Skip to content

armgit / dify

Public Imported from GitHub
0

Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.

17.8 KB
Raw Blame
import base64
import hashlib
import os
import posixpath
import uuid
from collections.abc import Generator, Sequence  # Changed Iterator to Generator
from contextlib import contextmanager, suppress
from dataclasses import dataclass
from datetime import datetime
from tempfile import NamedTemporaryFile
from typing import Literal
from zipfile import ZIP_DEFLATED, ZipFile

from sqlalchemy import Engine, select
from sqlalchemy.orm import Session, sessionmaker
from werkzeug.exceptions import NotFound

from configs import dify_config
from constants import (
    AUDIO_EXTENSIONS,
    DOCUMENT_EXTENSIONS,
    IMAGE_EXTENSIONS,
    VIDEO_EXTENSIONS,
)
from core.rag.extractor.extract_processor import ExtractProcessor
from enums import DeploymentEdition
from extensions.ext_storage import storage
from extensions.storage.storage_type import StorageType
from graphon.file import helpers as file_helpers
from libs.datetime_utils import naive_utc_now
from libs.helper import extract_tenant_id
from models import Account
from models.enums import CreatorUserRole
from models.model import EndUser, UploadFile

from .errors.file import BlockedFileExtensionError, FileNotExistsError, FileTooLargeError, UnsupportedFileTypeError

PREVIEW_WORDS_LIMIT = 3000


@dataclass(frozen=True, slots=True)
class FileUploadActor:
    """Creator identity, independent of the tenant receiving the upload."""

    id: str
    creator_role: CreatorUserRole


@dataclass(frozen=True, slots=True)
class FileUploadResult:
    id: str
    name: str
    size: int
    extension: str
    mime_type: str
    created_by: str
    created_at: datetime
    tenant_id: str
    source_url: str


class FileService:
    _session_maker: sessionmaker[Session]

    def __init__(self, session_factory: sessionmaker | Engine | None = None):
        match session_factory:
            case Engine():
                self._session_maker = sessionmaker(bind=session_factory)
            case sessionmaker():
                self._session_maker = session_factory
            case _:
                raise AssertionError("must be a sessionmaker or an Engine.")

    def upload_file(
        self,
        *,
        filename: str,
        content: bytes,
        mimetype: str,
        user: Account | EndUser | FileUploadActor,
        tenant_id: str | None = None,
        source: Literal["datasets", "knowledge_fs"] | None = None,
        source_url: str = "",
        default_file_size_limit: int | None = None,
    ) -> UploadFile:
        """Persist a file with the given creator and resource tenant.

        For Account or EndUser, tenant_id=None (including when omitted) uses
        the account's current workspace or the end user's tenant. An explicit
        tenant_id takes precedence. FileUploadActor carries no tenant, so it
        requires an explicit tenant_id.

        default_file_size_limit is the non-media limit in MiB; None uses
        dify_config.UPLOAD_FILE_SIZE_LIMIT, and 0 permits only empty files.
        Images, video and audio always use their dedicated configured limits.
        """
        if isinstance(user, FileUploadActor) and tenant_id is None:
            raise TypeError("tenant_id is required when uploading with FileUploadActor")

        # get file extension
        extension = os.path.splitext(filename)[1].lstrip(".").lower()

        # Only reject path separators here. The original filename is stored as metadata,
        # while the storage key is UUID-based.
        if any(c in filename for c in ["/", "\\"]):
            raise ValueError("Filename contains invalid characters")

        if len(filename) > 200:
            filename = filename.split(".")[0][:200] + "." + extension

        # check if extension is in blacklist
        if extension and extension in dify_config.UPLOAD_FILE_EXTENSION_BLACKLIST:
            raise BlockedFileExtensionError(f"File extension '.{extension}' is not allowed for security reasons")

        if source == "datasets" and extension not in DOCUMENT_EXTENSIONS:
            raise UnsupportedFileTypeError()

        # get file size
        file_size = len(content)

        # check if the file size is exceeded
        if not FileService.is_file_size_within_limit(
            extension=extension,
            file_size=file_size,
            default_file_size_limit=default_file_size_limit,
        ):
            raise FileTooLargeError

