fsspec 2025.3.2__py3-none-any.whl → 2025.5.1__py3-none-any.whl

fsspec/__init__.py

@@ -1,5 +1,3 @@
-from importlib.metadata import entry_points
-
 from . import caching
 from ._version import __version__  # noqa: F401
 from .callbacks import Callback
@@ -38,6 +36,10 @@ __all__ = [
 
 
 def process_entries():
+    try:
+        from importlib.metadata import entry_points
+    except ImportError:
+        return
     if entry_points is not None:
         try:
             eps = entry_points()

fsspec/_version.py

@@ -17,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '2025.3.2'
-__version_tuple__ = version_tuple = (2025, 3, 2)
+__version__ = version = '2025.5.1'
+__version_tuple__ = version_tuple = (2025, 5, 1)

fsspec/generic.py

@@ -16,15 +16,13 @@ logger = logging.getLogger("fsspec.generic")
 
 
 def set_generic_fs(protocol, **storage_options):
+    """Populate the dict used for method=="generic" lookups"""
     _generic_fs[protocol] = filesystem(protocol, **storage_options)
 
 
-default_method = "default"
-
-
-def _resolve_fs(url, method=None, protocol=None, storage_options=None):
+def _resolve_fs(url, method, protocol=None, storage_options=None):
     """Pick instance of backend FS"""
-    method = method or default_method
+    url = url[0] if isinstance(url, (list, tuple)) else url
     protocol = protocol or split_protocol(url)[0]
     storage_options = storage_options or {}
     if method == "default":
@@ -159,7 +157,7 @@ class GenericFileSystem(AsyncFileSystem):
 
     protocol = "generic"  # there is no real reason to ever use a protocol with this FS
 
-    def __init__(self, default_method="default", **kwargs):
+    def __init__(self, default_method="default", storage_options=None, **kwargs):
         """
 
         Parameters
@@ -171,22 +169,25 @@ class GenericFileSystem(AsyncFileSystem):
               configured via the config system
             - "generic": takes instances from the `_generic_fs` dict in this module,
               which you must populate before use. Keys are by protocol
+            - "options": expects storage_options, a dict mapping protocol to
+              kwargs to use when constructing the filesystem
             - "current": takes the most recently instantiated version of each FS
         """
         self.method = default_method
+        self.st_opts = storage_options
         super().__init__(**kwargs)
 
     def _parent(self, path):
-        fs = _resolve_fs(path, self.method)
+        fs = _resolve_fs(path, self.method, storage_options=self.st_opts)
         return fs.unstrip_protocol(fs._parent(path))
 
     def _strip_protocol(self, path):
         # normalization only
-        fs = _resolve_fs(path, self.method)
+        fs = _resolve_fs(path, self.method, storage_options=self.st_opts)
         return fs.unstrip_protocol(fs._strip_protocol(path))
 
     async def _find(self, path, maxdepth=None, withdirs=False, detail=False, **kwargs):
-        fs = _resolve_fs(path, self.method)
+        fs = _resolve_fs(path, self.method, storage_options=self.st_opts)
         if fs.async_impl:
             out = await fs._find(
                 path, maxdepth=maxdepth, withdirs=withdirs, detail=True, **kwargs
@@ -251,7 +252,7 @@ class GenericFileSystem(AsyncFileSystem):
         value,
         **kwargs,
     ):
-        fs = _resolve_fs(path, self.method)
+        fs = _resolve_fs(path, self.method, storage_options=self.st_opts)
         if fs.async_impl:
             return await fs._pipe_file(path, value, **kwargs)
         else:
@@ -269,7 +270,7 @@ class GenericFileSystem(AsyncFileSystem):
 
     async def _makedirs(self, path, exist_ok=False):
         logger.debug("Make dir %s", path)
-        fs = _resolve_fs(path, self.method)
+        fs = _resolve_fs(path, self.method, storage_options=self.st_opts)
         if fs.async_impl:
             await fs._makedirs(path, exist_ok=exist_ok)
         else:
@@ -288,6 +289,7 @@ class GenericFileSystem(AsyncFileSystem):
         url2,
         blocksize=2**20,
         callback=DEFAULT_CALLBACK,
+        tempdir: Optional[str] = None,
         **kwargs,
     ):
         fs = _resolve_fs(url, self.method)
@@ -295,35 +297,10 @@ class GenericFileSystem(AsyncFileSystem):
         if fs is fs2:
             # pure remote
             if fs.async_impl:
-                return await fs._cp_file(url, url2, **kwargs)
+                return await fs._copy(url, url2, **kwargs)
             else:
-                return fs.cp_file(url, url2, **kwargs)
-        kw = {"blocksize": 0, "cache_type": "none"}
-        try:
-            f1 = (
-                await fs.open_async(url, "rb")
-                if hasattr(fs, "open_async")
-                else fs.open(url, "rb", **kw)
-            )
-            callback.set_size(await maybe_await(f1.size))
-            f2 = (
-                await fs2.open_async(url2, "wb")
-                if hasattr(fs2, "open_async")
-                else fs2.open(url2, "wb", **kw)
-            )
-            while f1.size is None or f2.tell() < f1.size:
-                data = await maybe_await(f1.read(blocksize))
-                if f1.size is None and not data:
-                    break
-                await maybe_await(f2.write(data))
-                callback.absolute_update(f2.tell())
-        finally:
-            try:
-                await maybe_await(f2.close())
-                await maybe_await(f1.close())
-            except NameError:
-                # fail while opening f1 or f2
-                pass
+                return fs.copy(url, url2, **kwargs)
+        await copy_file_op(fs, [url], fs2, [url2], tempdir, 1, on_error="raise")
 
