""
try:
success, _ = await pipeline_enqueue_file(rag, file_path, track_id)
if success:
await rag.apipeline_process_enqueue_documents()
except Exception as e:
logger.error(f"Error indexing file {file_path.name}: {str(e)}")
logger.error(traceback.format_exc())
async def pipeline_index_files(
rag: LightRAG,
file_paths: List[Path],
track_id: str = None,
from_scan: bool = False,
):
"""Index multiple files sequentially to avoid high CPU load
Args:
rag: LightRAG instance
file_paths: Paths to the files to index
track_id: Optional tracking ID to pass to all files
from_scan: True only when invoked by the scan-owned background task.
Forwarded to ``pipeline_enqueue_file`` so the per-file enqueue
calls bypass the scanning guard inside
``apipeline_enqueue_documents`` (whose ``scanning`` flag the
scan task itself owns).
"""
if not file_paths:
return
try:
enqueued = False
# Use get_pinyin_sort_key for Chinese pinyin sorting
sorted_file_paths = sorted(
file_paths, key=lambda p: get_pinyin_sort_key(str(p))
)
# Process files sequentially with track_id
for file_path in sorted_file_paths:
success, _ = await pipeline_enqueue_file(
rag,
file_path,
track_id,
from_scan=from_scan,
)
if success:
enqueued = True
# Process the queue only if at least one file was successfully enqueued
if enqueued:
await rag.apipeline_process_enqueue_documents()
except Exception as e:
logger.error(f"Error indexing files: {str(e)}")
logger.error(traceback.format_exc())
_STRATEGY_TO_PROCESS_OPTION: Dict[str, str] = {
"fixed_token": PROCESS_OPTION_CHUNK_FIXED,
"recursive_character": PROCESS_OPTION_CHUNK_RECURSIVE,
"semantic_vector": PROCESS_OPTION_CHUNK_VECTOR,
"paragraph_semantic": PROCESS_OPTION_CHUNK_PARAGRAH,
}
def _resolve_text_chunking(
chunking: Optional[TextChunkingConfig], rag: LightRAG
) -> tuple[str, dict]:
"""Freeze a ``chunking`` request into ``(process_options, chunk_options)``.
When ``chunking`` is ``None`` this reproduces today's behavior exactly:
fixed-token strategy with the snapshot built from
``rag.addon_params['chunker']``.
Otherwise the validated, strategy-specific params are merged into the
selected strategy's sub-dict. ``chunk_token_size`` rides along inside
``params`` like any other key — every strategy (F included, after the
``process_single_document`` cleanup) reads its size from its own
sub-dict, with the top-level snapshot value as the shared fallback.
Raises:
ValueError: when the request lowers ``chunk_token_size`` below the
*effective* ``chunk_overlap_token_size``. The overlap is often
inherited from ``addon_params``/env (the overlay fills
``fixed_token``/``recursive_character``/``paragraph_semantic``
overlap with ``CHUNK_*_OVERLAP_SIZE`` / ``CHUNK_OVERLAP_SIZE``),
so this can only be checked here against the resolved snapshot,
not in the request model. Callers on the request path invoke
this synchronously so the failure surfaces as HTTP 422 before any
background work is scheduled.
"""
if chunking is None:
# No request-driven config: reproduce today's behavior verbatim,
# including not introducing new validation on the default path.
process_options = PROCESS_OPTION_CHUNK_FIXED
return process_options, resolve_chunk_options(
rag.addon_params, process_options=process_options
)
process_options = _STRATEGY_TO_PROCESS_OPTION[chunking.strategy]
chunk_options = resolve_chunk_options(
rag.addon_params, process_options=process_options
)
strategy_key = chunk_strategy_key(process_options)
chunk_options[strategy_key].update(chunking.params)
_validate_effective_chunk_overlap(chunk_options, strategy_key, chunking.strategy)
_validate_effective_semantic_amount(chunk_options, strategy_key)
return process_options, chunk_options
def _validate_effective_chunk_overlap(
chunk_options: dict, strategy_key: str, strategy_name: str
) -> None:
"""Reject a resolved snapshot whose overlap is >= its chunk size.
Operates on the fully-resolved ``chunk_options`` so it catches the case
the request model cannot: ``chunk_token_size`` supplied in the request
while ``chunk_overlap_token_size`` is inherited from addon_params/env
(e.g. ``chunk_token_size=50`` with the default overlap ``100``). The
effective size is the strategy sub-dict value, falling back to the
top-level snapshot size; the effective overlap is the sub-dict value
(``semantic_vector`` carries none, so it is skipped).
"""
sub = chunk_options.get(strategy_key) or {}
# Fixed-token delimiter-only mode (split_by_character set AND
# split_by_character_only=True) never applies overlap:
# chunking_by_token_size only validates each delimiter segment against
# chunk_token_size and raises on an oversize segment — the overlap field
# is unused. Enforcing overlap < size there would wrongly 422 a valid
# request such as paragraph splitting with a small chunk_token_size.
# (split_by_character_only is itself a no-op when split_by_character is
# falsy, so both must be effective for overlap to be skipped.)
if (
strategy_key == "fixed_token"
and sub.get("split_by_character")
and sub.get("split_by_character_only")
):
return
overlap = sub.get("chunk_overlap_token_size")
if overlap is None:
return
size = sub.get("chunk_token_size")
if size is None:
size = chunk_options.get("chunk_token_size")
if size is not None and overlap >= size:
raise ValueError(
f"chunking for strategy '{strategy_name}': effective "
f"chunk_overlap_token_size ({overlap}) must be < chunk_token_size "
f"({size}). The overlap is inherited from addon_params/env when "
f"not set in the request; raise chunk_token_size or lower "
f"chunk_overlap_token_size."
)
def _validate_effective_semantic_amount(chunk_options: dict, strategy_key: str) -> None:
"""Reject a resolved semantic_vector snapshot whose breakpoint amount
exceeds the percentile/gradient ceiling.
Uses the *effective* ``breakpoint_threshold_type`` from the merged
snapshot — the request model cannot, because the type may be inherited
from ``addon_params``/``CHUNK_V_BREAKPOINT_THRESHOLD_TYPE`` while the
request overrides only ``breakpoint_threshold_amount``. ``percentile`` /
``gradient`` feed ``np.percentile`` (q must be in ``[0, 100]``);
``standard_deviation`` / ``interquartile`` are multipliers with no upper
bound, so a request amount > 100 is valid for them.
"""
if strategy_key != "semantic_vector":
return
sub = chunk_options.get(strategy_key) or {}
amt = sub.get("breakpoint_threshold_amount")
if amt is None:
return
kind = sub.get("breakpoint_threshold_type") or "percentile"
if kind in ("percentile", "gradient") and amt > 100:
raise ValueError(
f"chunking for strategy 'semantic_vector': "
f"breakpoint_threshold_amount ({amt}) must be within (0, 100] for "
f"breakpoint_threshold_type '{kind}'. The type is inherited from "
f"addon_params/env when not set in the request."
)
async def pipeline_index_texts(
rag: LightRAG,
texts: List[str],
file_sources: List[str] = None,
track_id: str = None,
chunking: Optional[TextChunkingConfig] = None,
):
"""Index a list of texts with track_id
Args:
rag: LightRAG instance
texts: The texts to index
file_sources: Sources of the texts
track_id: Optional tracking ID
chunking: Optional chunking strategy + params (already validated by
the request model); when None, default fixed-token chunking is used
"""
if not texts:
return
if not file_sources or len(file_sources) != len(texts):
raise ValueError("A valid file source is required for each text")
normalized_file_sources = [normalize_file_path(source) for source in file_sources]
if any(source == UNKNOWN_FILE_SOURCE for source in normalized_file_sources):
raise ValueError("A valid file source is required for each text")
if len(set(normalized_file_sources)) != len(normalized_file_sources):
raise ValueError("File sources must be unique by filename")
process_options, chunk_options = _resolve_text_chunking(chunking, rag)
await rag.apipeline_enqueue_documents(
input=texts,
file_paths=normalized_file_sources,
track_id=track_id,
process_options=process_options,
chunk_options=chunk_options,
)
await rag.apipeline_process_enqueue_documents()
async def run_scanning_process(
rag: LightRAG, doc_manager: DocumentManager, track_id: str = None
):
"""Background task to scan and index documents
Args:
rag: LightRAG instance
doc_manager: DocumentManager instance
track_id: Optional tracking ID to pass to all scanned files
"""
# The scan endpoint set ``scanning=True`` AND
# ``scanning_exclusive=True`` synchronously before scheduling this
# task. ``scanning`` covers the whole lifecycle (refuses
# overlapping scans); ``scanning_exclusive`` covers only the
# classification phase below — we clear it before invoking
# pipeline_index_files so concurrent uploads can land while the
# scan-driven processing finishes. Both MUST be cleared in
# finally so subsequent uploads / scans can proceed even if the
# body raises. When pipeline_status is not initialised (mocked
# test rigs), the flags were never set so there's nothing to
# clear — track that here to skip the namespace fetch.
from lightrag.exceptions import PipelineNotInitializedError
from lightrag.kg.shared_storage import get_namespace_data, get_namespace_lock
pipeline_status = None
pipeline_status_lock = None
try:
pipeline_status = await get_namespace_data(
"pipeline_status", workspace=rag.workspace
)
pipeline_status_lock = get_namespace_lock(
"pipeline_status", workspace=rag.workspace
)
except PipelineNotInitializedError:
pass
try:
new_files = doc_manager.scan_directory_for_new_files()
total_files = len(new_files)
logger.info(f"Found {total_files} files to index.")
