fsspec 2025.2.0__py3-none-any.whl → 2025.3.0__py3-none-any.whl

fsspec/_version.py

@@ -1,8 +1,13 @@
-# file generated by setuptools_scm
+# file generated by setuptools-scm
 # don't change, don't track in version control
+
+__all__ = ["__version__", "__version_tuple__", "version", "version_tuple"]
+
 TYPE_CHECKING = False
 if TYPE_CHECKING:
-    from typing import Tuple, Union
+    from typing import Tuple
+    from typing import Union
+
     VERSION_TUPLE = Tuple[Union[int, str], ...]
 else:
     VERSION_TUPLE = object
@@ -12,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '2025.2.0'
-__version_tuple__ = version_tuple = (2025, 2, 0)
+__version__ = version = '2025.3.0'
+__version_tuple__ = version_tuple = (2025, 3, 0)

fsspec/core.py

@@ -452,7 +452,7 @@ def open(
     newline: bytes or None
         Used for line terminator in text mode. If None, uses system default;
         if blank, uses no translation.
-    expand: bool or Nonw
+    expand: bool or None
         Whether to regard file paths containing special glob characters as needing
         expansion (finding the first match) or absolute. Setting False allows using
         paths which do embed such characters. If None (default), this argument

fsspec/implementations/asyn_wrapper.py

@@ -2,7 +2,7 @@ import asyncio
 import functools
 import inspect
 
-from fsspec.asyn import AsyncFileSystem
+from fsspec.asyn import AsyncFileSystem, running_async
 
 
 def async_wrapper(func, obj=None):
@@ -42,10 +42,14 @@ class AsyncFileSystemWrapper(AsyncFileSystem):
         The synchronous filesystem instance to wrap.
     """
 
-    def __init__(self, sync_fs, *args, **kwargs):
-        super().__init__(*args, **kwargs)
-        self.asynchronous = True
-        self.sync_fs = sync_fs
+    protocol = "async_wrapper"
+    cachable = False
+
+    def __init__(self, fs, *args, asynchronous=None, **kwargs):
+        if asynchronous is None:
+            asynchronous = running_async()
+        super().__init__(*args, asynchronous=asynchronous, **kwargs)
+        self.sync_fs = fs
         self.protocol = self.sync_fs.protocol
         self._wrap_all_sync_methods()
 

fsspec/implementations/dirfs.py

@@ -36,8 +36,6 @@ class DirFileSystem(AsyncFileSystem):
         super().__init__(**storage_options)
         if fs is None:
             fs = filesystem(protocol=target_protocol, **(target_options or {}))
-        if (path is not None) ^ (fo is not None) is False:
-            raise ValueError("Provide path or fo, not both")
         path = path or fo
 
         if self.asynchronous and not fs.async_impl:
@@ -233,10 +231,16 @@ class DirFileSystem(AsyncFileSystem):
         return self.fs.exists(self._join(path))
 
     async def _info(self, path, **kwargs):
-        return await self.fs._info(self._join(path), **kwargs)
+        info = await self.fs._info(self._join(path), **kwargs)
+        info = info.copy()
+        info["name"] = self._relpath(info["name"])
+        return info
 
     def info(self, path, **kwargs):
-        return self.fs.info(self._join(path), **kwargs)
+        info = self.fs.info(self._join(path), **kwargs)
+        info = info.copy()
+        info["name"] = self._relpath(info["name"])
+        return info
 
     async def _ls(self, path, detail=True, **kwargs):
         ret = (await self.fs._ls(self._join(path), detail=detail, **kwargs)).copy()

fsspec/implementations/http.py

@@ -522,6 +522,30 @@ class HTTPFileSystem(AsyncFileSystem):
         except (FileNotFoundError, ValueError):
             return False
 
+    async def _pipe_file(self, path, value, mode="overwrite", **kwargs):
+        """
+        Write bytes to a remote file over HTTP.
+
+        Parameters
+        ----------
+        path : str
+            Target URL where the data should be written
+        value : bytes
+            Data to be written
+        mode : str
+            How to write to the file - 'overwrite' or 'append'
+        **kwargs : dict
+            Additional parameters to pass to the HTTP request
+        """
+        url = self._strip_protocol(path)
+        headers = kwargs.pop("headers", {})
+        headers["Content-Length"] = str(len(value))
+
+        session = await self.set_session()
+
+        async with session.put(url, data=value, headers=headers, **kwargs) as r:
+            r.raise_for_status()
+
 
 class HTTPFile(AbstractBufferedFile):
     """