     async def _make_many_dirs(self, urls, exist_ok=True):
         fs = _resolve_fs(urls[0], self.method)
@@ -347,17 +324,22 @@ class GenericFileSystem(AsyncFileSystem):
         tempdir: Optional[str] = None,
         **kwargs,
     ):
+        # TODO: special case for one FS being local, which can use get/put
+        # TODO: special case for one being memFS, which can use cat/pipe
         if recursive:
-            raise NotImplementedError
-        fs = _resolve_fs(path1[0], self.method)
-        fs2 = _resolve_fs(path2[0], self.method)
-        # not expanding paths atm., assume call is from rsync()
+            raise NotImplementedError("Please use fsspec.generic.rsync")
+        path1 = [path1] if isinstance(path1, str) else path1
+        path2 = [path2] if isinstance(path2, str) else path2
+
+        fs = _resolve_fs(path1, self.method)
+        fs2 = _resolve_fs(path2, self.method)
+
         if fs is fs2:
-            # pure remote
             if fs.async_impl:
                 return await fs._copy(path1, path2, **kwargs)
             else:
                 return fs.copy(path1, path2, **kwargs)
+
         await copy_file_op(
             fs, path1, fs2, path2, tempdir, batch_size, on_error=on_error
         )
@@ -377,31 +359,33 @@ async def copy_file_op(
                 fs2,
                 u2,
                 os.path.join(tempdir, uuid.uuid4().hex),
-                on_error=on_error,
             )
             for u1, u2 in zip(url1, url2)
         ]
-        await _run_coros_in_chunks(coros, batch_size=batch_size)
+        out = await _run_coros_in_chunks(
+            coros, batch_size=batch_size, return_exceptions=True
+        )
     finally:
         shutil.rmtree(tempdir)
+    if on_error == "return":
+        return out
+    elif on_error == "raise":
+        for o in out:
+            if isinstance(o, Exception):
+                raise o
 
 
 async def _copy_file_op(fs1, url1, fs2, url2, local, on_error="ignore"):
-    ex = () if on_error == "raise" else Exception
-    logger.debug("Copy %s -> %s", url1, url2)
-    try:
-        if fs1.async_impl:
-            await fs1._get_file(url1, local)
-        else:
-            fs1.get_file(url1, local)
-        if fs2.async_impl:
-            await fs2._put_file(local, url2)
-        else:
-            fs2.put_file(local, url2)
-        os.unlink(local)
-        logger.debug("Copy %s -> %s; done", url1, url2)
-    except ex as e:
-        logger.debug("ignoring cp exception for %s: %s", url1, e)
+    if fs1.async_impl:
+        await fs1._get_file(url1, local)
+    else:
+        fs1.get_file(url1, local)
+    if fs2.async_impl:
+        await fs2._put_file(local, url2)
+    else:
+        fs2.put_file(local, url2)
+    os.unlink(local)
+    logger.debug("Copy %s -> %s; done", url1, url2)
 
 
 async def maybe_await(cor):

fsspec/implementations/asyn_wrapper.py

@@ -2,6 +2,7 @@ import asyncio
 import functools
 import inspect
 
+import fsspec
 from fsspec.asyn import AsyncFileSystem, running_async
 
 
@@ -42,14 +43,24 @@ class AsyncFileSystemWrapper(AsyncFileSystem):
         The synchronous filesystem instance to wrap.
     """
 
-    protocol = "async_wrapper"
+    protocol = "asyncwrapper", "async_wrapper"
     cachable = False
 