if new_files:
# Group canonical-equivalent files so we can prefer hint-bearing
# variants over plain ones. Within each group sort order is
# preserved as a deterministic tiebreaker.
files_by_canonical_name: dict[str, list[Path]] = {}
for file_path in sorted(
new_files, key=lambda p: get_pinyin_sort_key(str(p))
):
canonical_name = normalize_file_path(str(file_path))
files_by_canonical_name.setdefault(canonical_name, []).append(file_path)
unique_files: list[Path] = []
for canonical_name, group in files_by_canonical_name.items():
# Prefer the first file carrying a supported parser hint so
# the user's explicit engine choice wins over plain variants;
# otherwise fall back to the first sorted entry.
chosen = next(
(f for f in group if filename_parser_hint(f.name) is not None),
group[0],
)
unique_files.append(chosen)
for duplicate in group:
if duplicate is chosen:
continue
warning = (
"Skipping duplicate file in scan batch: "
f"{duplicate.name} duplicates {chosen.name} "
f"(canonical: {canonical_name})"
)
await record_scan_warning(rag, warning)
try:
await move_file_to_parsed_dir(duplicate)
except Exception as move_error:
logger.error(
f"Failed to move duplicate scan file {duplicate.name} to {PARSED_DIR_NAME}: {move_error}"
)
# Partition unique_files into:
# * processed_files — already PROCESSED, archived and skipped.
# * resume_files — same canonical basename matches an existing
# non-PROCESSED doc_status row (PARSING /
# FAILED / PROCESSING / ANALYZING / PENDING).
# These must NOT go through pipeline_enqueue_file
# because apipeline_enqueue_documents would
# treat the same canonical name as a duplicate
# (returning None) and pipeline_enqueue_file
# would then archive the source as if it were
# a duplicate — corrupting pending-parse cases
# that still need the source on disk. The
# pipeline's resume logic, triggered via
# apipeline_process_enqueue_documents, will
# advance them based on their existing
# doc_status row.
# * new_files — no existing record; standard enqueue path.
new_files: list[Path] = []
resume_files: list[Path] = []
processed_files: list[str] = []
for file_path in unique_files:
filename = file_path.name
# Inline the canonical-basename lookup so we keep both the
# doc_id and the data: the FAILED-without-full_docs sub-case
# below needs the doc_id to delete the stale stub.
basename = normalize_file_path(str(file_path))
existing_match = (
await rag.doc_status.get_doc_by_file_basename(basename)
if basename != UNKNOWN_FILE_SOURCE
else None
)
existing_doc_id, existing_doc_data = (
existing_match if existing_match else (None, None)
)
if (
existing_doc_data
and get_doc_status_value(existing_doc_data)
== DocStatus.PROCESSED.value
):
# File is already PROCESSED, skip it with warning and archive it.
processed_files.append(filename)
warning = f"Skipping already processed file: {filename}"
await record_scan_warning(rag, warning)
try:
await move_file_to_parsed_dir(file_path)
except Exception as move_error:
logger.error(
f"Failed to move already processed file {filename} to {PARSED_DIR_NAME}: {move_error}"
)
elif existing_doc_data:
# FAILED rows recorded by apipeline_enqueue_error_documents
# never write a full_docs entry — extraction blew up before
# any content was stored. _validate_and_fix_document_consistency
# preserves them for manual review and removes them from the
# processing list, so the resume path can never advance them.
# When the user fixes the file and re-scans we want a real
# retry: drop the stale stub and treat the file as new so
# the standard enqueue path re-extracts content.
status_value = get_doc_status_value(existing_doc_data)
if status_value == DocStatus.FAILED.value:
full_doc = await rag.full_docs.get_by_id(existing_doc_id)
if full_doc is None:
try:
await rag.doc_status.delete([existing_doc_id])
except Exception as delete_error:
logger.error(
"Failed to delete stale failed-extraction "
f"doc_status stub {existing_doc_id} "
f"({filename}): {delete_error}"
)
# Fall through to resume — at worst the row
# remains preserved (current behaviour) rather
# than re-enqueued.
resume_files.append(file_path)
continue
logger.info(
"Retrying previously failed extraction; "
f"removed stale doc_status stub: {filename} "
f"(doc_id: {existing_doc_id})"
)
new_files.append(file_path)
continue
logger.info(
"Resuming previously unfinished file from scan: "
f"{filename} (Status: {status_value})"
)
resume_files.append(file_path)
else:
new_files.append(file_path)
# Classification phase complete — release ``scanning_exclusive``
# so concurrent uploads/inserts can land in doc_status while
# the scan-driven processing finishes. ``scanning`` stays
# True for the rest of the task lifecycle (releases in
# finally) so the /scan endpoint still refuses overlapping
# scans. Any per-file enqueue or duplicate detected during
# the processing phase is handled by
# apipeline_enqueue_documents' in-batch dedup, identical to
# the upload-during-busy case.
if pipeline_status is not None and pipeline_status_lock is not None:
async with pipeline_status_lock:
pipeline_status["scanning_exclusive"] = False
# New files take the standard enqueue + process path. When at
# least one new file is successfully enqueued, pipeline_index_files
# internally invokes apipeline_process_enqueue_documents, which
# selects work by doc_status state and so will also pick up any
# resume_files in the same run.
if new_files:
await pipeline_index_files(
rag,
new_files,
track_id,
from_scan=True,
)
# Resume targets must always trigger the pipeline explicitly:
# pipeline_index_files only runs apipeline_process_enqueue_documents
# after at least one new file successfully enqueues, so when every
# new file is rejected (unsupported extension, empty body, content
# / filename duplicate, ...) the resume rows would otherwise stay
# stuck until an unrelated indexing run. When new files DID
# enqueue, the inner call already drained the queue and this is a
# cheap no-op that returns "No documents to process".
if resume_files:
await rag.apipeline_process_enqueue_documents()
total_active = len(new_files) + len(resume_files)
if total_active or processed_files:
summary_parts: list[str] = []
if total_active:
summary_parts.append(f"{total_active} files Processed")
if processed_files:
summary_parts.append(f"{len(processed_files)} skipped")
logger.info(f"Scanning process completed: {' '.join(summary_parts)}.")
else:
logger.info(
"No files to process after filtering already processed files."
)
else:
# No new files to index — classification is trivially done;
# release ``scanning_exclusive`` before driving the queue so
# concurrent uploads can land while process_enqueue runs.
if pipeline_status is not None and pipeline_status_lock is not None:
async with pipeline_status_lock:
pipeline_status["scanning_exclusive"] = False
logger.info(
"No upload file found, check if there are any documents in the queue..."
)
await rag.apipeline_process_enqueue_documents()
except Exception as e:
logger.error(f"Error during scanning process: {str(e)}")
logger.error(traceback.format_exc())
finally:
# Always release both scanning flags so future uploads / scans
# are not blocked by a crashed task. Skip when pipeline_status
# was never initialised for this workspace (test rigs).
if pipeline_status is not None and pipeline_status_lock is not None:
async with pipeline_status_lock:
pipeline_status["scanning"] = False
pipeline_status["scanning_exclusive"] = False
async def background_delete_documents(
rag: LightRAG,
doc_manager: DocumentManager,
doc_ids: List[str],
delete_file: bool = False,
delete_llm_cache: bool = False,
):
"""Background task to delete multiple documents"""
from lightrag.kg.shared_storage import (
get_namespace_data,
get_namespace_lock,
)
pipeline_status = await get_namespace_data(
"pipeline_status", workspace=rag.workspace
)
pipeline_status_lock = get_namespace_lock(
"pipeline_status", workspace=rag.workspace
)
total_docs = len(doc_ids)
successful_deletions = []
failed_deletions = []
# The /documents/delete_document endpoint has already reserved the
# destructive slot synchronously: ``busy=True`` and
# ``destructive_busy=True`` were set before the client got
# ``deletion_started``, after checking busy + scanning +
# pending_enqueues>0 atomically. Here we only update the
# job-info fields; the busy reservation was acquired by the
# endpoint and is released in the finally block below.
async with pipeline_status_lock:
pipeline_status.update(
{
# Job name can not be changed, it's verified in adelete_by_doc_id()
"job_name": f"Deleting {total_docs} Documents",
"job_start": datetime.now().isoformat(),
"docs": total_docs,
"batchs": total_docs,
"cur_batch": 0,
"latest_message": "Starting document deletion process",
}
)
# Use slice assignment to clear the list in place
pipeline_status["history_messages"][:] = ["Starting document deletion process"]
if delete_llm_cache:
pipeline_status["history_messages"].append(
"LLM cache cleanup requested for this deletion job"
)
try:
# Loop through each document ID and delete them one by one
for i, doc_id in enumerate(doc_ids, 1):
# Check for cancellation at the start of each document deletion
async with pipeline_status_lock:
if pipeline_status.get("cancellation_requested", False):
cancel_msg = f"Deletion cancelled by user at document {i}/{total_docs}. {len(successful_deletions)} deleted, {total_docs - i + 1} remaining."