fsspec/implementations/http_sync.py

@@ -0,0 +1,932 @@
+"""This file is largely copied from http.py"""
+
+import io
+import logging
+import re
+import urllib.error
+import urllib.parse
+from copy import copy
+from json import dumps, loads
+from urllib.parse import urlparse
+
+try:
+    import yarl
+except (ImportError, ModuleNotFoundError, OSError):
+    yarl = False
+
+from fsspec.callbacks import _DEFAULT_CALLBACK
+from fsspec.registry import register_implementation
+from fsspec.spec import AbstractBufferedFile, AbstractFileSystem
+from fsspec.utils import DEFAULT_BLOCK_SIZE, isfilelike, nullcontext, tokenize
+
+from ..caching import AllBytes
+
+# https://stackoverflow.com/a/15926317/3821154
+ex = re.compile(r"""<(a|A)\s+(?:[^>]*?\s+)?(href|HREF)=["'](?P<url>[^"']+)""")
+ex2 = re.compile(r"""(?P<url>http[s]?://[-a-zA-Z0-9@:%_+.~#?&/=]+)""")
+logger = logging.getLogger("fsspec.http")
+
+
+class JsHttpException(urllib.error.HTTPError): ...
+
+
+class StreamIO(io.BytesIO):
+    # fake class, so you can set attributes on it
+    # will eventually actually stream
+    ...
+
+
+class ResponseProxy:
+    """Looks like a requests response"""
+
+    def __init__(self, req, stream=False):
+        self.request = req
+        self.stream = stream
+        self._data = None
+        self._headers = None
+
+    @property
+    def raw(self):
+        if self._data is None:
+            b = self.request.response.to_bytes()
+            if self.stream:
+                self._data = StreamIO(b)
+            else:
+                self._data = b
+        return self._data
+
+    def close(self):
+        if hasattr(self, "_data"):
+            del self._data
+
+    @property
+    def headers(self):
+        if self._headers is None:
+            self._headers = dict(
+                [
+                    _.split(": ")
+                    for _ in self.request.getAllResponseHeaders().strip().split("\r\n")
+                ]
+            )
+        return self._headers
+
+    @property
+    def status_code(self):
+        return int(self.request.status)
+
+    def raise_for_status(self):
+        if not self.ok:
+            raise JsHttpException(
+                self.url, self.status_code, self.reason, self.headers, None
+            )
+
+    def iter_content(self, chunksize, *_, **__):
+        while True:
+            out = self.raw.read(chunksize)
+            if out:
+                yield out
+            else:
+                break
+
+    @property
+    def reason(self):
+        return self.request.statusText
+
+    @property
+    def ok(self):
+        return self.status_code < 400
+
+    @property
+    def url(self):
+        return self.request.response.responseURL
+
+    @property
+    def text(self):
+        # TODO: encoding from headers
+        return self.content.decode()
+
+    @property
+    def content(self):
+        self.stream = False
+        return self.raw
+
+    @property
+    def json(self):
+        return loads(self.text)
+
+
+class RequestsSessionShim:
+    def __init__(self):
+        self.headers = {}
+
+    def request(
+        self,
+        method,
+        url,
+        params=None,
+        data=None,
+        headers=None,
+        cookies=None,
+        files=None,
+        auth=None,
+        timeout=None,
+        allow_redirects=None,
+        proxies=None,
+        hooks=None,
+        stream=None,
+        verify=None,
+        cert=None,
+        json=None,
+    ):
+        from js import Blob, XMLHttpRequest
+
+        logger.debug("JS request: %s %s", method, url)
+
+        if cert or verify or proxies or files or cookies or hooks:
+            raise NotImplementedError
+        if data and json:
+            raise ValueError("Use json= or data=, not both")
+        req = XMLHttpRequest.new()
+        extra = auth if auth else ()
+        if params:
+            url = f"{url}?{urllib.parse.urlencode(params)}"
+        req.open(method, url, False, *extra)
+        if timeout:
+            req.timeout = timeout
+        if headers:
+            for k, v in headers.items():
+                req.setRequestHeader(k, v)
+
+        req.setRequestHeader("Accept", "application/octet-stream")
+        req.responseType = "arraybuffer"
+        if json:
+            blob = Blob.new([dumps(data)], {type: "application/json"})
+            req.send(blob)
+        elif data:
+            if isinstance(data, io.IOBase):
+                data = data.read()
+            blob = Blob.new([data], {type: "application/octet-stream"})
+            req.send(blob)
+        else:
+            req.send(None)
+        return ResponseProxy(req, stream=stream)
+
+    def get(self, url, **kwargs):
+        return self.request("GET", url, **kwargs)
+
+    def head(self, url, **kwargs):
+        return self.request("HEAD", url, **kwargs)
+
+    def post(self, url, **kwargs):
+        return self.request("POST}", url, **kwargs)
+
+    def put(self, url, **kwargs):
+        return self.request("PUT", url, **kwargs)
+
+    def patch(self, url, **kwargs):
+        return self.request("PATCH", url, **kwargs)
+
+    def delete(self, url, **kwargs):
+        return self.request("DELETE", url, **kwargs)
+
+
+class HTTPFileSystem(AbstractFileSystem):
+    """
+    Simple File-System for fetching data via HTTP(S)
+
+    This is the BLOCKING version of the normal HTTPFileSystem. It uses
+    requests in normal python and the JS runtime in pyodide.
+
+    ***This implementation is extremely experimental, do not use unless
+    you are testing pyodide/pyscript integration***
+    """
+
+    protocol = ("http", "https", "sync_http", "sync_https")
+    sep = "/"
+
+    def __init__(
+        self,
+        simple_links=True,
+        block_size=None,
+        same_scheme=True,
+        cache_type="readahead",
+        cache_options=None,
+        client_kwargs=None,
+        encoded=False,
+        **storage_options,
+    ):
+        """
+
+        Parameters
+        ----------
+        block_size: int
+            Blocks to read bytes; if 0, will default to raw requests file-like
+            objects instead of HTTPFile instances
+        simple_links: bool
+            If True, will consider both HTML <a> tags and anything that looks