-    def __init__(self, fs, *args, asynchronous=None, **kwargs):
+    def __init__(
+        self,
+        fs=None,
+        asynchronous=None,
+        target_protocol=None,
+        target_options=None,
+        **kwargs,
+    ):
         if asynchronous is None:
             asynchronous = running_async()
-        super().__init__(*args, asynchronous=asynchronous, **kwargs)
-        self.sync_fs = fs
+        super().__init__(asynchronous=asynchronous, **kwargs)
+        if fs is not None:
+            self.sync_fs = fs
+        else:
+            self.sync_fs = fsspec.filesystem(target_protocol, **target_options)
         self.protocol = self.sync_fs.protocol
         self._wrap_all_sync_methods()
 

fsspec/implementations/cached.py

@@ -335,6 +335,11 @@ class CachingFileSystem(AbstractFileSystem):
             self._metadata.update_file(path, detail)
             logger.debug("Creating local sparse file for %s", path)
 
+        # explicitly submitting the size to the open call will avoid extra
+        # operations when opening. This is particularly relevant
+        # for any file that is read over a network, e.g. S3.
+        size = detail.get("size", None)
+
         # call target filesystems open
         self._mkcache()
         f = self.fs._open(
@@ -344,8 +349,15 @@ class CachingFileSystem(AbstractFileSystem):
             autocommit=autocommit,
             cache_options=cache_options,
             cache_type="none",
+            size=size,
             **kwargs,
         )
+
+        # set size if not already set
+        if size is None:
+            detail["size"] = f.size
+            self._metadata.update_file(path, detail)
+
         if self.compression:
             comp = (
                 infer_compression(path)

fsspec/implementations/dbfs.py

@@ -14,12 +14,13 @@ class DatabricksException(Exception):
     Helper class for exceptions raised in this module.
     """
 
-    def __init__(self, error_code, message):
+    def __init__(self, error_code, message, details=None):
         """Create a new DatabricksException"""
         super().__init__(message)
 
         self.error_code = error_code
         self.message = message
+        self.details = details
 
 
 class DatabricksFileSystem(AbstractFileSystem):
@@ -80,7 +81,7 @@ class DatabricksFileSystem(AbstractFileSystem):
                     raise FileNotFoundError(e.message) from e
 
                 raise
-            files = r["files"]
+            files = r.get("files", [])
             out = [
                 {
                     "name": o["path"],