logger.info(cancel_msg)
pipeline_status["latest_message"] = cancel_msg
pipeline_status["history_messages"].append(cancel_msg)
# Add remaining documents to failed list with cancellation reason
failed_deletions.extend(
doc_ids[i - 1 :]
) # i-1 because enumerate starts at 1
break # Exit the loop, remaining documents unchanged
start_msg = f"Deleting document {i}/{total_docs}: {doc_id}"
logger.info(start_msg)
pipeline_status["cur_batch"] = i
pipeline_status["latest_message"] = start_msg
pipeline_status["history_messages"].append(start_msg)
file_path = "#"
try:
result = await rag.adelete_by_doc_id(
doc_id, delete_llm_cache=delete_llm_cache
)
file_path = (
getattr(result, "file_path", "-") if "result" in locals() else "-"
)
if result.status == "success":
successful_deletions.append(doc_id)
success_msg = (
f"Document deleted {i}/{total_docs}: {doc_id}[{file_path}]"
)
logger.info(success_msg)
async with pipeline_status_lock:
pipeline_status["history_messages"].append(success_msg)
# Handle file deletion if requested and source information is available
if (
delete_file
and result.file_path
and result.file_path != UNKNOWN_FILE_SOURCE
):
try:
deleted_files, file_delete_errors = (
delete_file_variants_by_file_path(
doc_manager.input_dir,
result.file_path,
)
)
for file_delete_error in file_delete_errors:
logger.warning(file_delete_error)
async with pipeline_status_lock:
pipeline_status["latest_message"] = (
file_delete_error
)
pipeline_status["history_messages"].append(
file_delete_error
)
if deleted_files:
file_delete_msg = (
"Successfully deleted source files: "
+ ", ".join(deleted_files)
)
logger.info(file_delete_msg)
async with pipeline_status_lock:
pipeline_status["latest_message"] = file_delete_msg
pipeline_status["history_messages"].append(
file_delete_msg
)
else:
file_error_msg = (
"File deletion skipped, missing or unsafe file: "
f"{result.file_path}"
)
logger.warning(file_error_msg)
async with pipeline_status_lock:
pipeline_status["latest_message"] = file_error_msg
pipeline_status["history_messages"].append(
file_error_msg
)
except Exception as file_error:
file_error_msg = f"Failed to delete file {result.file_path}: {str(file_error)}"
logger.error(file_error_msg)
async with pipeline_status_lock:
pipeline_status["latest_message"] = file_error_msg
pipeline_status["history_messages"].append(
file_error_msg
)
elif delete_file:
no_file_msg = (
f"File deletion skipped, missing file path: {doc_id}"
)
logger.warning(no_file_msg)
async with pipeline_status_lock:
pipeline_status["latest_message"] = no_file_msg
pipeline_status["history_messages"].append(no_file_msg)
else:
failed_deletions.append(doc_id)
error_msg = f"Failed to delete {i}/{total_docs}: {doc_id}[{file_path}] - {result.message}"
logger.error(error_msg)
async with pipeline_status_lock:
pipeline_status["latest_message"] = error_msg
pipeline_status["history_messages"].append(error_msg)
except Exception as e:
failed_deletions.append(doc_id)
error_msg = f"Error deleting document {i}/{total_docs}: {doc_id}[{file_path}] - {str(e)}"
logger.error(error_msg)
logger.error(traceback.format_exc())
async with pipeline_status_lock:
pipeline_status["latest_message"] = error_msg
pipeline_status["history_messages"].append(error_msg)
except Exception as e:
error_msg = f"Critical error during batch deletion: {str(e)}"
logger.error(error_msg)
logger.error(traceback.format_exc())
async with pipeline_status_lock:
pipeline_status["history_messages"].append(error_msg)
finally:
# Final summary and check for pending requests
async with pipeline_status_lock:
pipeline_status["busy"] = False
pipeline_status["destructive_busy"] = False
pipeline_status["pending_requests"] = False # Reset pending requests flag
pipeline_status["cancellation_requested"] = (
False # Always reset cancellation flag
)
completion_msg = f"Deletion completed: {len(successful_deletions)} successful, {len(failed_deletions)} failed"
pipeline_status["latest_message"] = completion_msg
pipeline_status["history_messages"].append(completion_msg)
# Check if there are pending document indexing requests
has_pending_request = pipeline_status.get("request_pending", False)
# If there are pending requests, start document processing pipeline
if has_pending_request:
try:
logger.info(
"Processing pending document indexing requests after deletion"
)
await rag.apipeline_process_enqueue_documents()
except Exception as e:
logger.error(f"Error processing pending documents after deletion: {e}")
def create_document_routes(
rag: LightRAG, doc_manager: DocumentManager, api_key: Optional[str] = None
):
# Fresh router per call — see the note above the temp_prefix constant.
router = APIRouter(
prefix="/documents",
tags=["documents"],
)
# Create combined auth dependency for document routes
combined_auth = get_combined_auth_dependency(api_key)
@router.post(
"/scan", response_model=ScanResponse, dependencies=[Depends(combined_auth)]
)
async def scan_for_new_documents(background_tasks: BackgroundTasks):
"""
Trigger the scanning process for new documents.
Refuses to start a new scan with
``status='scanning_skipped_pipeline_busy'`` (and does not
schedule a background task) when any of these is set:
- ``pipeline_status["busy"]`` — the processing loop or another
destructive job is running.
- ``pipeline_status["scanning"]`` — another scan is already
running (any phase: classification or processing).
- ``pipeline_status["pending_enqueues"] > 0`` — an /upload,
/text or /texts endpoint has reserved a slot whose bg task
has not yet written to doc_status; starting a scan now would
race scan's classification reads against that pending write.
Both ``scanning`` and ``scanning_exclusive`` are acquired
synchronously here so a subsequent fast-follow request hits the
guard rather than racing against the not-yet-started task.
``run_scanning_process`` clears ``scanning_exclusive`` once
classification is done, allowing concurrent uploads to land
while the scan-driven processing finishes.
Returns:
ScanResponse: A response object containing the scanning status and track_id
"""
from lightrag.exceptions import PipelineNotInitializedError
from lightrag.kg.shared_storage import get_namespace_data, get_namespace_lock
# Generate track_id with "scan" prefix for scanning operation
track_id = generate_track_id("scan")
try:
pipeline_status = await get_namespace_data(
"pipeline_status", workspace=rag.workspace
)
except PipelineNotInitializedError:
# Workspace pipeline_status not yet bootstrapped (e.g. mocked
# test rigs). Treat as idle and allow the scan to proceed; the
# scanning flag has nowhere to live so it is effectively skipped.
background_tasks.add_task(run_scanning_process, rag, doc_manager, track_id)
return ScanResponse(
status="scanning_started",
message="Scanning process has been initiated in the background",
track_id=track_id,
)
pipeline_status_lock = get_namespace_lock(
"pipeline_status", workspace=rag.workspace
)
# Atomically acquire the scanning flag. Scan is the exclusive
# writer in this contract — it reads doc_status to make
# classification decisions (PROCESSED / resume / retry-as-new /
# archive) and would race with concurrent writers — so refuse if:
# * pipeline is processing (busy=True): scan + processing both
# read/mutate doc_status; serialise.
# * another scan is in flight (scanning=True).
# * any /upload, /text, /texts endpoint has reserved a
# pending-enqueue slot (see _reserve_enqueue_slot): the bg
# task has not yet written doc_status and we would otherwise
# race with its mid-flight write.
async with pipeline_status_lock:
if pipeline_status.get("busy"):
logger.warning(
"Scan request skipped: pipeline is busy processing documents"
)
return ScanResponse(
status="scanning_skipped_pipeline_busy",
message=(
"Pipeline is currently busy processing documents. "
"Wait for the running job to finish before triggering another scan."
),
track_id=track_id,
)
if pipeline_status.get("scanning"):
logger.warning(
"Scan request skipped: another scan is already in progress"
)
return ScanResponse(
status="scanning_skipped_pipeline_busy",
message=(
"Another scan is already in progress. "
"Wait for it to finish before triggering a new one."