        # generate file key
        file_uuid = str(uuid.uuid4())

        if isinstance(user, FileUploadActor):
            resource_tenant_id = tenant_id
            creator_role = user.creator_role
        else:
            resource_tenant_id = tenant_id if tenant_id is not None else extract_tenant_id(user)
            creator_role = CreatorUserRole.ACCOUNT if isinstance(user, Account) else CreatorUserRole.END_USER

        file_key = "upload_files/" + (resource_tenant_id or "") + "/" + file_uuid + "." + extension

        # save file to storage
        storage.save(file_key, content)

        # save file to db
        upload_file = UploadFile(
            tenant_id=resource_tenant_id or "",
            storage_type=StorageType(dify_config.STORAGE_TYPE),
            key=file_key,
            name=filename,
            size=file_size,
            extension=extension,
            mime_type=mimetype,
            created_by_role=creator_role,
            created_by=user.id,
            created_at=naive_utc_now(),
            used=False,
            hash=hashlib.sha3_256(content).hexdigest(),
            source_url=source_url,
        )

        with self._session_maker(expire_on_commit=False) as session:
            session.add(upload_file)
            session.commit()

        if not upload_file.source_url:
            upload_file.source_url = file_helpers.get_signed_file_url(upload_file_id=upload_file.id)

        return upload_file

    def upload_file_for_actor(
        self,
        *,
        actor: FileUploadActor,
        resource_tenant_id: str,
        filename: str,
        content: bytes,
        mimetype: str,
        source: Literal["datasets"] | None = None,
        source_url: str = "",
        default_file_size_limit: int | None = None,
    ) -> FileUploadResult:
        """Upload into the explicitly admitted tenant and return detached values.

        default_file_size_limit overrides the non-media limit in MiB, including
        0 for empty files only. None uses dify_config.UPLOAD_FILE_SIZE_LIMIT.
        Images, video and audio retain their dedicated configured limits.
        """
        upload_file = self.upload_file(
            filename=filename,
            content=content,
            mimetype=mimetype,
            user=actor,
            tenant_id=resource_tenant_id,
            source=source,
            source_url=source_url,
            default_file_size_limit=default_file_size_limit,
        )
        return FileUploadResult(
            id=upload_file.id,
            name=upload_file.name,
            size=upload_file.size,
            extension=upload_file.extension,
            mime_type=upload_file.mime_type,
            created_by=upload_file.created_by,
            created_at=upload_file.created_at,
            tenant_id=upload_file.tenant_id,
            source_url=upload_file.source_url,
        )

    @staticmethod
    def is_file_size_within_limit(
        *,
        extension: str,
        file_size: int,
        default_file_size_limit: int | None = None,
    ) -> bool:
        return file_size <= FileService.file_size_limit(
            extension=extension,
            default_file_size_limit=default_file_size_limit,
        )

    @staticmethod
    def file_size_limit(
        *,
        extension: str,
        default_file_size_limit: int | None = None,
    ) -> int:
        """Return the size an extension is allowed, in bytes.

        default_file_size_limit is a non-media override in MiB. None uses
        dify_config.UPLOAD_FILE_SIZE_LIMIT; 0 allows only empty files.
        Images, video and audio use their dedicated limits regardless of this
        override.
        """

        if extension in IMAGE_EXTENSIONS:
            file_size_limit = dify_config.UPLOAD_IMAGE_FILE_SIZE_LIMIT
        elif extension in VIDEO_EXTENSIONS:
            file_size_limit = dify_config.UPLOAD_VIDEO_FILE_SIZE_LIMIT
        elif extension in AUDIO_EXTENSIONS:
            file_size_limit = dify_config.UPLOAD_AUDIO_FILE_SIZE_LIMIT
        else:
            # Context-specific uploads may override the default limit without changing media-specific limits.
            file_size_limit = (
                default_file_size_limit if default_file_size_limit is not None else dify_config.UPLOAD_FILE_SIZE_LIMIT
            )