fsspec/implementations/gist.py

@@ -0,0 +1,232 @@
+import requests
+
+from ..spec import AbstractFileSystem
+from ..utils import infer_storage_options
+from .memory import MemoryFile
+
+
+class GistFileSystem(AbstractFileSystem):
+    """
+    Interface to files in a single GitHub Gist.
+
+    Provides read-only access to a gist's files. Gists do not contain
+    subdirectories, so file listing is straightforward.
+
+    Parameters
+    ----------
+    gist_id : str
+        The ID of the gist you want to access (the long hex value from the URL).
+    filenames : list[str] (optional)
+        If provided, only make a file system representing these files, and do not fetch
+        the list of all files for this gist.
+    sha : str (optional)
+        If provided, fetch a particular revision of the gist. If omitted,
+        the latest revision is used.
+    username : str (optional)
+        GitHub username for authentication (required if token is given).
+    token : str (optional)
+        GitHub personal access token (required if username is given).
+    timeout : (float, float) or float, optional
+        Connect and read timeouts for requests (default 60s each).
+    kwargs : dict
+        Stored on `self.request_kw` and passed to `requests.get` when fetching Gist
+        metadata or reading ("opening") a file.
+    """
+
+    protocol = "gist"
+    gist_url = "https://api.github.com/gists/{gist_id}"
+    gist_rev_url = "https://api.github.com/gists/{gist_id}/{sha}"
+
+    def __init__(
+        self,
+        gist_id,
+        filenames=None,
+        sha=None,
+        username=None,
+        token=None,
+        timeout=None,
+        **kwargs,
+    ):
+        super().__init__()
+        self.gist_id = gist_id
+        self.filenames = filenames
+        self.sha = sha  # revision of the gist (optional)
+        if (username is None) ^ (token is None):
+            # Both or neither must be set
+            if username or token:
+                raise ValueError("Auth requires both username and token, or neither.")
+        self.username = username
+        self.token = token
+        self.request_kw = kwargs
+        # Default timeouts to 60s connect/read if none provided
+        self.timeout = timeout if timeout is not None else (60, 60)
+
+        # We use a single-level "directory" cache, because a gist is essentially flat
+        self.dircache[""] = self._fetch_file_list()
+
+    @property
+    def kw(self):
+        """Auth parameters passed to 'requests' if we have username/token."""
+        if self.username is not None and self.token is not None:
+            return {"auth": (self.username, self.token), **self.request_kw}
+        return self.request_kw
+
+    def _fetch_gist_metadata(self):
+        """
+        Fetch the JSON metadata for this gist (possibly for a specific revision).
+        """
+        if self.sha:
+            url = self.gist_rev_url.format(gist_id=self.gist_id, sha=self.sha)
+        else:
+            url = self.gist_url.format(gist_id=self.gist_id)
+
+        r = requests.get(url, timeout=self.timeout, **self.kw)
+        if r.status_code == 404:
+            raise FileNotFoundError(
+                f"Gist not found: {self.gist_id}@{self.sha or 'latest'}"
+            )
+        r.raise_for_status()
+        return r.json()
+
+    def _fetch_file_list(self):
+        """
+        Returns a list of dicts describing each file in the gist. These get stored
+        in self.dircache[""].
+        """
+        meta = self._fetch_gist_metadata()
+        if self.filenames:
+            available_files = meta.get("files", {})
+            files = {}
+            for fn in self.filenames:
+                if fn not in available_files:
+                    raise FileNotFoundError(fn)
+                files[fn] = available_files[fn]
+        else:
+            files = meta.get("files", {})
+
+        out = []
+        for fname, finfo in files.items():
+            if finfo is None:
+                # Occasionally GitHub returns a file entry with null if it was deleted
+                continue
+            # Build a directory entry
+            out.append(
+                {
+                    "name": fname,  # file's name
+                    "type": "file",  # gists have no subdirectories
+                    "size": finfo.get("size", 0),  # file size in bytes
+                    "raw_url": finfo.get("raw_url"),
+                }
+            )
+        return out
+
+    @classmethod
+    def _strip_protocol(cls, path):
+        """
+        Remove 'gist://' from the path, if present.
+        """
+        # The default infer_storage_options can handle gist://username:token@id/file
+        # or gist://id/file, but let's ensure we handle a normal usage too.
+        # We'll just strip the protocol prefix if it exists.
+        path = infer_storage_options(path).get("path", path)
+        return path.lstrip("/")
+
+    @staticmethod
+    def _get_kwargs_from_urls(path):
+        """
+        Parse 'gist://' style URLs into GistFileSystem constructor kwargs.
+        For example:
+          gist://:TOKEN@<gist_id>/file.txt
+          gist://username:TOKEN@<gist_id>/file.txt
+        """
+        so = infer_storage_options(path)
+        out = {}
+        if "username" in so and so["username"]:
+            out["username"] = so["username"]
+        if "password" in so and so["password"]:
+            out["token"] = so["password"]
+        if "host" in so and so["host"]:
+            # We interpret 'host' as the gist ID
+            out["gist_id"] = so["host"]
+
+        # Extract SHA and filename from path
+        if "path" in so and so["path"]:
+            path_parts = so["path"].rsplit("/", 2)[-2:]
+            if len(path_parts) == 2:
+                if path_parts[0]:  # SHA present
+                    out["sha"] = path_parts[0]
+                if path_parts[1]:  # filename also present
+                    out["filenames"] = [path_parts[1]]
+
+        return out
+
+    def ls(self, path="", detail=False, **kwargs):
+        """
+        List files in the gist. Gists are single-level, so any 'path' is basically
+        the filename, or empty for all files.
+
+        Parameters
+        ----------
+        path : str, optional
+            The filename to list. If empty, returns all files in the gist.
+        detail : bool, default False
+            If True, return a list of dicts; if False, return a list of filenames.
+        """
+        path = self._strip_protocol(path or "")
+        # If path is empty, return all
+        if path == "":
+            results = self.dircache[""]
+        else:
+            # We want just the single file with this name
+            all_files = self.dircache[""]
+            results = [f for f in all_files if f["name"] == path]
+            if not results:
+                raise FileNotFoundError(path)
+        if detail:
+            return results
+        else:
+            return sorted(f["name"] for f in results)
+
+    def _open(self, path, mode="rb", block_size=None, **kwargs):
+        """
+        Read a single file from the gist.
+        """
+        if mode != "rb":
+            raise NotImplementedError("GitHub Gist FS is read-only (no write).")
+
+        path = self._strip_protocol(path)
+        # Find the file entry in our dircache
+        matches = [f for f in self.dircache[""] if f["name"] == path]
+        if not matches:
+            raise FileNotFoundError(path)
+        finfo = matches[0]
+
+        raw_url = finfo.get("raw_url")
+        if not raw_url:
+            raise FileNotFoundError(f"No raw_url for file: {path}")
+
+        r = requests.get(raw_url, timeout=self.timeout, **self.kw)
+        if r.status_code == 404:
+            raise FileNotFoundError(path)
+        r.raise_for_status()
+        return MemoryFile(path, None, r.content)
+
+    def cat(self, path, recursive=False, on_error="raise", **kwargs):
+        """
+        Return {path: contents} for the given file or files. If 'recursive' is True,
+        and path is empty, returns all files in the gist.
+        """
+        paths = self.expand_path(path, recursive=recursive)
+        out = {}
+        for p in paths:
+            try:
+                with self.open(p, "rb") as f:
+                    out[p] = f.read()
+            except FileNotFoundError as e:
+                if on_error == "raise":
+                    raise e
+                elif on_error == "omit":
+                    pass  # skip
+                else:
+                    out[p] = e
+        return out

fsspec/implementations/github.py

@@ -1,4 +1,5 @@
 import base64
+import re
 
 import requests
 
@@ -6,8 +7,6 @@ from ..spec import AbstractFileSystem
 from ..utils import infer_storage_options
 from .memory import MemoryFile
 