),
track_id=track_id,
)
pending_enqueues = pipeline_status.get("pending_enqueues", 0)
if pending_enqueues > 0:
logger.warning(
"Scan request skipped: "
f"{pending_enqueues} pending enqueue(s) reserved by "
"upload/insert endpoints"
)
return ScanResponse(
status="scanning_skipped_pipeline_busy",
message=(
"Document upload/insert is being enqueued. "
"Wait for in-flight work to complete before triggering a scan."
),
track_id=track_id,
)
# ``scanning`` covers the whole scan task lifecycle (used by
# this endpoint to refuse overlapping scans).
# ``scanning_exclusive`` is True only during the
# classification phase: run_scanning_process clears it once
# classification is done so concurrent uploads can land
# while the scan-driven processing finishes.
pipeline_status["scanning"] = True
pipeline_status["scanning_exclusive"] = True
# Start the scanning process in the background with track_id. The
# task is responsible for clearing both flags in its finally block.
background_tasks.add_task(run_scanning_process, rag, doc_manager, track_id)
return ScanResponse(
status="scanning_started",
message="Scanning process has been initiated in the background",
track_id=track_id,
)
@router.post(
"/upload", response_model=InsertResponse, dependencies=[Depends(combined_auth)]
)
async def upload_to_input_dir(
background_tasks: BackgroundTasks, file: UploadFile = File(...)
):
"""
Upload a file to the input directory and index it.
This API endpoint accepts a file through an HTTP POST request, checks if the
uploaded file is of a supported type, saves it in the specified input directory,
indexes it for retrieval, and returns a success status with relevant details.
**File Size Limit:**
- Configurable via `MAX_UPLOAD_SIZE` environment variable (default: 100MB)
- Set to `None` or `0` for unlimited upload size
- Returns HTTP 413 (Request Entity Too Large) if file exceeds limit
**Duplicate Detection Behavior:**
This endpoint handles two types of duplicate scenarios differently:
1. **Filename Duplicate (Synchronous Detection)**:
- Detected immediately, before any file is written.
- File name is treated as the unique document key. Both
``doc_status`` and the INPUT directory are checked under the
canonical (parser-hint stripped) basename so ``abc.docx`` and
``abc.[native].docx`` map to the same record.
- **HTTP 409** is returned when a same-name record already exists.
The response detail names the conflict source ("Document
storage already contains ..." or "Input directory already
contains ..."). Clients must delete the existing document
(``DELETE /documents/{doc_id}``) before re-uploading; there is
no longer a 200 ``status="duplicated"`` soft-fail response.
2. **Content Duplicate (Asynchronous Detection)**:
- Detected during background processing after content extraction
- Returns `status="success"` with a new track_id immediately
- The duplicate is detected later when processing the file content
- Use `/documents/track_status/{track_id}` to check the final result:
- Document will have `status="FAILED"`
- `error_msg` contains "Content already exists. Original doc_id: xxx"
- `metadata.is_duplicate=true` with reference to original document
- `metadata.original_doc_id` points to the existing document
- `metadata.original_track_id` shows the original upload's track_id
**Why Different Behavior?**
- Filename check is fast (simple lookup), done synchronously
- Content extraction is expensive (PDF/DOCX parsing), done asynchronously
- This design prevents blocking the client during expensive operations
**Concurrency Constraint:**
- The endpoint refuses with HTTP 409 only while one of the
following exclusive-writer states is set:
``pipeline_status["scanning_exclusive"]`` (a scan is in its
classification phase, reading and possibly mutating doc_status)
or ``pipeline_status["destructive_busy"]`` (``/documents/clear``
or per-doc delete is dropping storages / removing input files).
Wait for the running job to finish before re-submitting.
- ``busy=True`` from the processing loop, and a scan in its
processing phase (``scanning=True`` with
``scanning_exclusive=False``), do NOT block uploads — uploads
are accepted concurrently and the running pipeline picks them
up via its ``request_pending`` mechanism.
Args:
background_tasks: FastAPI BackgroundTasks for async processing
file (UploadFile): The file to be uploaded. It must have an allowed extension.
Returns:
InsertResponse: A response object containing the upload status and a message.
- status="success": File accepted and queued for processing
Raises:
HTTPException: 400 unsupported file type, 409 same-name
conflict or scan-classifying / destructive job in
flight, 413 file too large, 500 other errors.
"""
slot_reserved = False
try:
# Reject upload while a scan is in its CLASSIFICATION
# phase or a destructive job (clear / per-doc delete) is
# in flight, AND reserve a pending-enqueue slot so a scan
# request that arrives before the bg task runs cannot
# transition scanning_exclusive=True under us. Concurrent
# processing (``busy=True``) and a scan in its processing
# phase (``scanning=True`` with
# ``scanning_exclusive=False``) are permitted: the running
# loop's ``request_pending`` mechanism picks up our doc
# after the current batch.
slot_reserved = await _reserve_enqueue_slot(rag)
# Sanitize filename to prevent Path Traversal attacks
safe_filename = sanitize_filename(file.filename, doc_manager.input_dir)
try:
filename_supported = doc_manager.is_supported_file(safe_filename)
except FilenameParserHintError as hint_error:
# Reject malformed hints synchronously with the detailed
# message (previously surfaced asynchronously as an error
# document after the upload was accepted).
raise HTTPException(status_code=400, detail=str(hint_error))
if not filename_supported:
raise HTTPException(
status_code=400,
detail=f"Unsupported file type. Supported types: {doc_manager.supported_extensions}",
)
# Check file size limit (if configured)
if (
global_args.max_upload_size is not None
and global_args.max_upload_size > 0
):
# Safe access to file size (not available in older Starlette versions)
file_size = getattr(file, "size", None)
# Pre-flight size check (only if size is available)
if file_size is not None:
if file_size > global_args.max_upload_size:
raise HTTPException(
status_code=413,
detail=f"File too large. Maximum size: {global_args.max_upload_size / 1024 / 1024:.1f}MB, uploaded: {file_size / 1024 / 1024:.1f}MB",
)
else:
# If size not available, we'll check during streaming
logger.debug(
f"File size not available in UploadFile for {safe_filename}, will check during streaming"
)
file_path = doc_manager.input_dir / safe_filename
# Strict name pre-check. Both the INPUT directory and doc_status
# must be free of any same-canonical-basename record before we
# accept the upload. Replacing an existing document requires an
# explicit DELETE first; we no longer write a "duplicated" 200
# response that silently no-ops.
existing_doc_data = await get_existing_doc_by_file_path_candidates(
rag.doc_status, file_path
)
if existing_doc_data:
status = get_doc_status_value(existing_doc_data) or "unknown"
raise HTTPException(
status_code=409,
detail=(
f"Document storage already contains '{safe_filename}' "
f"(Status: {status}). Delete the existing record before re-uploading."
),
)
# INPUT directory check, using canonical parser-hint names.
# Fast path: exact filename match avoids iterdir on large input directories.
canonical_filename = normalize_file_path(safe_filename)
if file_path.exists():
existing_input_file: Path | None = file_path
else:
existing_input_file = find_existing_file_by_file_path(
doc_manager.input_dir, canonical_filename
)
if existing_input_file:
raise HTTPException(
status_code=409,
detail=(
f"Input directory already contains a file with the same "
f"canonical basename ('{existing_input_file.name}'). "
f"Remove or rename it before re-uploading."
),
)
# Async streaming write with size check
bytes_written = 0
chunk_size = 1024 * 1024 # 1MB chunks
needs_cleanup = False
async with aiofiles.open(file_path, "wb") as out_file:
while True:
# Read chunk from upload stream
chunk = await file.read(chunk_size)
if not chunk:
break
# Check size limit during streaming (if not checked before)
if (
global_args.max_upload_size is not None
and global_args.max_upload_size > 0
):
bytes_written += len(chunk)
if bytes_written > global_args.max_upload_size:
needs_cleanup = True
break
# Write chunk to file
await out_file.write(chunk)
# Cleanup after file is closed
if needs_cleanup:
try:
file_path.unlink()
except Exception as cleanup_error:
logger.error(
f"Error cleaning up oversized file {safe_filename}: {cleanup_error}"
)
raise HTTPException(
status_code=413,
detail=f"File too large. Maximum size: {global_args.max_upload_size / 1024 / 1024:.1f}MB, uploaded: {bytes_written / 1024 / 1024:.1f}MB",
)
track_id = generate_track_id("upload")
# Bg task: enqueue + trigger processing, then release the slot.
# ``pipeline_index_file`` does both: it calls
# ``pipeline_enqueue_file`` (writes doc_status / full_docs) and
# then ``apipeline_process_enqueue_documents``. The latter is
# safe to invoke even when the loop is already busy — it
# collapses to a ``request_pending=True`` nudge and returns,
# so concurrent uploads/inserts cooperate via the running
# loop's request_pending mechanism.
async def _indexing_task():
try:
await pipeline_index_file(rag, file_path, track_id)
finally:
await _release_enqueue_slot(rag)
background_tasks.add_task(_indexing_task)
# Ownership of the slot transferred to the bg task — the
# finally block below must NOT release it again.
slot_reserved = False
return InsertResponse(
status="success",
message=f"File '{safe_filename}' uploaded successfully. Processing will continue in background.",
track_id=track_id,
)
except HTTPException:
# Re-raise HTTP exceptions (400, 413, etc.)
raise
except Exception as e:
logger.error(f"Error /documents/upload: {file.filename}: {str(e)}")
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=str(e))
finally:
# If we reserved a slot but never scheduled the bg task
# (e.g. early validation rejection or streaming-write
# failure), release here. No drain coordination needed —
# any sibling bg task triggers its own processing pass.
if slot_reserved:
await _release_enqueue_slot(rag)
@router.post(
"/text", response_model=InsertResponse, dependencies=[Depends(combined_auth)]
)
async def insert_text(
request: InsertTextRequest, background_tasks: BackgroundTasks
):
"""
Insert text into the RAG system.