+            like a URL; if False, will consider only the former.
+        same_scheme: True
+            When doing ls/glob, if this is True, only consider paths that have
+            http/https matching the input URLs.
+        size_policy: this argument is deprecated
+        client_kwargs: dict
+            Passed to aiohttp.ClientSession, see
+            https://docs.aiohttp.org/en/stable/client_reference.html
+            For example, ``{'auth': aiohttp.BasicAuth('user', 'pass')}``
+        storage_options: key-value
+            Any other parameters passed on to requests
+        cache_type, cache_options: defaults used in open
+        """
+        super().__init__(self, **storage_options)
+        self.block_size = block_size if block_size is not None else DEFAULT_BLOCK_SIZE
+        self.simple_links = simple_links
+        self.same_schema = same_scheme
+        self.cache_type = cache_type
+        self.cache_options = cache_options
+        self.client_kwargs = client_kwargs or {}
+        self.encoded = encoded
+        self.kwargs = storage_options
+
+        try:
+            import js  # noqa: F401
+
+            logger.debug("Starting JS session")
+            self.session = RequestsSessionShim()
+            self.js = True
+        except Exception as e:
+            import requests
+
+            logger.debug("Starting cpython session because of: %s", e)
+            self.session = requests.Session(**(client_kwargs or {}))
+            self.js = False
+
+        request_options = copy(storage_options)
+        self.use_listings_cache = request_options.pop("use_listings_cache", False)
+        request_options.pop("listings_expiry_time", None)
+        request_options.pop("max_paths", None)
+        request_options.pop("skip_instance_cache", None)
+        self.kwargs = request_options
+
+    @property
+    def fsid(self):
+        return "http_sync"
+
+    def encode_url(self, url):
+        if yarl:
+            return yarl.URL(url, encoded=self.encoded)
+        return url
+
+    @classmethod
+    def _strip_protocol(cls, path: str) -> str:
+        """For HTTP, we always want to keep the full URL"""
+        path = path.replace("http_sync://", "http://").replace(
+            "https_sync://", "https://"
+        )
+        return path
+
+    @classmethod
+    def _parent(cls, path):
+        # override, since _strip_protocol is different for URLs
+        par = super()._parent(path)
+        if len(par) > 7:  # "http://..."
+            return par
+        return ""
+
+    def _ls_real(self, url, detail=True, **kwargs):
+        # ignoring URL-encoded arguments
+        kw = self.kwargs.copy()
+        kw.update(kwargs)
+        logger.debug(url)
+        r = self.session.get(self.encode_url(url), **self.kwargs)
+        self._raise_not_found_for_status(r, url)
+        text = r.text
+        if self.simple_links:
+            links = ex2.findall(text) + [u[2] for u in ex.findall(text)]
+        else:
+            links = [u[2] for u in ex.findall(text)]
+        out = set()
+        parts = urlparse(url)
+        for l in links:
+            if isinstance(l, tuple):
+                l = l[1]
+            if l.startswith("/") and len(l) > 1:
+                # absolute URL on this server
+                l = parts.scheme + "://" + parts.netloc + l
+            if l.startswith("http"):
+                if self.same_schema and l.startswith(url.rstrip("/") + "/"):
+                    out.add(l)
+                elif l.replace("https", "http").startswith(
+                    url.replace("https", "http").rstrip("/") + "/"
+                ):
+                    # allowed to cross http <-> https
+                    out.add(l)
+            else:
+                if l not in ["..", "../"]:
+                    # Ignore FTP-like "parent"
+                    out.add("/".join([url.rstrip("/"), l.lstrip("/")]))
+        if not out and url.endswith("/"):
+            out = self._ls_real(url.rstrip("/"), detail=False)
+        if detail:
+            return [
+                {
+                    "name": u,
+                    "size": None,
+                    "type": "directory" if u.endswith("/") else "file",
+                }
+                for u in out
+            ]
+        else:
+            return sorted(out)
+
+    def ls(self, url, detail=True, **kwargs):
+        if self.use_listings_cache and url in self.dircache:
+            out = self.dircache[url]
+        else:
+            out = self._ls_real(url, detail=detail, **kwargs)
+            self.dircache[url] = out
+        return out
+
+    def _raise_not_found_for_status(self, response, url):
+        """
+        Raises FileNotFoundError for 404s, otherwise uses raise_for_status.
+        """
+        if response.status_code == 404:
+            raise FileNotFoundError(url)
+        response.raise_for_status()
+
+    def cat_file(self, url, start=None, end=None, **kwargs):
+        kw = self.kwargs.copy()
+        kw.update(kwargs)
+        logger.debug(url)
+
+        if start is not None or end is not None:
+            if start == end:
+                return b""
+            headers = kw.pop("headers", {}).copy()
+
+            headers["Range"] = self._process_limits(url, start, end)
+            kw["headers"] = headers
+        r = self.session.get(self.encode_url(url), **kw)
+        self._raise_not_found_for_status(r, url)
+        return r.content
+
+    def get_file(
+        self, rpath, lpath, chunk_size=5 * 2**20, callback=_DEFAULT_CALLBACK, **kwargs
+    ):
+        kw = self.kwargs.copy()
+        kw.update(kwargs)
+        logger.debug(rpath)
+        r = self.session.get(self.encode_url(rpath), **kw)
+        try:
+            size = int(
+                r.headers.get("content-length", None)
+                or r.headers.get("Content-Length", None)
+            )
+        except (ValueError, KeyError, TypeError):
+            size = None
+
+        callback.set_size(size)
+        self._raise_not_found_for_status(r, rpath)
+        if not isfilelike(lpath):
+            lpath = open(lpath, "wb")
+        for chunk in r.iter_content(chunk_size, decode_unicode=False):
+            lpath.write(chunk)