        return file_size_limit * 1024 * 1024

    def get_file_base64(self, file_id: str) -> str:
        with self._session_maker(expire_on_commit=False) as session:
            upload_file = session.scalar(select(UploadFile).where(UploadFile.id == file_id).limit(1))
            if not upload_file:
                raise NotFound("File not found")
            upload_file_key = upload_file.key

        blob = storage.load_once(upload_file_key)
        return base64.b64encode(blob).decode()

    def get_file_presigned_url(self, *, file_id: str, tenant_id: str) -> str:
        """Generate a direct storage URL for a tenant-owned upload file."""
        with self._session_maker(expire_on_commit=False) as session:
            upload_file = self.get_upload_file_by_id(tenant_id, file_id, session=session)
            if upload_file is None:
                raise NotFound("File not found")

            file_key = upload_file.key
            content_type = upload_file.mime_type

        return storage.generate_presigned_url(
            file_key,
            expires_in=dify_config.FILES_ACCESS_TIMEOUT,
            content_type=content_type,
        )

    def get_icon_url(self, file_id: str, tenant_id: str) -> str:
        try:
            if dify_config.DEPLOYMENT_EDITION == DeploymentEdition.CLOUD and (
                StorageType(dify_config.STORAGE_TYPE) == StorageType.S3
            ):
                return self.get_file_presigned_url(file_id=file_id, tenant_id=tenant_id)
            with self._session_maker(expire_on_commit=False) as session:
                upload_file = self.get_upload_file_by_id(tenant_id, file_id, session=session)
            if upload_file is None:
                raise NotFound("File not found")
        except NotFound as exc:
            raise FileNotExistsError("File reference not found") from exc
        return file_helpers.get_signed_file_url(upload_file_id=file_id)

    def upload_text(self, text: str, text_name: str, user_id: str, tenant_id: str) -> UploadFile:
        if len(text_name) > 200:
            text_name = text_name[:200]
        # user uuid as file name
        file_uuid = str(uuid.uuid4())
        file_key = "upload_files/" + tenant_id + "/" + file_uuid + ".txt"
        content = text.encode("utf-8")

        # save file to storage
        storage.save(file_key, content)

        # save file to db
        upload_file = UploadFile(
            tenant_id=tenant_id,
            storage_type=StorageType(dify_config.STORAGE_TYPE),
            key=file_key,
            name=text_name,
            size=len(content),
            extension="txt",
            mime_type="text/plain",
            created_by=user_id,
            created_by_role=CreatorUserRole.ACCOUNT,
            created_at=naive_utc_now(),
            used=True,
            used_by=user_id,
            used_at=naive_utc_now(),
        )

        with self._session_maker(expire_on_commit=False) as session:
            session.add(upload_file)
            session.commit()

        return upload_file

    def get_file_preview(self, file_id: str, tenant_id: str) -> str:
        """
        Return a short text preview extracted from a document file.
        """
        with self._session_maker(expire_on_commit=False) as session:
            upload_file = session.scalar(
                select(UploadFile).where(UploadFile.id == file_id, UploadFile.tenant_id == tenant_id).limit(1)
            )

        if not upload_file:
            raise NotFound("File not found")

        # extract text from file
        extension = upload_file.extension
        if extension.lower() not in DOCUMENT_EXTENSIONS:
            raise UnsupportedFileTypeError()

        text = ExtractProcessor.load_from_upload_file(upload_file, return_text=True)
        return text[0:PREVIEW_WORDS_LIMIT] if text else ""

    def get_file_content(self, file_id: str) -> str:
        with self._session_maker(expire_on_commit=False) as session:
            upload_file: UploadFile | None = session.scalar(select(UploadFile).where(UploadFile.id == file_id).limit(1))

        if not upload_file:
            raise NotFound("File not found")
        content = storage.load(upload_file.key)

        return content.decode("utf-8")

    def delete_file(self, file_id: str):
        with self._session_maker() as session, session.begin():
            upload_file = session.scalar(select(UploadFile).where(UploadFile.id == file_id))

            if not upload_file:
                return
            storage.delete(upload_file.key)
            session.delete(upload_file)

    @staticmethod
    def get_upload_file_by_id(tenant_id: str, upload_file_id: str, *, session: Session) -> UploadFile | None:
        return session.scalar(
            select(UploadFile)
            .where(
                UploadFile.tenant_id == tenant_id,
                UploadFile.id == upload_file_id,
            )
            .limit(1)
        )

    @staticmethod
    def get_upload_files_by_ids(
        tenant_id: str, upload_file_ids: Sequence[str], *, session: Session
    ) -> dict[str, UploadFile]:
        """
        Fetch `UploadFile` rows for a tenant in a single batch query.