This endpoint allows you to insert text data into the RAG system for later retrieval
and use in generating responses.
**Concurrency Constraint:**
- Refuses with HTTP 409 only while
``pipeline_status["scanning_exclusive"]`` (a scan is in its
classification phase) or ``pipeline_status["destructive_busy"]``
(clear / per-doc delete is in flight) is set. ``busy=True``
from the processing loop, and a scan in its processing phase,
do NOT block — the running pipeline picks up the new doc via
``request_pending``.
Args:
request (InsertTextRequest): The request body containing the text to be inserted.
background_tasks: FastAPI BackgroundTasks for async processing
Returns:
InsertResponse: A response object containing the status of the operation.
Raises:
HTTPException: 400 invalid file_source, 409 same-name conflict
or scan/destructive job in flight, 500 other errors.
"""
slot_reserved = False
try:
# Reject text insertion while a scan is in progress AND reserve
# a pending-enqueue slot — see /upload for the rationale.
slot_reserved = await _reserve_enqueue_slot(rag)
# Check if file_source already exists in doc_status storage
if not is_valid_file_source(request.file_source):
raise HTTPException(
status_code=400,
detail="A valid file_source is required for text insertion",
)
normalized_file_source = normalize_file_path(request.file_source)
existing_doc_data = await get_existing_doc_by_file_path_candidates(
rag.doc_status, normalized_file_source
)
if existing_doc_data:
status = get_doc_status_value(existing_doc_data) or "unknown"
raise HTTPException(
status_code=409,
detail=(
f"Document storage already contains '{normalized_file_source}' "
f"(Status: {status}). Delete the existing record before re-inserting."
),
)
# Resolve + validate chunking synchronously so an invalid
# effective config (e.g. chunk_token_size below the inherited
# overlap) fails with HTTP 422 here, before any background work is
# scheduled. pipeline_index_texts re-resolves from the same
# addon_params inside the task.
try:
_resolve_text_chunking(request.chunking, rag)
except ValueError as exc:
raise HTTPException(status_code=422, detail=str(exc))
# Generate track_id for text insertion
track_id = generate_track_id("insert")
async def _indexing_task():
try:
await pipeline_index_texts(
rag,
[request.text],
file_sources=[normalized_file_source],
track_id=track_id,
chunking=request.chunking,
)
finally:
await _release_enqueue_slot(rag)
background_tasks.add_task(_indexing_task)
slot_reserved = False
return InsertResponse(
status="success",
message="Text successfully received. Processing will continue in background.",
track_id=track_id,
)
except HTTPException:
raise
except Exception as e:
logger.error(f"Error /documents/text: {str(e)}")
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=str(e))
finally:
if slot_reserved:
await _release_enqueue_slot(rag)
@router.post(
"/texts",
response_model=InsertResponse,
dependencies=[Depends(combined_auth)],
)
async def insert_texts(
request: InsertTextsRequest, background_tasks: BackgroundTasks
):
"""
Insert multiple texts into the RAG system.
This endpoint allows you to insert multiple text entries into the RAG system
in a single request.
**Concurrency Constraint:**
- Refuses with HTTP 409 only while
``pipeline_status["scanning_exclusive"]`` (a scan is in its
classification phase) or ``pipeline_status["destructive_busy"]``
(clear / per-doc delete is in flight) is set. ``busy=True``
from the processing loop, and a scan in its processing phase,
do NOT block — the running pipeline picks up the new docs via
``request_pending``.
Args:
request (InsertTextsRequest): The request body containing the list of texts.
background_tasks: FastAPI BackgroundTasks for async processing
Returns:
InsertResponse: A response object containing the status of the operation.
Raises:
HTTPException: 400 invalid file_sources, 409 same-name
conflict or scan/destructive job in flight, 500 other
errors.
"""
slot_reserved = False
try:
# Reject batch text insertion while a scan is in progress AND
# reserve a pending-enqueue slot — see /upload for the rationale.
slot_reserved = await _reserve_enqueue_slot(rag)
# Check if any file_sources already exist in doc_status storage
if not request.file_sources or len(request.file_sources) != len(
request.texts
):
raise HTTPException(
status_code=400,
detail="A valid file_source is required for each text",
)
normalized_file_sources = [
normalize_file_path(file_source) for file_source in request.file_sources
]
if any(
file_source == UNKNOWN_FILE_SOURCE
for file_source in normalized_file_sources
):
raise HTTPException(
status_code=400,
detail="A valid file_source is required for each text",
)
if len(set(normalized_file_sources)) != len(normalized_file_sources):
raise HTTPException(
status_code=400,
detail="file_sources must be unique by filename",
)
for file_source in normalized_file_sources:
existing_doc_data = await get_existing_doc_by_file_path_candidates(
rag.doc_status, file_source
)
if existing_doc_data:
status = get_doc_status_value(existing_doc_data) or "unknown"
raise HTTPException(
status_code=409,
detail=(
f"Document storage already contains '{file_source}' "
f"(Status: {status}). Delete the existing record before re-inserting."
),
)
# Resolve + validate the shared chunking synchronously so an
# invalid effective config (e.g. chunk_token_size below the
# inherited overlap) fails with HTTP 422 here, before any
# background work is scheduled. pipeline_index_texts re-resolves
# from the same addon_params inside the task.
try:
_resolve_text_chunking(request.chunking, rag)
except ValueError as exc:
raise HTTPException(status_code=422, detail=str(exc))
# Generate track_id for texts insertion
track_id = generate_track_id("insert")
async def _indexing_task():
try:
await pipeline_index_texts(
rag,
request.texts,
file_sources=normalized_file_sources,
track_id=track_id,
chunking=request.chunking,
)
finally:
await _release_enqueue_slot(rag)
background_tasks.add_task(_indexing_task)
slot_reserved = False
return InsertResponse(
status="success",
message="Texts successfully received. Processing will continue in background.",
track_id=track_id,
)
except HTTPException:
raise
except Exception as e:
logger.error(f"Error /documents/texts: {str(e)}")
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=str(e))
finally:
if slot_reserved:
await _release_enqueue_slot(rag)
@router.delete(
"", response_model=ClearDocumentsResponse, dependencies=[Depends(combined_auth)]
)
async def clear_documents():
"""
Clear all documents from the RAG system.
This endpoint deletes all documents, entities, relationships, and files from the system.
It uses the storage drop methods to properly clean up all data and removes all files
from the input directory.
**Concurrency Constraint:**
- Atomically reserves the destructive slot (sets ``busy=True``
and ``destructive_busy=True``) before dropping anything.
Refuses with ``status="busy"`` when ANY of these is set:
``pipeline_status["busy"]`` (processing loop or another
destructive job in flight), ``pipeline_status["scanning"]``
(a scan is anywhere in its lifecycle), or
``pipeline_status["pending_enqueues"] > 0`` (an /upload,
/text or /texts has reserved a slot whose bg task has not
yet written to doc_status).
Returns:
ClearDocumentsResponse: A response object containing the status and message.
- status="success": All documents and files were successfully cleared.
- status="partial_success": Document clear job exit with some errors.
- status="busy": Operation could not be completed because another
writer (busy / scanning / pending enqueue) holds the pipeline.
- status="fail": All storage drop operations failed, with message
- message: Detailed information about the operation results, including counts
of deleted files and any errors encountered.
Raises:
HTTPException: Raised when a serious error occurs during the clearing process,
with status code 500 and error details in the detail field.