+            callback.relative_update(len(chunk))
+
+    def put_file(
+        self,
+        lpath,
+        rpath,
+        chunk_size=5 * 2**20,
+        callback=_DEFAULT_CALLBACK,
+        method="post",
+        **kwargs,
+    ):
+        def gen_chunks():
+            # Support passing arbitrary file-like objects
+            # and use them instead of streams.
+            if isinstance(lpath, io.IOBase):
+                context = nullcontext(lpath)
+                use_seek = False  # might not support seeking
+            else:
+                context = open(lpath, "rb")
+                use_seek = True
+
+            with context as f:
+                if use_seek:
+                    callback.set_size(f.seek(0, 2))
+                    f.seek(0)
+                else:
+                    callback.set_size(getattr(f, "size", None))
+
+                chunk = f.read(chunk_size)
+                while chunk:
+                    yield chunk
+                    callback.relative_update(len(chunk))
+                    chunk = f.read(chunk_size)
+
+        kw = self.kwargs.copy()
+        kw.update(kwargs)
+
+        method = method.lower()
+        if method not in ("post", "put"):
+            raise ValueError(
+                f"method has to be either 'post' or 'put', not: {method!r}"
+            )
+
+        meth = getattr(self.session, method)
+        resp = meth(rpath, data=gen_chunks(), **kw)
+        self._raise_not_found_for_status(resp, rpath)
+
+    def _process_limits(self, url, start, end):
+        """Helper for "Range"-based _cat_file"""
+        size = None
+        suff = False
+        if start is not None and start < 0:
+            # if start is negative and end None, end is the "suffix length"
+            if end is None:
+                end = -start
+                start = ""
+                suff = True
+            else:
+                size = size or self.info(url)["size"]
+                start = size + start
+        elif start is None:
+            start = 0
+        if not suff:
+            if end is not None and end < 0:
+                if start is not None:
+                    size = size or self.info(url)["size"]
+                    end = size + end
+            elif end is None:
+                end = ""
+            if isinstance(end, int):
+                end -= 1  # bytes range is inclusive
+        return f"bytes={start}-{end}"
+
+    def exists(self, path, **kwargs):
+        kw = self.kwargs.copy()
+        kw.update(kwargs)
+        try:
+            logger.debug(path)
+            r = self.session.get(self.encode_url(path), **kw)
+            return r.status_code < 400
+        except Exception:
+            return False
+
+    def isfile(self, path, **kwargs):
+        return self.exists(path, **kwargs)
+
+    def _open(
+        self,
+        path,
+        mode="rb",
+        block_size=None,
+        autocommit=None,  # XXX: This differs from the base class.
+        cache_type=None,
+        cache_options=None,
+        size=None,
+        **kwargs,
+    ):
+        """Make a file-like object
+
+        Parameters
+        ----------
+        path: str
+            Full URL with protocol
+        mode: string
+            must be "rb"
+        block_size: int or None
+            Bytes to download in one request; use instance value if None. If
+            zero, will return a streaming Requests file-like instance.
+        kwargs: key-value
+            Any other parameters, passed to requests calls
+        """
+        if mode != "rb":
+            raise NotImplementedError
+        block_size = block_size if block_size is not None else self.block_size
+        kw = self.kwargs.copy()
+        kw.update(kwargs)
+        size = size or self.info(path, **kwargs)["size"]
+        if block_size and size:
+            return HTTPFile(
+                self,
+                path,
+                session=self.session,
+                block_size=block_size,
+                mode=mode,
+                size=size,
+                cache_type=cache_type or self.cache_type,
+                cache_options=cache_options or self.cache_options,
+                **kw,
+            )
+        else:
+            return HTTPStreamFile(
+                self,
+                path,
+                mode=mode,
+                session=self.session,
+                **kw,
+            )
+
+    def ukey(self, url):
+        """Unique identifier; assume HTTP files are static, unchanging"""
+        return tokenize(url, self.kwargs, self.protocol)
+
+    def info(self, url, **kwargs):
+        """Get info of URL
+
+        Tries to access location via HEAD, and then GET methods, but does
+        not fetch the data.
+
+        It is possible that the server does not supply any size information, in
+        which case size will be given as None (and certain operations on the
+        corresponding file will not work).
+        """
+        info = {}
+        for policy in ["head", "get"]:
+            try:
+                info.update(
+                    _file_info(
+                        self.encode_url(url),
+                        size_policy=policy,
+                        session=self.session,
+                        **self.kwargs,
+                        **kwargs,
+                    )
+                )
+                if info.get("size") is not None:
+                    break
+            except Exception as exc:
+                if policy == "get":
+                    # If get failed, then raise a FileNotFoundError
+                    raise FileNotFoundError(url) from exc
+                logger.debug(str(exc))
+
+        return {"name": url, "size": None, **info, "type": "file"}
+
+    def glob(self, path, maxdepth=None, **kwargs):
+        """
+        Find files by glob-matching.
+
+        This implementation is idntical to the one in AbstractFileSystem,
+        but "?" is not considered as a character for globbing, because it is
+        so common in URLs, often identifying the "query" part.
+        """
+        import re
+
+        ends = path.endswith("/")
+        path = self._strip_protocol(path)
