Aluode/PerceptionLabPortable

1from __future__ import annotations2 3import importlib4import types5import warnings6 7__all__ = ["registry", "get_filesystem_class", "default"]8 9# internal, mutable10_registry: dict[str, type] = {}11 12# external, immutable13registry = types.MappingProxyType(_registry)14default = "file"15 16 17def register_implementation(name, cls, clobber=False, errtxt=None):18    """Add implementation class to the registry19 20    Parameters21    ----------22    name: str23        Protocol name to associate with the class24    cls: class or str25        if a class: fsspec-compliant implementation class (normally inherits from26        ``fsspec.AbstractFileSystem``, gets added straight to the registry. If a27        str, the full path to an implementation class like package.module.class,28        which gets added to known_implementations,29        so the import is deferred until the filesystem is actually used.30    clobber: bool (optional)31        Whether to overwrite a protocol with the same name; if False, will raise32        instead.33    errtxt: str (optional)34        If given, then a failure to import the given class will result in this35        text being given.36    """37    if isinstance(cls, str):38        if name in known_implementations and clobber is False:39            if cls != known_implementations[name]["class"]:40                raise ValueError(41                    f"Name ({name}) already in the known_implementations and clobber "42                    f"is False"43                )44        else:45            known_implementations[name] = {46                "class": cls,47                "err": errtxt or f"{cls} import failed for protocol {name}",48            }49 50    else:51        if name in registry and clobber is False:52            if _registry[name] is not cls:53                raise ValueError(54                    f"Name ({name}) already in the registry and clobber is False"55                )56        else:57            _registry[name] = cls58 59 60# protocols mapped to the class which implements them. This dict can be61# updated with register_implementation62known_implementations = {63    "abfs": {64        "class": "adlfs.AzureBlobFileSystem",65        "err": "Install adlfs to access Azure Datalake Gen2 and Azure Blob Storage",66    },67    "adl": {68        "class": "adlfs.AzureDatalakeFileSystem",69        "err": "Install adlfs to access Azure Datalake Gen1",70    },71    "arrow_hdfs": {72        "class": "fsspec.implementations.arrow.HadoopFileSystem",73        "err": "pyarrow and local java libraries required for HDFS",74    },75    "async_wrapper": {76        "class": "fsspec.implementations.asyn_wrapper.AsyncFileSystemWrapper",77    },78    "asynclocal": {79        "class": "morefs.asyn_local.AsyncLocalFileSystem",80        "err": "Install 'morefs[asynclocalfs]' to use AsyncLocalFileSystem",81    },82    "asyncwrapper": {83        "class": "fsspec.implementations.asyn_wrapper.AsyncFileSystemWrapper",84    },85    "az": {86        "class": "adlfs.AzureBlobFileSystem",87        "err": "Install adlfs to access Azure Datalake Gen2 and Azure Blob Storage",88    },89    "blockcache": {"class": "fsspec.implementations.cached.CachingFileSystem"},90    "box": {91        "class": "boxfs.BoxFileSystem",92        "err": "Please install boxfs to access BoxFileSystem",93    },94    "cached": {"class": "fsspec.implementations.cached.CachingFileSystem"},95    "dask": {96        "class": "fsspec.implementations.dask.DaskWorkerFileSystem",97        "err": "Install dask distributed to access worker file system",98    },99    "data": {"class": "fsspec.implementations.data.DataFileSystem"},100    "dbfs": {101        "class": "fsspec.implementations.dbfs.DatabricksFileSystem",102        "err": "Install the requests package to use the DatabricksFileSystem",103    },104    "dir": {"class": "fsspec.implementations.dirfs.DirFileSystem"},105    "dropbox": {106        "class": "dropboxdrivefs.DropboxDriveFileSystem",107        "err": (108            'DropboxFileSystem requires "dropboxdrivefs","requests" and "'109            '"dropbox" to be installed'110        ),111    },112    "dvc": {113        "class": "dvc.api.DVCFileSystem",114        "err": "Install