-# TODO: add GIST backend, would be very similar
-
 
 class GithubFileSystem(AbstractFileSystem):
     """Interface to files in github
@@ -265,3 +264,70 @@ class GithubFileSystem(AbstractFileSystem):
             cache_options=cache_options,
             **kwargs,
         )
+
+    def rm(self, path, recursive=False, maxdepth=None, message=None):
+        path = self.expand_path(path, recursive=recursive, maxdepth=maxdepth)
+        for p in reversed(path):
+            self.rm_file(p, message=message)
+
+    def rm_file(self, path, message=None, **kwargs):
+        """
+        Remove a file from a specified branch using a given commit message.
+
+        Since Github DELETE operation requires a branch name, and we can't reliably
+        determine whether the provided SHA refers to a branch, tag, or commit, we
+        assume it's a branch. If it's not, the user will encounter an error when
+        attempting to retrieve the file SHA or delete the file.
+
+        Parameters
+        ----------
+        path: str
+            The file's location relative to the repository root.
+        message: str, optional
+            The commit message for the deletion.
+        """
+
+        if not self.username:
+            raise ValueError("Authentication required")
+
+        path = self._strip_protocol(path)
+
+        # Attempt to get SHA from cache or Github API
+        sha = self._get_sha_from_cache(path)
+        if not sha:
+            url = self.content_url.format(
+                org=self.org, repo=self.repo, path=path.lstrip("/"), sha=self.root
+            )
+            r = requests.get(url, timeout=self.timeout, **self.kw)
+            if r.status_code == 404:
+                raise FileNotFoundError(path)
+            r.raise_for_status()
+            sha = r.json()["sha"]
+
+        # Delete the file
+        delete_url = self.content_url.format(
+            org=self.org, repo=self.repo, path=path, sha=self.root
+        )
+        branch = self.root
+        data = {
+            "message": message or f"Delete {path}",
+            "sha": sha,
+            **({"branch": branch} if branch else {}),
+        }
+
+        r = requests.delete(delete_url, json=data, timeout=self.timeout, **self.kw)
+        error_message = r.json().get("message", "")
+        if re.search(r"Branch .+ not found", error_message):
+            error = "Remove only works when the filesystem is initialised from a branch or default (None)"
+            raise ValueError(error)
+        r.raise_for_status()
+
+        self.invalidate_cache(path)
+
+    def _get_sha_from_cache(self, path):
+        for entries in self.dircache.values():
+            for entry in entries:
+                entry_path = entry.get("name")
+                if entry_path and entry_path == path and "sha" in entry:
+                    return entry["sha"]
+        return None

fsspec/implementations/local.py

@@ -6,6 +6,7 @@ import os.path as osp
 import shutil
 import stat
 import tempfile
+from functools import lru_cache
 
 from fsspec import AbstractFileSystem
 from fsspec.compression import compr
@@ -154,7 +155,11 @@ class LocalFileSystem(AbstractFileSystem):
     def put_file(self, path1, path2, callback=None, **kwargs):
         return self.cp_file(path1, path2, **kwargs)
 
-    def mv(self, path1, path2, **kwargs):
+    def mv(self, path1, path2, recursive: bool = True, **kwargs):
+        """Move files/directories
+        For the specific case of local, all ops on directories are recursive and
+        the recursive= kwarg is ignored.
+        """
         path1 = self._strip_protocol(path1)
         path2 = self._strip_protocol(path2)
         shutil.move(path1, path2)
@@ -350,6 +355,19 @@ def trailing_sep(path):
     return path.endswith(os.sep) or (os.altsep is not None and path.endswith(os.altsep))
 
 
+@lru_cache(maxsize=1)
+def get_umask(mask: int = 0o666) -> int:
+    """Get the current umask.
+
+    Follows https://stackoverflow.com/a/44130549 to get the umask.
+    Temporarily sets the umask to the given value, and then resets it to the
+    original value.
+    """
+    value = os.umask(mask)
+    os.umask(value)
+    return value
+
+
 class LocalFileOpener(io.IOBase):
     def __init__(
         self, path, mode, autocommit=True, fs=None, compression=None, **kwargs
@@ -412,7 +430,22 @@ class LocalFileOpener(io.IOBase):
     def commit(self):
         if self.autocommit:
             raise RuntimeError("Can only commit if not already set to autocommit")
-        shutil.move(self.temp, self.path)
+        try:
+            shutil.move(self.temp, self.path)
+        except PermissionError as e:
+            # shutil.move raises PermissionError if os.rename
+            # and the default copy2 fallback with shutil.copystats fail.
+            # The file should be there nonetheless, but without copied permissions.
+            # If it doesn't exist, there was no permission to create the file.
+            if not os.path.exists(self.path):
+                raise e
+        else:
+            # If PermissionError is not raised, permissions can be set.
+            try:
+                mask = 0o666
+                os.chmod(self.path, mask & ~get_umask(mask))
+            except RuntimeError:
+                pass
 