+        indstar = path.find("*") if path.find("*") >= 0 else len(path)
+        indbrace = path.find("[") if path.find("[") >= 0 else len(path)
+
+        ind = min(indstar, indbrace)
+
+        detail = kwargs.pop("detail", False)
+
+        if not has_magic(path):
+            root = path
+            depth = 1
+            if ends:
+                path += "/*"
+            elif self.exists(path):
+                if not detail:
+                    return [path]
+                else:
+                    return {path: self.info(path)}
+            else:
+                if not detail:
+                    return []  # glob of non-existent returns empty
+                else:
+                    return {}
+        elif "/" in path[:ind]:
+            ind2 = path[:ind].rindex("/")
+            root = path[: ind2 + 1]
+            depth = None if "**" in path else path[ind2 + 1 :].count("/") + 1
+        else:
+            root = ""
+            depth = None if "**" in path else path[ind + 1 :].count("/") + 1
+
+        allpaths = self.find(
+            root, maxdepth=maxdepth or depth, withdirs=True, detail=True, **kwargs
+        )
+        # Escape characters special to python regex, leaving our supported
+        # special characters in place.
+        # See https://www.gnu.org/software/bash/manual/html_node/Pattern-Matching.html
+        # for shell globbing details.
+        pattern = (
+            "^"
+            + (
+                path.replace("\\", r"\\")
+                .replace(".", r"\.")
+                .replace("+", r"\+")
+                .replace("//", "/")
+                .replace("(", r"\(")
+                .replace(")", r"\)")
+                .replace("|", r"\|")
+                .replace("^", r"\^")
+                .replace("$", r"\$")
+                .replace("{", r"\{")
+                .replace("}", r"\}")
+                .rstrip("/")
+            )
+            + "$"
+        )
+        pattern = re.sub("[*]{2}", "=PLACEHOLDER=", pattern)
+        pattern = re.sub("[*]", "[^/]*", pattern)
+        pattern = re.compile(pattern.replace("=PLACEHOLDER=", ".*"))
+        out = {
+            p: allpaths[p]
+            for p in sorted(allpaths)
+            if pattern.match(p.replace("//", "/").rstrip("/"))
+        }
+        if detail:
+            return out
+        else:
+            return list(out)
+
+    def isdir(self, path):
+        # override, since all URLs are (also) files
+        try:
+            return bool(self.ls(path))
+        except (FileNotFoundError, ValueError):
+            return False
+
+
+class HTTPFile(AbstractBufferedFile):
+    """
+    A file-like object pointing to a remove HTTP(S) resource
+
+    Supports only reading, with read-ahead of a predermined block-size.
+
+    In the case that the server does not supply the filesize, only reading of
+    the complete file in one go is supported.
+
+    Parameters
+    ----------
+    url: str
+        Full URL of the remote resource, including the protocol
+    session: requests.Session or None
+        All calls will be made within this session, to avoid restarting
+        connections where the server allows this
+    block_size: int or None
+        The amount of read-ahead to do, in bytes. Default is 5MB, or the value
+        configured for the FileSystem creating this file
+    size: None or int
+        If given, this is the size of the file in bytes, and we don't attempt
+        to call the server to find the value.
+    kwargs: all other key-values are passed to requests calls.
+    """
+
+    def __init__(
+        self,
+        fs,
+        url,
+        session=None,
+        block_size=None,
+        mode="rb",
+        cache_type="bytes",
+        cache_options=None,
+        size=None,
+        **kwargs,
+    ):
+        if mode != "rb":
+            raise NotImplementedError("File mode not supported")
+        self.url = url
+        self.session = session
+        self.details = {"name": url, "size": size, "type": "file"}
+        super().__init__(
+            fs=fs,
+            path=url,
+            mode=mode,
+            block_size=block_size,
+            cache_type=cache_type,
+            cache_options=cache_options,
+            **kwargs,
+        )
+
+    def read(self, length=-1):
+        """Read bytes from file
+
+        Parameters
+        ----------
+        length: int
+            Read up to this many bytes. If negative, read all content to end of
+            file. If the server has not supplied the filesize, attempting to
+            read only part of the data will raise a ValueError.
+        """
+        if (
+            (length < 0 and self.loc == 0)  # explicit read all
+            # but not when the size is known and fits into a block anyways
+            and not (self.size is not None and self.size <= self.blocksize)
+        ):
+            self._fetch_all()
+        if self.size is None:
+            if length < 0:
+                self._fetch_all()
+        else:
+            length = min(self.size - self.loc, length)
+        return super().read(length)
+
+    def _fetch_all(self):
+        """Read whole file in one shot, without caching
+
+        This is only called when position is still at zero,
+        and read() is called without a byte-count.
+        """
+        logger.debug(f"Fetch all for {self}")
+        if not isinstance(self.cache, AllBytes):
+            r = self.session.get(self.fs.encode_url(self.url), **self.kwargs)
+            r.raise_for_status()
+            out = r.content
+            self.cache = AllBytes(size=len(out), fetcher=None, blocksize=None, data=out)
+            self.size = len(out)
+
+    def _parse_content_range(self, headers):
+        """Parse the Content-Range header"""
+        s = headers.get("Content-Range", "")
+        m = re.match(r"bytes (\d+-\d+|\*)/(\d+|\*)", s)
+        if not m:
+            return None, None, None
+
+        if m[1] == "*":
+            start = end = None
+        else:
+            start, end = [int(x) for x in m[1].split("-")]
+        total = None if m[2] == "*" else int(m[2])
+        return start, end, total
+
+    def _fetch_range(self, start, end):
+        """Download a block of data
+