        This is a generic `UploadFile` lookup helper (not dataset/document specific), so it lives in `FileService`.
        """
        if not upload_file_ids:
            return {}

        # Normalize and deduplicate ids before using them in the IN clause.
        upload_file_id_list: list[str] = [str(upload_file_id) for upload_file_id in upload_file_ids]
        unique_upload_file_ids: list[str] = list(set(upload_file_id_list))

        # Fetch upload files in one query for efficient batch access.
        upload_files: Sequence[UploadFile] = session.scalars(
            select(UploadFile).where(
                UploadFile.tenant_id == tenant_id,
                UploadFile.id.in_(unique_upload_file_ids),
            )
        ).all()
        return {str(upload_file.id): upload_file for upload_file in upload_files}

    @staticmethod
    def _sanitize_zip_entry_name(name: str) -> str:
        """
        Sanitize a ZIP entry name to avoid path traversal and weird separators.

        We keep this conservative: the upload flow already rejects `/` and `\\`, but older rows (or imported data)
        could still contain unsafe names.
        """
        # Drop any directory components and prevent empty names. ZIP names are always POSIX-style, so split on
        # `/` explicitly rather than through `os.path`, whose separators differ on Windows hosts.
        base = posixpath.basename(name).strip() or "file"

        # ZIP uses forward slashes as separators; remove any residual separator characters.
        return base.replace("/", "_").replace("\\", "_")

    @staticmethod
    def _dedupe_zip_entry_name(original_name: str, used_names: set[str]) -> str:
        """
        Return a unique ZIP entry name, inserting suffixes before the extension.
        """
        # Keep the original name when it's not already used.
        if original_name not in used_names:
            return original_name

        # Insert suffixes before the extension (e.g., "doc.txt" -> "doc (1).txt").
        stem, extension = os.path.splitext(original_name)
        suffix = 1
        while True:
            candidate = f"{stem} ({suffix}){extension}"
            if candidate not in used_names:
                return candidate
            suffix += 1

    @staticmethod
    @contextmanager
    def build_upload_files_zip_tempfile(
        *,
        upload_files: Sequence[UploadFile],
    ) -> Generator[str]:  # Changed from Iterator[str]
        """
        Build a ZIP from `UploadFile`s and yield a tempfile path.

        We yield a path (rather than an open file handle) to avoid "read of closed file" issues when Flask/Werkzeug
        streams responses. The caller is expected to keep this context open until the response is fully sent, then
        close it (e.g., via `response.call_on_close(...)`) to delete the tempfile.
        """
        used_names: set[str] = set()

        # Build a ZIP in a temp file and keep it on disk until the caller finishes streaming it.
        tmp_path: str | None = None
        try:
            with NamedTemporaryFile(mode="w+b", suffix=".zip", delete=False) as tmp:
                tmp_path = tmp.name
                with ZipFile(tmp, mode="w", compression=ZIP_DEFLATED) as zf:
                    for upload_file in upload_files:
                        # Ensure the entry name is safe and unique.
                        safe_name = FileService._sanitize_zip_entry_name(upload_file.name)
                        arcname = FileService._dedupe_zip_entry_name(safe_name, used_names)
                        used_names.add(arcname)

                        # Stream file bytes from storage into the ZIP entry.
                        with zf.open(arcname, "w") as entry:
                            for chunk in storage.load(upload_file.key, stream=True):
                                entry.write(chunk)

                # Flush so `send_file(path, ...)` can re-open it safely on all platforms.
                tmp.flush()

            assert tmp_path is not None
            yield tmp_path
        finally:
            # Remove the temp file when the context is closed (typically after the response finishes streaming).
            if tmp_path is not None:
                with suppress(FileNotFoundError):
                    os.remove(tmp_path)