     def discard(self):
         if self.autocommit:

fsspec/registry.py

@@ -72,13 +72,13 @@ known_implementations = {
         "class": "fsspec.implementations.arrow.HadoopFileSystem",
         "err": "pyarrow and local java libraries required for HDFS",
     },
-    "async_wrapper": {
-        "class": "fsspec.asyn_wrapper.AsyncWrapperFileSystem",
-    },
     "asynclocal": {
         "class": "morefs.asyn_local.AsyncLocalFileSystem",
         "err": "Install 'morefs[asynclocalfs]' to use AsyncLocalFileSystem",
     },
+    "asyncwrapper": {
+        "class": "fsspec.implementations.asyn_wrapper.AsyncFileSystemWrapper",
+    },
     "az": {
         "class": "adlfs.AzureBlobFileSystem",
         "err": "Install adlfs to access Azure Datalake Gen2 and Azure Blob Storage",
@@ -122,6 +122,10 @@ known_implementations = {
         "err": "Please install gdrivefs for access to Google Drive",
     },
     "generic": {"class": "fsspec.generic.GenericFileSystem"},
+    "gist": {
+        "class": "fsspec.implementations.gist.GistFileSystem",
+        "err": "Install the requests package to use the gist FS",
+    },
     "git": {
         "class": "fsspec.implementations.git.GitFileSystem",
         "err": "Install pygit2 to browse local git repos",
@@ -180,6 +184,10 @@ known_implementations = {
         "class": "ossfs.OSSFileSystem",
         "err": "Install ossfs to access Alibaba Object Storage System",
     },
+    "pyscript": {
+        "class": "pyscript_fsspec_client.client.PyscriptFileSystem",
+        "err": "Install requests (cpython) or run in pyscript",
+    },
     "reference": {"class": "fsspec.implementations.reference.ReferenceFileSystem"},
     "root": {
         "class": "fsspec_xrootd.XRootDFileSystem",

fsspec/spec.py

@@ -548,17 +548,45 @@ class AbstractFileSystem(metaclass=_Cached):
             return sizes
 
     def glob(self, path, maxdepth=None, **kwargs):
-        """
-        Find files by glob-matching.
+        """Find files by glob-matching.
 
-        If the path ends with '/', only folders are returned.
+        Pattern matching capabilities for finding files that match the given pattern.
 
-        We support ``"**"``,
-        ``"?"`` and ``"[..]"``. We do not support ^ for pattern negation.
+        Parameters
+        ----------
+        path: str
+            The glob pattern to match against
+        maxdepth: int or None
+            Maximum depth for ``'**'`` patterns. Applied on the first ``'**'`` found.
+            Must be at least 1 if provided.
+        kwargs:
+            Additional arguments passed to ``find`` (e.g., detail=True)
 
-        The `maxdepth` option is applied on the first `**` found in the path.
+        Returns
+        -------
+        List of matched paths, or dict of paths and their info if detail=True
 
-        kwargs are passed to ``ls``.
+        Notes
+        -----
+        Supported patterns:
+        - '*': Matches any sequence of characters within a single directory level
+        - ``'**'``: Matches any number of directory levels (must be an entire path component)
+        - '?': Matches exactly one character
+        - '[abc]': Matches any character in the set
+        - '[a-z]': Matches any character in the range
+        - '[!abc]': Matches any character NOT in the set
+
+        Special behaviors:
+        - If the path ends with '/', only folders are returned
+        - Consecutive '*' characters are compressed into a single '*'
+        - Empty brackets '[]' never match anything
+        - Negated empty brackets '[!]' match any single character
+        - Special characters in character classes are escaped properly
+
+        Limitations:
+        - ``'**'`` must be a complete path component (e.g., ``'a/**/b'``, not ``'a**b'``)
+        - No brace expansion ('{a,b}.txt')
+        - No extended glob patterns ('+(pattern)', '!(pattern)')
         """
         if maxdepth is not None and maxdepth < 1:
             raise ValueError("maxdepth must be at least 1")

{fsspec-2025.3.2.dist-info → fsspec-2025.5.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fsspec
-Version: 2025.3.2
+Version: 2025.5.1
 Summary: File-system specification
 Project-URL: Changelog, https://filesystem-spec.readthedocs.io/en/latest/changelog.html
 Project-URL: Documentation, https://filesystem-spec.readthedocs.io/en/latest/