+        The expectation is that the server returns only the requested bytes,
+        with HTTP code 206. If this is not the case, we first check the headers,
+        and then stream the output - if the data size is bigger than we
+        requested, an exception is raised.
+        """
+        logger.debug(f"Fetch range for {self}: {start}-{end}")
+        kwargs = self.kwargs.copy()
+        headers = kwargs.pop("headers", {}).copy()
+        headers["Range"] = f"bytes={start}-{end - 1}"
+        logger.debug("%s : %s", self.url, headers["Range"])
+        r = self.session.get(self.fs.encode_url(self.url), headers=headers, **kwargs)
+        if r.status_code == 416:
+            # range request outside file
+            return b""
+        r.raise_for_status()
+
+        # If the server has handled the range request, it should reply
+        # with status 206 (partial content). But we'll guess that a suitable
+        # Content-Range header or a Content-Length no more than the
+        # requested range also mean we have got the desired range.
+        cl = r.headers.get("Content-Length", r.headers.get("content-length", end + 1))
+        response_is_range = (
+            r.status_code == 206
+            or self._parse_content_range(r.headers)[0] == start
+            or int(cl) <= end - start
+        )
+
+        if response_is_range:
+            # partial content, as expected
+            out = r.content
+        elif start > 0:
+            raise ValueError(
+                "The HTTP server doesn't appear to support range requests. "
+                "Only reading this file from the beginning is supported. "
+                "Open with block_size=0 for a streaming file interface."
+            )
+        else:
+            # Response is not a range, but we want the start of the file,
+            # so we can read the required amount anyway.
+            cl = 0
+            out = []
+            for chunk in r.iter_content(2**20, False):
+                out.append(chunk)
+                cl += len(chunk)
+            out = b"".join(out)[: end - start]
+        return out
+
+
+magic_check = re.compile("([*[])")
+
+
+def has_magic(s):
+    match = magic_check.search(s)
+    return match is not None
+
+
+class HTTPStreamFile(AbstractBufferedFile):
+    def __init__(self, fs, url, mode="rb", session=None, **kwargs):
+        self.url = url
+        self.session = session
+        if mode != "rb":
+            raise ValueError
+        self.details = {"name": url, "size": None}
+        super().__init__(fs=fs, path=url, mode=mode, cache_type="readahead", **kwargs)
+
+        r = self.session.get(self.fs.encode_url(url), stream=True, **kwargs)
+        self.fs._raise_not_found_for_status(r, url)
+        self.it = r.iter_content(1024, False)
+        self.leftover = b""
+
+        self.r = r
+
+    def seek(self, *args, **kwargs):
+        raise ValueError("Cannot seek streaming HTTP file")
+
+    def read(self, num=-1):
+        bufs = [self.leftover]
+        leng = len(self.leftover)
+        while leng < num or num < 0:
+            try:
+                out = self.it.__next__()
+            except StopIteration:
+                break
+            if out:
+                bufs.append(out)
+            else:
+                break
+            leng += len(out)
+        out = b"".join(bufs)
+        if num >= 0:
+            self.leftover = out[num:]
+            out = out[:num]
+        else:
+            self.leftover = b""
+        self.loc += len(out)
+        return out
+
+    def close(self):
+        self.r.close()
+        self.closed = True
+
+
+def get_range(session, url, start, end, **kwargs):
+    # explicit get a range when we know it must be safe
+    kwargs = kwargs.copy()
+    headers = kwargs.pop("headers", {}).copy()
+    headers["Range"] = f"bytes={start}-{end - 1}"
+    r = session.get(url, headers=headers, **kwargs)
+    r.raise_for_status()
+    return r.content
+
+
+def _file_info(url, session, size_policy="head", **kwargs):
+    """Call HEAD on the server to get details about the file (size/checksum etc.)
+
+    Default operation is to explicitly allow redirects and use encoding
+    'identity' (no compression) to get the true size of the target.
+    """
+    logger.debug("Retrieve file size for %s", url)
+    kwargs = kwargs.copy()
+    ar = kwargs.pop("allow_redirects", True)
+    head = kwargs.get("headers", {}).copy()
+    # TODO: not allowed in JS
+    # head["Accept-Encoding"] = "identity"
+    kwargs["headers"] = head
+
+    info = {}
+    if size_policy == "head":
+        r = session.head(url, allow_redirects=ar, **kwargs)
+    elif size_policy == "get":
+        r = session.get(url, allow_redirects=ar, **kwargs)
+    else:
+        raise TypeError(f'size_policy must be "head" or "get", got {size_policy}')
+    r.raise_for_status()
+
+    # TODO:
+    #  recognise lack of 'Accept-Ranges',
+    #                 or 'Accept-Ranges': 'none' (not 'bytes')
+    #  to mean streaming only, no random access => return None
+    if "Content-Length" in r.headers:
+        info["size"] = int(r.headers["Content-Length"])
+    elif "Content-Range" in r.headers:
+        info["size"] = int(r.headers["Content-Range"].split("/")[1])
+    elif "content-length" in r.headers:
+        info["size"] = int(r.headers["content-length"])
+    elif "content-range" in r.headers:
+        info["size"] = int(r.headers["content-range"].split("/")[1])
+
+    for checksum_field in ["ETag", "Content-MD5", "Digest"]:
+        if r.headers.get(checksum_field):
+            info[checksum_field] = r.headers[checksum_field]
+
+    return info
+
+
+# importing this is enough to register it
+def register():
+    register_implementation("http", HTTPFileSystem, clobber=True)
+    register_implementation("https", HTTPFileSystem, clobber=True)
+    register_implementation("sync_http", HTTPFileSystem, clobber=True)
+    register_implementation("sync_https", HTTPFileSystem, clobber=True)
+
+
+register()
+
+
+def unregister():
+    from fsspec.implementations.http import HTTPFileSystem
+
+    register_implementation("http", HTTPFileSystem, clobber=True)
+    register_implementation("https", HTTPFileSystem, clobber=True)