"""
from lightrag.kg.shared_storage import (
get_namespace_data,
get_namespace_lock,
)
# Get pipeline status and lock
pipeline_status = await get_namespace_data(
"pipeline_status", workspace=rag.workspace
)
pipeline_status_lock = get_namespace_lock(
"pipeline_status", workspace=rag.workspace
)
# Atomically reserve the destructive slot. Checks busy +
# scanning + pending_enqueues>0 in a single critical section
# before flipping busy=True and destructive_busy=True together.
# ``destructive_busy`` blocks reservation and the enqueue
# last-line guard: clear is about to drop every storage and
# remove every input file, so a concurrent upload accepted in
# this window would write to storages mid-drop and silently
# lose the document.
acquired, reason = await _acquire_destructive_busy(rag)
if not acquired:
return ClearDocumentsResponse(status="busy", message=reason)
async with pipeline_status_lock:
pipeline_status.update(
{
"job_name": "Clearing Documents",
"job_start": datetime.now().isoformat(),
"docs": 0,
"batchs": 0,
"cur_batch": 0,
"request_pending": False, # Clear any previous request
"latest_message": "Starting document clearing process",
}
)
# Cleaning history_messages without breaking it as a shared list object
del pipeline_status["history_messages"][:]
pipeline_status["history_messages"].append(
"Starting document clearing process"
)
try:
# Use drop method to clear all data
drop_tasks = []
storages = [
rag.text_chunks,
rag.full_docs,
rag.full_entities,
rag.full_relations,
rag.entity_chunks,
rag.relation_chunks,
rag.entities_vdb,
rag.relationships_vdb,
rag.chunks_vdb,
rag.chunk_entity_relation_graph,
rag.doc_status,
]
# Log storage drop start
if "history_messages" in pipeline_status:
pipeline_status["history_messages"].append(
"Starting to drop storage components"
)
for storage in storages:
if storage is not None:
drop_tasks.append(storage.drop())
# Wait for all drop tasks to complete
drop_results = await asyncio.gather(*drop_tasks, return_exceptions=True)
# Check for errors and log results
errors = []
storage_success_count = 0
storage_error_count = 0
for i, result in enumerate(drop_results):
storage_name = storages[i].__class__.__name__
if isinstance(result, Exception):
error_msg = f"Error dropping {storage_name}: {str(result)}"
errors.append(error_msg)
logger.error(error_msg)
storage_error_count += 1
elif isinstance(result, dict) and result.get("status") != "success":
# drop() reports a non-raising failure as {"status": "error"}
# (e.g. a backend that could not safely clear a kept legacy
# store). Honor it so the clear is not counted as successful
# while stale data remains and could be re-migrated/resurface.
error_msg = (
f"Error dropping {storage_name}: "
f"{result.get('message', 'unknown error')}"
)
errors.append(error_msg)
logger.error(error_msg)
storage_error_count += 1
else:
namespace = storages[i].namespace
workspace = storages[i].workspace
logger.info(
f"Successfully dropped {storage_name}: {workspace}/{namespace}"
)
storage_success_count += 1
# Log storage drop results
if "history_messages" in pipeline_status:
if storage_error_count > 0:
pipeline_status["history_messages"].append(
f"Dropped {storage_success_count} storage components with {storage_error_count} errors"
)
else:
pipeline_status["history_messages"].append(
f"Successfully dropped all {storage_success_count} storage components"
)
# If all storage operations failed, return error status and don't proceed with file deletion
if storage_success_count == 0 and storage_error_count > 0:
error_message = "All storage drop operations failed. Aborting document clearing process."
logger.error(error_message)
if "history_messages" in pipeline_status:
pipeline_status["history_messages"].append(error_message)
return ClearDocumentsResponse(status="fail", message=error_message)
# Log file deletion start
if "history_messages" in pipeline_status:
pipeline_status["history_messages"].append(
"Starting to delete files in input directory"
)
# Delete only files in the current directory, preserve files in subdirectories
deleted_files_count = 0
file_errors_count = 0
for file_path in doc_manager.input_dir.glob("*"):
if file_path.is_file():
try:
file_path.unlink()
deleted_files_count += 1
except Exception as e:
logger.error(f"Error deleting file {file_path}: {str(e)}")
file_errors_count += 1
# Log file deletion results
if "history_messages" in pipeline_status:
if file_errors_count > 0:
pipeline_status["history_messages"].append(
f"Deleted {deleted_files_count} files with {file_errors_count} errors"
)
errors.append(f"Failed to delete {file_errors_count} files")
else:
pipeline_status["history_messages"].append(
f"Successfully deleted {deleted_files_count} files"
)
# Prepare final result message
final_message = ""
if errors:
final_message = f"Cleared documents with some errors. Deleted {deleted_files_count} files."
status = "partial_success"
else:
final_message = f"All documents cleared successfully. Deleted {deleted_files_count} files."
status = "success"
# Log final result
if "history_messages" in pipeline_status:
pipeline_status["history_messages"].append(final_message)
# Return response based on results
return ClearDocumentsResponse(status=status, message=final_message)
except Exception as e:
error_msg = f"Error clearing documents: {str(e)}"
logger.error(error_msg)
logger.error(traceback.format_exc())
if "history_messages" in pipeline_status:
pipeline_status["history_messages"].append(error_msg)
raise HTTPException(status_code=500, detail=str(e))
finally:
# Reset busy + destructive_busy after completion so the next
# reservation / scan sees an idle pipeline.
async with pipeline_status_lock:
pipeline_status["busy"] = False
pipeline_status["destructive_busy"] = False
completion_msg = "Document clearing process completed"
pipeline_status["latest_message"] = completion_msg
if "history_messages" in pipeline_status:
pipeline_status["history_messages"].append(completion_msg)
@router.get(
"/pipeline_status",
dependencies=[Depends(combined_auth)],
response_model=PipelineStatusResponse,
)
async def get_pipeline_status() -> PipelineStatusResponse:
"""
Get the current status of the document indexing pipeline.
This endpoint returns information about the current state of the document processing pipeline,
including the processing status, progress information, and history messages.
Returns:
PipelineStatusResponse: A response object containing:
- autoscanned (bool): Whether auto-scan has started
- busy (bool): Whether the pipeline is currently busy
- job_name (str): Current job name (e.g., indexing files/indexing texts)
- job_start (str, optional): Job start time as ISO format string
- docs (int): Total number of documents to be indexed
- batchs (int): Number of batches for processing documents
- cur_batch (int): Current processing batch
- request_pending (bool): Flag for pending request for processing
- latest_message (str): Latest message from pipeline processing
- history_messages (List[str], optional): List of history messages (limited to latest 1000 entries,
with truncation message if more than 1000 messages exist)
Raises:
HTTPException: If an error occurs while retrieving pipeline status (500)
"""
try:
from lightrag.kg.shared_storage import (
get_namespace_data,
get_namespace_lock,
get_all_update_flags_status,
)
pipeline_status = await get_namespace_data(
"pipeline_status", workspace=rag.workspace
)
pipeline_status_lock = get_namespace_lock(
"pipeline_status", workspace=rag.workspace
)
# Get update flags status for all namespaces
update_status = await get_all_update_flags_status(workspace=rag.workspace)
# Convert MutableBoolean objects to regular boolean values
processed_update_status = {}
for namespace, flags in update_status.items():
processed_flags = []
for flag in flags:
# Handle both multiprocess and single process cases
if hasattr(flag, "value"):
processed_flags.append(bool(flag.value))
else:
processed_flags.append(bool(flag))
processed_update_status[namespace] = processed_flags
async with pipeline_status_lock:
# Convert to regular dict if it's a Manager.dict
status_dict = dict(pipeline_status)
# Add processed update_status to the status dictionary
status_dict["update_status"] = processed_update_status
# Convert history_messages to a regular list if it's a Manager.list
# and limit to latest 1000 entries with truncation message if needed
if "history_messages" in status_dict:
history_list = list(status_dict["history_messages"])
total_count = len(history_list)
if total_count > 1000:
# Calculate truncated message count
truncated_count = total_count - 1000
# Take only the latest 1000 messages
latest_messages = history_list[-1000:]
# Add truncation message at the beginning
truncation_message = (
f"[Truncated history messages: {truncated_count}/{total_count}]"
)
status_dict["history_messages"] = [
truncation_message
] + latest_messages
else:
# No truncation needed, return all messages
status_dict["history_messages"] = history_list
# Ensure job_start is properly formatted as a string with timezone information
if "job_start" in status_dict and status_dict["job_start"]:
# Use format_datetime to ensure consistent formatting
status_dict["job_start"] = format_datetime(status_dict["job_start"])
return PipelineStatusResponse(**status_dict)
except Exception as e:
logger.error(f"Error getting pipeline status: {str(e)}")
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=str(e))
# TODO: Deprecated, use /documents/paginated instead
@router.get(
"", response_model=DocsStatusesResponse, dependencies=[Depends(combined_auth)]
)
async def documents() -> DocsStatusesResponse:
"""
Get the status of all documents in the system. This endpoint is deprecated; use /documents/paginated instead.