{fsspec-2025.3.2.dist-info → fsspec-2025.5.1.dist-info}/RECORD

@@ -1,5 +1,5 @@
-fsspec/__init__.py,sha256=l9MJaNNV2d4wKpCtMvXDr55n92DkdrAayGy3F9ICjzk,1998
-fsspec/_version.py,sha256=Qu6vPGmiev5F31vB8drXeCnOmd04CpjXkul_ZRHIKdg,517
+fsspec/__init__.py,sha256=L7qwNBU1iMNQd8Of87HYSNFT9gWlNMSESaJC8fY0AaQ,2053
+fsspec/_version.py,sha256=wECC04X-mZMQuTtKrW1mQDmwI4m9ylB6qWwnMihRB6I,517
 fsspec/archive.py,sha256=vM6t_lgV6lBWbBYwpm3S4ofBQFQxUPr5KkDQrrQcQro,2411
 fsspec/asyn.py,sha256=VJ2jBdYgUjV4_dETpKeCp2wF1XHAdeUET95d2HqNZck,36776
 fsspec/caching.py,sha256=n_SbdT-l92Kqo3e1BQgef0uEWJD0raP5-Qd8Ewp8CHY,34292
@@ -11,33 +11,34 @@ fsspec/core.py,sha256=1tLctwr7sF1VO3djc_UkjhJ8IAEy0TUMH_bb07Sw17E,23828
 fsspec/dircache.py,sha256=YzogWJrhEastHU7vWz-cJiJ7sdtLXFXhEpInGKd4EcM,2717
 fsspec/exceptions.py,sha256=pauSLDMxzTJMOjvX1WEUK0cMyFkrFxpWJsyFywav7A8,331
 fsspec/fuse.py,sha256=Q-3NOOyLqBfYa4Db5E19z_ZY36zzYHtIs1mOUasItBQ,10177
-fsspec/generic.py,sha256=AFbo-mHBt5QJV1Aplg5CJuUiiJ4bNQhcKRuwkZJdWac,13761
+fsspec/generic.py,sha256=6NIZX_Un78zkTtw46RoYubDPg_gPgO1blsjxytAVzu4,13449
 fsspec/gui.py,sha256=xBnHL2-r0LVwhDAtnHoPpXts7jd4Z32peawCJiI-7lI,13975
 fsspec/json.py,sha256=65sQ0Y7mTj33u_Y4IId5up4abQ3bAel4E4QzbKMiQSg,3826
 fsspec/mapping.py,sha256=m2ndB_gtRBXYmNJg0Ie1-BVR75TFleHmIQBzC-yWhjU,8343
 fsspec/parquet.py,sha256=6ibAmG527L5JNFS0VO8BDNlxHdA3bVYqdByeiFgpUVM,19448
-fsspec/registry.py,sha256=5kz-61fkb29lgSLzXWvk67ft6ooeYJR24Hs1583rD8w,11570
-fsspec/spec.py,sha256=l7ZEbgLsnrFuS-yrGl9re6ia1Yts1_10RqGV_mT-5P8,76032
+fsspec/registry.py,sha256=iPIyCIDcSKxgX7ppEFKHEmTubt8Z-YYpN0Eb3K94S3k,11893
+fsspec/spec.py,sha256=7cOUe5PC5Uyf56HtGBUHEoym8ktPj-BI8G4HR8Xd_C8,77298
 fsspec/transaction.py,sha256=xliRG6U2Zf3khG4xcw9WiB-yAoqJSHEGK_VjHOdtgo0,2398
 fsspec/utils.py,sha256=A11t25RnpiQ30RO6xeR0Qqlu3fGj8bnc40jg08tlYSI,22980
 fsspec/implementations/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 fsspec/implementations/arrow.py,sha256=721Dikne_lV_0tlgk9jyKmHL6W-5MT0h2LKGvOYQTPI,8623
-fsspec/implementations/asyn_wrapper.py,sha256=PNkYdHiLVWwk-GJok5O6dTnhPwDaSU9QTtBTE9CIRec,3082
+fsspec/implementations/asyn_wrapper.py,sha256=uZROca8lqiGOf5EILoAUjfalWoUU5CtseiE46l_3lkQ,3326
 fsspec/implementations/cache_mapper.py,sha256=W4wlxyPxZbSp9ItJ0pYRVBMh6bw9eFypgP6kUYuuiI4,2421
 fsspec/implementations/cache_metadata.py,sha256=pcOJYcBQY5OaC7Yhw0F3wjg08QLYApGmoISCrbs59ks,8511
-fsspec/implementations/cached.py,sha256=2EqeHZi6TKZzr222ZyXHjuYBnx_1g_HWXj1CJBFsXbc,33173
+fsspec/implementations/cached.py,sha256=FKEstAQxn5CyA5yM8NSl154ffm_k2wQ6_Za6C7ygIWg,33592
 fsspec/implementations/dask.py,sha256=CXZbJzIVOhKV8ILcxuy3bTvcacCueAbyQxmvAkbPkrk,4466