fsspec/implementations/local.py

@@ -57,20 +57,21 @@ class LocalFileSystem(AbstractFileSystem):
 
     def ls(self, path, detail=False, **kwargs):
         path = self._strip_protocol(path)
-        info = self.info(path)
-        if info["type"] == "directory":
+        path_info = self.info(path)
+        infos = []
+        if path_info["type"] == "directory":
             with os.scandir(path) as it:
-                infos = []
                 for f in it:
                     try:
-                        infos.append(self.info(f))
+                        # Only get the info if requested since it is a bit expensive (the stat call inside)
+                        # The strip_protocol is also used in info() and calls make_path_posix to always return posix paths
+                        info = self.info(f) if detail else self._strip_protocol(f.path)
+                        infos.append(info)
                     except FileNotFoundError:
                         pass
         else:
-            infos = [info]
+            infos = [path_info] if detail else [path_info["name"]]
 
-        if not detail:
-            return [i["name"] for i in infos]
         return infos
 
     def info(self, path, **kwargs):

fsspec/implementations/reference.py

@@ -140,13 +140,13 @@ class LazyReferenceMapper(collections.abc.MutableMapping):
 
         self.root = root
         self.chunk_sizes = {}
-        self.out_root = out_root or self.root
         self.cat_thresh = categorical_threshold
         self.engine = engine
         self.cache_size = cache_size
         self.url = self.root + "/{field}/refs.{record}.parq"
         # TODO: derive fs from `root`
         self.fs = fsspec.filesystem("file") if fs is None else fs
+        self.out_root = self.fs.unstrip_protocol(out_root or self.root)
 
         from importlib.util import find_spec
 
@@ -498,7 +498,6 @@ class LazyReferenceMapper(collections.abc.MutableMapping):
             }
         else:
             raise NotImplementedError(f"{self.engine} not supported")
-
         df.to_parquet(
             fn,
             engine=self.engine,
@@ -766,7 +765,7 @@ class ReferenceFileSystem(AsyncFileSystem):
         # Wrap any non-async filesystems to ensure async methods are available below
         for k, f in self.fss.items():
             if not f.async_impl:
-                self.fss[k] = AsyncFileSystemWrapper(f)
+                self.fss[k] = AsyncFileSystemWrapper(f, asynchronous=self.asynchronous)
             elif self.asynchronous ^ f.asynchronous:
                 raise ValueError(
                     "Reference-FS's target filesystem must have same value"

fsspec/registry.py

@@ -72,6 +72,9 @@ known_implementations = {
         "class": "fsspec.implementations.arrow.HadoopFileSystem",
         "err": "pyarrow and local java libraries required for HDFS",
     },
+    "async_wrapper": {
+        "class": "morefs.asyn_wrapper.AsyncWrapperFileSystem",
+    },
     "asynclocal": {
         "class": "morefs.asyn_local.AsyncLocalFileSystem",
         "err": "Install 'morefs[asynclocalfs]' to use AsyncLocalFileSystem",

{fsspec-2025.2.0.dist-info → fsspec-2025.3.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fsspec
-Version: 2025.2.0
+Version: 2025.3.0
 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-2025.2.0.dist-info → fsspec-2025.3.0.dist-info}/RECORD