dvc to access DVCFileSystem",115    },116    "file": {"class": "fsspec.implementations.local.LocalFileSystem"},117    "filecache": {"class": "fsspec.implementations.cached.WholeFileCacheFileSystem"},118    "ftp": {"class": "fsspec.implementations.ftp.FTPFileSystem"},119    "gcs": {120        "class": "gcsfs.GCSFileSystem",121        "err": "Please install gcsfs to access Google Storage",122    },123    "gdrive": {124        "class": "gdrive_fsspec.GoogleDriveFileSystem",125        "err": "Please install gdrive_fs for access to Google Drive",126    },127    "generic": {"class": "fsspec.generic.GenericFileSystem"},128    "gist": {129        "class": "fsspec.implementations.gist.GistFileSystem",130        "err": "Install the requests package to use the gist FS",131    },132    "git": {133        "class": "fsspec.implementations.git.GitFileSystem",134        "err": "Install pygit2 to browse local git repos",135    },136    "github": {137        "class": "fsspec.implementations.github.GithubFileSystem",138        "err": "Install the requests package to use the github FS",139    },140    "gs": {141        "class": "gcsfs.GCSFileSystem",142        "err": "Please install gcsfs to access Google Storage",143    },144    "hdfs": {145        "class": "fsspec.implementations.arrow.HadoopFileSystem",146        "err": "pyarrow and local java libraries required for HDFS",147    },148    "hf": {149        "class": "huggingface_hub.HfFileSystem",150        "err": "Install huggingface_hub to access HfFileSystem",151    },152    "http": {153        "class": "fsspec.implementations.http.HTTPFileSystem",154        "err": 'HTTPFileSystem requires "requests" and "aiohttp" to be installed',155    },156    "https": {157        "class": "fsspec.implementations.http.HTTPFileSystem",158        "err": 'HTTPFileSystem requires "requests" and "aiohttp" to be installed',159    },160    "jlab": {161        "class": "fsspec.implementations.jupyter.JupyterFileSystem",162        "err": "Jupyter FS requires requests to be installed",163    },164    "jupyter": {165        "class": "fsspec.implementations.jupyter.JupyterFileSystem",166        "err": "Jupyter FS requires requests to be installed",167    },168    "lakefs": {169        "class": "lakefs_spec.LakeFSFileSystem",170        "err": "Please install lakefs-spec to access LakeFSFileSystem",171    },172    "libarchive": {173        "class": "fsspec.implementations.libarchive.LibArchiveFileSystem",174        "err": "LibArchive requires to be installed",175    },176    "local": {"class": "fsspec.implementations.local.LocalFileSystem"},177    "memory": {"class": "fsspec.implementations.memory.MemoryFileSystem"},178    "oci": {179        "class": "ocifs.OCIFileSystem",180        "err": "Install ocifs to access OCI Object Storage",181    },182    "ocilake": {183        "class": "ocifs.OCIFileSystem",184        "err": "Install ocifs to access OCI Data Lake",185    },186    "oss": {187        "class": "ossfs.OSSFileSystem",188        "err": "Install ossfs to access Alibaba Object Storage System",189    },190    "pyscript": {191        "class": "pyscript_fsspec_client.client.PyscriptFileSystem",192        "err": "Install requests (cpython) or run in pyscript",193    },194    "reference": {"class": "fsspec.implementations.reference.ReferenceFileSystem"},195    "root": {196        "class": "fsspec_xrootd.XRootDFileSystem",197        "err": (198            "Install fsspec-xrootd to access xrootd storage system. "199            "Note: 'root' is the protocol name for xrootd storage systems, "200            "not referring to root directories"201        ),202    },203    "s3": {"class": "s3fs.S3FileSystem", "err": "Install s3fs to access S3"},204    "s3a": {"class": "s3fs.S3FileSystem", "err": "Install s3fs to access S3"},205    "sftp": {206        "class": "fsspec.implementations.sftp.SFTPFileSystem",207        "err": 'SFTPFileSystem requires "paramiko" to be installed',208    },209    "simplecache": {"class": "fsspec.implementations.cached.SimpleCacheFileSystem"},210    "smb": {211        "class": "fsspec.implementations.smb.SMBFileSystem",212        "err": 'SMB requires "smbprotocol" or "smbprotocol[kerberos]" installed',213    },214    "ssh": {215        "class": "fsspec.implementations.sftp.SFTPFileSystem",216        "err": 'SFTPFileSystem requires "paramiko" to be installed',217    },218    "tar": {"class": "fsspec.implementations.tar.TarFileSystem"},219    "tos": {220        "class": "tosfs.TosFileSystem",221        "err": "Install tosfs to access ByteDance volcano engine Tinder Object Storage",222    },223    "tosfs": {224        "class": "tosfs.TosFileSystem",225        "err": "Install tosfs to access ByteDance volcano engine Tinder Object Storage",226    },227    "wandb": {"class": "wandbfs.WandbFS", "err": "Install wandbfs to access wandb"},228    "webdav": {229        "class": "webdav4.fsspec.WebdavFileSystem",230        "err": "Install webdav4 to access WebDAV",231    },232    "webhdfs": {233        "class": "fsspec.implementations.webhdfs.WebHDFS",234        "err": 'webHDFS access requires "requests" to be installed',235    },236    "zip": {"class": "fsspec.implementations.zip.ZipFileSystem"},237}238 239assert list(known_implementations) == sorted(known_implementations), (240    "Not in alphabetical order"241)242 243 244def get_filesystem_class(protocol):245    """Fetch named protocol implementation from the registry246 247    The dict ``known_implementations`` maps protocol names to the locations248    of classes implementing the corresponding file-system. When used for the249    first time, appropriate imports will happen and the class will be placed in250    the registry. All subsequent calls will fetch directly from the registry.251 252    Some protocol implementations require additional dependencies, and so the253    import may fail. In this case, the string in the "err" field of the254    ``known_implementations`` will be given as the error message.255    """256    if not protocol:257        protocol = default258 259    if protocol not in registry:260        if protocol not in known_implementations:261            raise ValueError(f"Protocol not known: {protocol}")262        bit = known_implementations[protocol]263        try:264            register_implementation(protocol, _import_class(bit["class"]))265        except ImportError as e:266            raise ImportError(bit.get("err")) from e267    cls = registry[protocol]268    if getattr(cls, "protocol", None) in ("abstract", None):269        cls.protocol = protocol270 271    return cls272 273 274s3_msg = """Your installed version of s3fs is very old and known to cause275severe performance issues, see also https://github.com/dask/dask/issues/10276276 277To fix, you should specify a lower version bound on s3fs, or278update the current installation.279"""280 281 282def _import_class(fqp: str):283    """Take a fully-qualified path and return the imported class or identifier.284 285    ``fqp`` is of the form "package.module.klass" or286    "package.module:subobject.klass".287 288    Warnings289    --------290    This can import arbitrary modules. Make sure you haven't installed any modules291    that may execute malicious code at import time.292    """293    if ":" in fqp:294        mod, name = fqp.rsplit(":", 1)295    else:296        mod, name = fqp.rsplit(".", 1)297 298    is_s3 = mod == "s3fs"299    mod = importlib.import_module(mod)300    if is_s3 and mod.__version__.split(".") < ["0", "5"]:301        warnings.warn(s3_msg)302    for part in name.split("."):303        mod = getattr(mod, part)304 305    if not isinstance(mod, type):306        raise TypeError(f"{fqp} is not a class")307 308    return mod309 310 311def filesystem(protocol, **storage_options):312    """Instantiate filesystems for given protocol and arguments313 314    ``storage_options`` are specific to the protocol being chosen, and are315    passed directly to the class.316    """317    if protocol == "arrow_hdfs":318        warnings.warn(319            "The 'arrow_hdfs' protocol has been deprecated and will be "320            "removed in the future. Specify it as 'hdfs'.",321            DeprecationWarning,322        )323 324    cls = get_filesystem_class(protocol)325    return cls(**storage_options)326 327 328def available_protocols():329    """Return a list of the implemented protocols.330 331    Note that any given protocol may require extra packages to be importable.332    """333    return list(known_implementations)334