To prevent excessive resource consumption, a maximum of 1,000 records is returned.
This endpoint retrieves the current status of all documents, grouped by their
processing status (PENDING, PROCESSING, PREPROCESSED, PROCESSED, FAILED). The results are
limited to 1000 total documents with fair distribution across all statuses.
Returns:
DocsStatusesResponse: A response object containing a dictionary where keys are
DocStatus values and values are lists of DocStatusResponse
objects representing documents in each status category.
Maximum 1000 documents total will be returned.
Raises:
HTTPException: If an error occurs while retrieving document statuses (500).
"""
try:
statuses = (
DocStatus.PENDING,
DocStatus.PARSING,
DocStatus.ANALYZING,
DocStatus.PROCESSING,
DocStatus.PREPROCESSED,
DocStatus.PROCESSED,
DocStatus.FAILED,
)
tasks = [rag.get_docs_by_status(status) for status in statuses]
results: List[Dict[str, DocProcessingStatus]] = await asyncio.gather(*tasks)
response = DocsStatusesResponse()
total_documents = 0
max_documents = 1000
# Convert results to lists for easier processing
status_documents = []
for idx, result in enumerate(results):
status = statuses[idx]
docs_list = []
for doc_id, doc_status in result.items():
docs_list.append((doc_id, doc_status))
status_documents.append((status, docs_list))
# Fair distribution: round-robin across statuses
status_indices = [0] * len(
status_documents
) # Track current index for each status
current_status_idx = 0
while total_documents < max_documents:
# Check if we have any documents left to process
has_remaining = False
for status_idx, (status, docs_list) in enumerate(status_documents):
if status_indices[status_idx] < len(docs_list):
has_remaining = True
break
if not has_remaining:
break
# Try to get a document from the current status
status, docs_list = status_documents[current_status_idx]
current_index = status_indices[current_status_idx]
if current_index < len(docs_list):
doc_id, doc_status = docs_list[current_index]
if status not in response.statuses:
response.statuses[status] = []
response.statuses[status].append(
DocStatusResponse(
id=doc_id,
content_summary=doc_status.content_summary,
content_length=doc_status.content_length,
status=doc_status.status,
created_at=format_datetime(doc_status.created_at),
updated_at=format_datetime(doc_status.updated_at),
track_id=doc_status.track_id,
chunks_count=doc_status.chunks_count,
error_msg=doc_status.error_msg,
metadata=doc_status.metadata,
file_path=normalize_file_path(doc_status.file_path),
)
)
status_indices[current_status_idx] += 1
total_documents += 1
# Move to next status (round-robin)
current_status_idx = (current_status_idx + 1) % len(status_documents)
return response
except Exception as e:
logger.error(f"Error GET /documents: {str(e)}")
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=str(e))
class DeleteDocByIdResponse(BaseModel):
"""Response model for single document deletion operation."""
status: Literal["deletion_started", "busy", "not_allowed"] = Field(
description="Status of the deletion operation"
)
message: str = Field(description="Message describing the operation result")
doc_id: str = Field(description="The ID of the document to delete")
@router.delete(
"/delete_document",
response_model=DeleteDocByIdResponse,
dependencies=[Depends(combined_auth)],
summary="Delete a document and all its associated data by its ID.",
)
async def delete_document(
delete_request: DeleteDocRequest,
background_tasks: BackgroundTasks,
) -> DeleteDocByIdResponse:
"""
Delete documents and all their associated data by their IDs using background processing.
Deletes specific documents and all their associated data, including their status,
text chunks, vector embeddings, and any related graph data. When requested,
cached LLM extraction responses are removed after graph deletion/rebuild completes.
The deletion process runs in the background to avoid blocking the client connection.
This operation is irreversible and will interact with the pipeline status.
**Concurrency Constraint:**
- Atomically reserves the destructive slot (sets ``busy=True``
and ``destructive_busy=True``) **synchronously** before
returning ``deletion_started``, so a /scan or /upload that
arrives before the bg task runs cannot race the delete.
Refuses with ``status="busy"`` when ANY of these is set:
``pipeline_status["busy"]``, ``pipeline_status["scanning"]``,
or ``pipeline_status["pending_enqueues"] > 0``.
Args:
delete_request (DeleteDocRequest): The request containing the document IDs and deletion options.
background_tasks: FastAPI BackgroundTasks for async processing
Returns:
DeleteDocByIdResponse: The result of the deletion operation.
- status="deletion_started": The document deletion has been initiated in the background.
- status="busy": Another writer (busy / scanning / pending enqueue) holds the
pipeline; nothing scheduled, retry after the running job finishes.
Raises:
HTTPException:
- 500: If an unexpected internal error occurs during initialization.
"""
doc_ids = delete_request.doc_ids
slot_acquired = False
try:
# Atomically reserve the destructive slot BEFORE returning
# ``deletion_started``. Without this, the bg task would set
# destructive_busy only when it later runs — leaving a
# window where a /scan or /upload can race the delete after
# the client has already received success. The check
# covers busy + scanning + pending_enqueues>0 in a single
# critical section.
acquired, reason = await _acquire_destructive_busy(rag)
if not acquired:
return DeleteDocByIdResponse(
status="busy",
message=reason or "Cannot delete documents while pipeline is busy",
doc_id=", ".join(doc_ids),
)
slot_acquired = True
background_tasks.add_task(
background_delete_documents,
rag,
doc_manager,
doc_ids,
delete_request.delete_file,
delete_request.delete_llm_cache,
)
# Ownership of the slot transferred to the bg task — it
# will release in its finally. The endpoint's finally
# below must NOT release it again.
slot_acquired = False
return DeleteDocByIdResponse(
status="deletion_started",
message=f"Document deletion for {len(doc_ids)} documents has been initiated. Processing will continue in background.",
doc_id=", ".join(doc_ids),
)
except Exception as e:
error_msg = f"Error initiating document deletion for {delete_request.doc_ids}: {str(e)}"
logger.error(error_msg)
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=error_msg)
finally:
# If we reserved but never scheduled the bg task (e.g. an
# unexpected error between acquire and add_task), release
# so the next reservation / scan / enqueue can proceed.
if slot_acquired:
await _release_destructive_busy(rag)
@router.post(
"/clear_cache",
response_model=ClearCacheResponse,
dependencies=[Depends(combined_auth)],
)
async def clear_cache(request: ClearCacheRequest):
"""
Clear all cache data from the LLM response cache storage.
This endpoint clears all cached LLM responses regardless of mode.
The request body is accepted for API compatibility but is ignored.
Args:
request (ClearCacheRequest): The request body (ignored for compatibility).
Returns:
ClearCacheResponse: A response object containing the status and message.
Raises:
HTTPException: If an error occurs during cache clearing (500).
"""
try:
# Call the aclear_cache method (no modes parameter)
await rag.aclear_cache()
# Prepare success message
message = "Successfully cleared all cache"
return ClearCacheResponse(status="success", message=message)
except Exception as e:
logger.error(f"Error clearing cache: {str(e)}")
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=str(e))
@router.get(
"/track_status/{track_id}",
response_model=TrackStatusResponse,
dependencies=[Depends(combined_auth)],
)
async def get_track_status(track_id: str) -> TrackStatusResponse:
"""
Get the processing status of documents by tracking ID.
This endpoint retrieves all documents associated with a specific tracking ID,
allowing users to monitor the processing progress of their uploaded files or inserted texts.
Args:
track_id (str): The tracking ID returned from upload, text, or texts endpoints
Returns:
TrackStatusResponse: A response object containing:
- track_id: The tracking ID
- documents: List of documents associated with this track_id
- total_count: Total number of documents for this track_id
Raises:
HTTPException: If track_id is invalid (400) or an error occurs (500).
"""
try:
# Validate track_id
if not track_id or not track_id.strip():
raise HTTPException(status_code=400, detail="Track ID cannot be empty")
track_id = track_id.strip()
# Get documents by track_id
docs_by_track_id = await rag.aget_docs_by_track_id(track_id)
# Convert to response format
documents = []
status_summary = {}
for doc_id, doc_status in docs_by_track_id.items():
documents.append(
DocStatusResponse(
id=doc_id,
content_summary=doc_status.content_summary,
content_length=doc_status.content_length,
status=doc_status.status,
created_at=format_datetime(doc_status.created_at),
updated_at=format_datetime(doc_status.updated_at),
track_id=doc_status.track_id,
chunks_count=doc_status.chunks_count,
error_msg=doc_status.error_msg,
metadata=doc_status.metadata,
file_path=normalize_file_path(doc_status.file_path),
)
)
# Build status summary
# Handle both DocStatus enum and string cases for robust deserialization
status_key = str(doc_status.status)
status_summary[status_key] = status_summary.get(status_key, 0) + 1
return TrackStatusResponse(
track_id=track_id,
documents=documents,
total_count=len(documents),
status_summary=status_summary,
)
except HTTPException:
raise
except Exception as e:
logger.error(f"Error getting track status for {track_id}: {str(e)}")
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=str(e))
@router.post(
"/paginated",
response_model=PaginatedDocsResponse,
dependencies=[Depends(combined_auth)],
)
async def get_documents_paginated(
request: DocumentsRequest,
) -> PaginatedDocsResponse:
"""
Get documents with pagination support.