@@ -1,5 +1,5 @@
 fsspec/__init__.py,sha256=l9MJaNNV2d4wKpCtMvXDr55n92DkdrAayGy3F9ICjzk,1998
-fsspec/_version.py,sha256=IE7d_vZlkju9WTb8xdQYMiqPyQOYnfC9HN9w8nHfkrY,417
+fsspec/_version.py,sha256=mWdw-j-kCQGfmn6901GQSyjFXi2pK3h-NrKTuLRVqCw,517
 fsspec/archive.py,sha256=vM6t_lgV6lBWbBYwpm3S4ofBQFQxUPr5KkDQrrQcQro,2411
 fsspec/asyn.py,sha256=rsnCsFUmBZmKJqg9m-IDWInoQtE4wV0rGDZEXZwuU3c,36500
 fsspec/caching.py,sha256=oHVy9zpy4Oqk5f1t3-Q31bbw0tsmfddGGKLJs__OdKA,32790
@@ -7,7 +7,7 @@ fsspec/callbacks.py,sha256=BDIwLzK6rr_0V5ch557fSzsivCElpdqhXr5dZ9Te-EE,9210
 fsspec/compression.py,sha256=jCSUMJu-zSNyrusnHT0wKXgOd1tTJR6vM126i5SR5Zc,4865
 fsspec/config.py,sha256=LF4Zmu1vhJW7Je9Q-cwkRc3xP7Rhyy7Xnwj26Z6sv2g,4279
 fsspec/conftest.py,sha256=fVfx-NLrH_OZS1TIpYNoPzM7efEcMoL62reHOdYeFCA,1245
-fsspec/core.py,sha256=bn-y3Mn9q8Gh3Ng_yAIDfIjyysQ95tuK78RlhlrqTb4,23828
+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
@@ -16,29 +16,30 @@ 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=QFyMiUV6fnksETJuapNplf6YjkNRIdHSOyd95IqPZe8,11473
+fsspec/registry.py,sha256=Bbur6KJilN62hx_lSRgCx9HlKrv91pLpdEPG7Vzme1M,11566
 fsspec/spec.py,sha256=l7ZEbgLsnrFuS-yrGl9re6ia1Yts1_10RqGV_mT-5P8,76032
 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=gmLy2voDAH9KRxhvd24UDPiOqX_NCK-3JY9rMX7R6Is,2935
+fsspec/implementations/asyn_wrapper.py,sha256=PNkYdHiLVWwk-GJok5O6dTnhPwDaSU9QTtBTE9CIRec,3082
 fsspec/implementations/cache_mapper.py,sha256=W4wlxyPxZbSp9ItJ0pYRVBMh6bw9eFypgP6kUYuuiI4,2421
 fsspec/implementations/cache_metadata.py,sha256=pcOJYcBQY5OaC7Yhw0F3wjg08QLYApGmoISCrbs59ks,8511
 fsspec/implementations/cached.py,sha256=KA6c4jqrGeeg8WNPLsh8FkL3KeRAQtGLzKw18vSF1CI,32820
 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/dirfs.py,sha256=ymakitNNQ07tW76EShyw3rC9RvIDHl4gtuOhE_h1vUg,12032
+fsspec/implementations/dirfs.py,sha256=f1sGnQ9Vf0xTxrXo4jDeBy4Qfq3RTqAEemqBSeb0hwY,12108
 fsspec/implementations/ftp.py,sha256=sorsczLp_2J3ukONsbZY-11sRZP6H5a3V7XXf6o6ip0,11936
 fsspec/implementations/git.py,sha256=4SElW9U5d3k3_ITlvUAx59Yk7XLNRTqkGa2C3hCUkWM,3754
 fsspec/implementations/github.py,sha256=eAn1kJ7VeWR6gVoVRLBYclF_rQDXSJU-xzMXpvPQWqs,8002
-fsspec/implementations/http.py,sha256=d7G7_pRTMHouKE42lvRNHqB5u4XQi0dm4wb-6U_IiF4,29361
+fsspec/implementations/http.py,sha256=_gLt0yGbVOYWvE9pK81WCC-3TgbOMOKJYllBU72ALo8,30138
+fsspec/implementations/http_sync.py,sha256=vHf2_O9RRlaW6k-R4hsS-91BkqyPIlc0zwo_ENNGU4U,30147
 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=YvR9b2MndSQIHszAMUkFvN65eWVbIfoGJJjAeS43ZS4,15259
+fsspec/implementations/local.py,sha256=g2iK8uWPGkSiI6bwmnIRXhJMQvTegCmXZ8Kb8ojhvAo,15543
 fsspec/implementations/memory.py,sha256=cLNrK9wk97sl4Tre9uVDXWj6mEHvvVVIgaVgNA5KVIg,10527
-fsspec/implementations/reference.py,sha256=1VbyjAxq_8xHSQo2UV4ohuuoSAreB3OY4vjK05DnHsY,48646
+fsspec/implementations/reference.py,sha256=t23prs_5ugXJnYhLxLlPLPyagrx4_ofZWR_oyX9wd3Q,48703
 fsspec/implementations/sftp.py,sha256=fMY9XZcmpjszQ2tCqO_TPaJesaeD_Dv7ptYzgUPGoO0,5631
 fsspec/implementations/smb.py,sha256=5fhu8h06nOLBPh2c48aT7WBRqh9cEcbIwtyu06wTjec,15236
 fsspec/implementations/tar.py,sha256=dam78Tp_CozybNqCY2JYgGBS3Uc9FuJUAT9oB0lolOs,4111
@@ -52,7 +53,7 @@ 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-2025.2.0.dist-info/METADATA,sha256=GqrJBTBVJNmkbcEfTOamzq2KqYYbc9vRM0jpSWTZT04,11747
-fsspec-2025.2.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-fsspec-2025.2.0.dist-info/licenses/LICENSE,sha256=LcNUls5TpzB5FcAIqESq1T53K0mzTN0ARFBnaRQH7JQ,1513
-fsspec-2025.2.0.dist-info/RECORD,,
+fsspec-2025.3.0.dist-info/METADATA,sha256=L2OjjikjNWpjvp8i5_ec515Cd09eUi8qZ_3peQmWG8M,11747
+fsspec-2025.3.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+fsspec-2025.3.0.dist-info/licenses/LICENSE,sha256=LcNUls5TpzB5FcAIqESq1T53K0mzTN0ARFBnaRQH7JQ,1513
+fsspec-2025.3.0.dist-info/RECORD,,