 fsspec/implementations/data.py,sha256=LDLczxRh8h7x39Zjrd-GgzdQHr78yYxDlrv2C9Uxb5E,1658
-fsspec/implementations/dbfs.py,sha256=XwpotuS_ncz3XK1dkUteww9GnTja7HoY91c0m4GUfwI,15092
+fsspec/implementations/dbfs.py,sha256=2Bp-0m9SqlaroDa0KbXxb5BobCyBJ7_5YQBISf3fxbQ,15145
 fsspec/implementations/dirfs.py,sha256=f1sGnQ9Vf0xTxrXo4jDeBy4Qfq3RTqAEemqBSeb0hwY,12108
 fsspec/implementations/ftp.py,sha256=sorsczLp_2J3ukONsbZY-11sRZP6H5a3V7XXf6o6ip0,11936
+fsspec/implementations/gist.py,sha256=Ost985hmFr50KsA-QD0shY3hP4KX5qJ9rb5C-X4ehK8,8341
 fsspec/implementations/git.py,sha256=4SElW9U5d3k3_ITlvUAx59Yk7XLNRTqkGa2C3hCUkWM,3754
-fsspec/implementations/github.py,sha256=tvLepeDtBsFHYocsD7XS1W6r8wxaoK57Us3kHGisnjU,9146
+fsspec/implementations/github.py,sha256=aCsZL8UvXZgdkcB1RUs3DdLeNrjLKcFsFYeQFDWbBFo,11653
 fsspec/implementations/http.py,sha256=_gLt0yGbVOYWvE9pK81WCC-3TgbOMOKJYllBU72ALo8,30138
 fsspec/implementations/http_sync.py,sha256=UydDqSdUBdhiJ1KufzV8rKGrTftFR4QmNV0safILb8g,30133
 fsspec/implementations/jupyter.py,sha256=B2uj7OEm7yIk-vRSsO37_ND0t0EBvn4B-Su43ibN4Pg,3811
 fsspec/implementations/libarchive.py,sha256=5_I2DiLXwQ1JC8x-K7jXu-tBwhO9dj7tFLnb0bTnVMQ,7102
-fsspec/implementations/local.py,sha256=g2iK8uWPGkSiI6bwmnIRXhJMQvTegCmXZ8Kb8ojhvAo,15543
+fsspec/implementations/local.py,sha256=38ylhASzWXSF0X41Bw8pYXyiM8xVu4UWDgE7l3Em8Uc,16768
 fsspec/implementations/memory.py,sha256=cLNrK9wk97sl4Tre9uVDXWj6mEHvvVVIgaVgNA5KVIg,10527
 fsspec/implementations/reference.py,sha256=t23prs_5ugXJnYhLxLlPLPyagrx4_ofZWR_oyX9wd3Q,48703
 fsspec/implementations/sftp.py,sha256=fMY9XZcmpjszQ2tCqO_TPaJesaeD_Dv7ptYzgUPGoO0,5631
@@ -53,7 +54,7 @@ fsspec/tests/abstract/mv.py,sha256=k8eUEBIrRrGMsBY5OOaDXdGnQUKGwDIfQyduB6YD3Ns,1
 fsspec/tests/abstract/open.py,sha256=Fi2PBPYLbRqysF8cFm0rwnB41kMdQVYjq8cGyDXp3BU,329
 fsspec/tests/abstract/pipe.py,sha256=LFzIrLCB5GLXf9rzFKJmE8AdG7LQ_h4bJo70r8FLPqM,402
 fsspec/tests/abstract/put.py,sha256=7aih17OKB_IZZh1Mkq1eBDIjobhtMQmI8x-Pw-S_aZk,21201
-fsspec-2025.3.2.dist-info/METADATA,sha256=4x2UMMjzVQTh6svjBem1yKpfCTuDqG-bSUWZiWRCFPg,11697
-fsspec-2025.3.2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-fsspec-2025.3.2.dist-info/licenses/LICENSE,sha256=LcNUls5TpzB5FcAIqESq1T53K0mzTN0ARFBnaRQH7JQ,1513
-fsspec-2025.3.2.dist-info/RECORD,,
+fsspec-2025.5.1.dist-info/METADATA,sha256=pL2uvv5MW4GapXukzVwdZGe3ghW4KRvh75lvDXTkMl4,11697
+fsspec-2025.5.1.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+fsspec-2025.5.1.dist-info/licenses/LICENSE,sha256=LcNUls5TpzB5FcAIqESq1T53K0mzTN0ARFBnaRQH7JQ,1513
+fsspec-2025.5.1.dist-info/RECORD,,