This endpoint retrieves documents with pagination, filtering, and sorting capabilities.
It provides better performance for large document collections by loading only the
requested page of data.
Args:
request (DocumentsRequest): The request body containing pagination parameters
Returns:
PaginatedDocsResponse: A response object containing:
- documents: List of documents for the current page
- pagination: Pagination information (page, total_count, etc.)
- status_counts: Count of documents by status for all documents
Raises:
HTTPException: If an error occurs while retrieving documents (500).
"""
trace_id = uuid4().hex[:8]
request_start = time.perf_counter()
status_filter_value = (
request.status_filter.value if request.status_filter is not None else None
)
workspace = getattr(rag, "workspace", None)
performance_timing_log(
"[documents/paginated][%s] Request start workspace=%s status_filter=%s page=%s page_size=%s sort_field=%s sort_direction=%s",
trace_id,
workspace,
status_filter_value,
request.page,
request.page_size,
request.sort_field,
request.sort_direction,
)
try:
async def _timed_call(operation_name: str, operation):
operation_start = time.perf_counter()
performance_timing_log(
"[documents/paginated][%s] %s started",
trace_id,
operation_name,
)
try:
result = await operation
except Exception:
elapsed = time.perf_counter() - operation_start
performance_timing_log(
"[documents/paginated][%s] %s failed after %.4fs",
trace_id,
operation_name,
elapsed,
)
raise
elapsed = time.perf_counter() - operation_start
performance_timing_log(
"[documents/paginated][%s] %s completed in %.4fs",
trace_id,
operation_name,
elapsed,
)
return result
query_task_create_start = time.perf_counter()
docs_task = asyncio.create_task(
_timed_call(
"get_docs_paginated",
rag.doc_status.get_docs_paginated(
status_filter=request.status_filter,
status_filters=request.status_filters,
page=request.page,
page_size=request.page_size,
sort_field=request.sort_field,
sort_direction=request.sort_direction,
),
)
)
status_counts_task = asyncio.create_task(
_timed_call(
"get_all_status_counts",
rag.doc_status.get_all_status_counts(),
)
)
query_task_create_elapsed = time.perf_counter() - query_task_create_start
performance_timing_log(
"[documents/paginated][%s] Query tasks created in %.4fs",
trace_id,
query_task_create_elapsed,
)
query_await_start = time.perf_counter()
(documents_with_ids, total_count), status_counts = await asyncio.gather(
docs_task, status_counts_task
)
query_await_elapsed = time.perf_counter() - query_await_start
performance_timing_log(
"[documents/paginated][%s] Query tasks awaited in %.4fs",
trace_id,
query_await_elapsed,
)
# Convert documents to response format
response_assembly_start = time.perf_counter()
doc_responses = []
for doc_id, doc in documents_with_ids:
doc_responses.append(
DocStatusResponse(
id=doc_id,
content_summary=doc.content_summary,
content_length=doc.content_length,
status=doc.status,
created_at=format_datetime(doc.created_at),
updated_at=format_datetime(doc.updated_at),
track_id=doc.track_id,
chunks_count=doc.chunks_count,
error_msg=doc.error_msg,
metadata=doc.metadata,
file_path=normalize_file_path(doc.file_path),
)
)
# Calculate pagination info
total_pages = (total_count + request.page_size - 1) // request.page_size
has_next = request.page < total_pages
has_prev = request.page > 1
pagination = PaginationInfo(
page=request.page,
page_size=request.page_size,
total_count=total_count,
total_pages=total_pages,
has_next=has_next,
has_prev=has_prev,
)
response = PaginatedDocsResponse(
documents=doc_responses,
pagination=pagination,
status_counts=status_counts,
)
response_assembly_elapsed = time.perf_counter() - response_assembly_start
total_elapsed = time.perf_counter() - request_start
performance_timing_log(
"[documents/paginated][%s] Response assembled in %.4fs",
trace_id,
response_assembly_elapsed,
)
performance_timing_log(
"[documents/paginated][%s] Request completed in %.4fs returned_rows=%s total_count=%s status_count_keys=%s",
trace_id,
total_elapsed,
len(doc_responses),
total_count,
sorted(status_counts.keys()),
)
return response
except Exception as e:
total_elapsed = time.perf_counter() - request_start
performance_timing_log(
"[documents/paginated][%s] Request failed after %.4fs",
trace_id,
total_elapsed,
)
logger.error(f"Error getting paginated documents: {str(e)}")
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=str(e))
@router.get(
"/status_counts",
response_model=StatusCountsResponse,
dependencies=[Depends(combined_auth)],
)
async def get_document_status_counts() -> StatusCountsResponse:
"""
Get counts of documents by status.
This endpoint retrieves the count of documents in each processing status
(PENDING, PROCESSING, PROCESSED, FAILED) for all documents in the system.
Returns:
StatusCountsResponse: A response object containing status counts
Raises:
HTTPException: If an error occurs while retrieving status counts (500).
"""
try:
status_counts = await rag.doc_status.get_all_status_counts()
return StatusCountsResponse(status_counts=status_counts)
except Exception as e:
logger.error(f"Error getting document status counts: {str(e)}")
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=str(e))
@router.post(
"/reprocess_failed",
response_model=ReprocessResponse,
dependencies=[Depends(combined_auth)],
)
async def reprocess_failed_documents(background_tasks: BackgroundTasks):
"""
Reprocess failed and pending documents.
This endpoint triggers the document processing pipeline which automatically
picks up and reprocesses documents in the following statuses:
- FAILED: Documents that failed during previous processing attempts
- PENDING: Documents waiting to be processed
- PROCESSING: Documents with abnormally terminated processing (e.g., server crashes)
This is useful for recovering from server crashes, network errors, LLM service
outages, or other temporary failures that caused document processing to fail.
The processing happens in the background and can be monitored by checking the
pipeline status. The reprocessed documents retain their original track_id from
initial upload, so use their original track_id to monitor progress.
Returns:
ReprocessResponse: Response with status and message.
track_id is always empty string because reprocessed documents retain
their original track_id from initial upload.
Raises:
HTTPException: If an error occurs while initiating reprocessing (500).
"""
try:
# Start the reprocessing in the background
# Note: Reprocessed documents retain their original track_id from initial upload
background_tasks.add_task(rag.apipeline_process_enqueue_documents)
logger.info("Reprocessing of failed documents initiated")
return ReprocessResponse(
status="reprocessing_started",
message="Reprocessing of failed documents has been initiated in background. Documents retain their original track_id.",
)
except Exception as e:
logger.error(f"Error initiating reprocessing of failed documents: {str(e)}")
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=str(e))
@router.post(
"/cancel_pipeline",
response_model=CancelPipelineResponse,
dependencies=[Depends(combined_auth)],
)
async def cancel_pipeline():
"""
Request cancellation of the currently running pipeline.
This endpoint sets a cancellation flag in the pipeline status. The pipeline will:
1. Check this flag at key processing points
2. Stop processing new documents
3. Cancel all running document processing tasks
4. Mark all PROCESSING documents as FAILED with reason "User cancelled"
The cancellation is graceful and ensures data consistency. Documents that have
completed processing will remain in PROCESSED status.
Returns:
CancelPipelineResponse: Response with status and message
- status="cancellation_requested": Cancellation flag has been set
- status="not_busy": Pipeline is not currently running
Raises:
HTTPException: If an error occurs while setting cancellation flag (500).
"""
try:
from lightrag.kg.shared_storage import (
get_namespace_data,
get_namespace_lock,
)
pipeline_status = await get_namespace_data(
"pipeline_status", workspace=rag.workspace
)
pipeline_status_lock = get_namespace_lock(
"pipeline_status", workspace=rag.workspace
)
async with pipeline_status_lock:
if not pipeline_status.get("busy", False):
return CancelPipelineResponse(
status="not_busy",
message="Pipeline is not currently running. No cancellation needed.",
)
# Set cancellation flag
pipeline_status["cancellation_requested"] = True
cancel_msg = "Pipeline cancellation requested by user"
logger.info(cancel_msg)
pipeline_status["latest_message"] = cancel_msg
pipeline_status["history_messages"].append(cancel_msg)
return CancelPipelineResponse(
status="cancellation_requested",
message="Pipeline cancellation has been requested. Documents will be marked as FAILED.",
)
except Exception as e:
logger.error(f"Error requesting pipeline cancellation: {str(e)}")
logger.error(traceback.format_exc())
raise HTTPException(status_code=500, detail=str(e))
return router