fsspec 2024.12.0__py3-none-any.whl → 2025.3.0__py3-none-any.whl

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

@@ -10,6 +10,7 @@ from itertools import chain
 from typing import TYPE_CHECKING, Literal
 
 import fsspec.core
+from fsspec.spec import AbstractBufferedFile
 
 try:
     import ujson as json
@@ -139,13 +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
 
@@ -394,10 +395,14 @@ class LazyReferenceMapper(collections.abc.MutableMapping):
                 self.write(field, record)
         else:
             # metadata or top-level
-            self._items[key] = value
-            new_value = json.loads(
-                value.decode() if isinstance(value, bytes) else value
-            )
+            if hasattr(value, "to_bytes"):
+                val = value.to_bytes().decode()
+            elif isinstance(value, bytes):
+                val = value.decode()
+            else:
+                val = value
+            self._items[key] = val
+            new_value = json.loads(val)
             self.zmetadata[key] = {**self.zmetadata.get(key, {}), **new_value}
 
     @staticmethod
@@ -493,7 +498,6 @@ class LazyReferenceMapper(collections.abc.MutableMapping):
             }
         else:
             raise NotImplementedError(f"{self.engine} not supported")
-
         df.to_parquet(
             fn,
             engine=self.engine,
@@ -595,8 +599,7 @@ class ReferenceFileSystem(AsyncFileSystem):
     async, and must allow start and end args in _cat_file. Later versions
     may allow multiple arbitrary URLs for the targets.
     This FileSystem is read-only. It is designed to be used with async
-    targets (for now). This FileSystem only allows whole-file access, no
-    ``open``. We do not get original file details from the target FS.
+    targets (for now). We do not get original file details from the target FS.
     Configuration is by passing a dict of references at init, or a URL to
     a JSON file containing the same; this dict
     can also contain concrete data for some set of paths.
@@ -606,6 +609,7 @@ class ReferenceFileSystem(AsyncFileSystem):
     """
 
     protocol = "reference"
+    cachable = False
 
     def __init__(
         self,
@@ -761,7 +765,12 @@ 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"
+                    "of asynchronous"
+                )
 
     def _cat_common(self, path, start=None, end=None):
         path = self._strip_protocol(path)
@@ -772,6 +781,8 @@ class ReferenceFileSystem(AsyncFileSystem):
             raise FileNotFoundError(path) from exc
         if isinstance(part, str):
             part = part.encode()
+        if hasattr(part, "to_bytes"):
+            part = part.to_bytes()
         if isinstance(part, bytes):
             logger.debug(f"Reference: {path}, type bytes")
             if part.startswith(b"base64:"):
@@ -1073,7 +1084,7 @@ class ReferenceFileSystem(AsyncFileSystem):
         self.dircache = {"": []}
         it = self.references.items()
         for path, part in it:
-            if isinstance(part, (bytes, str)):
+            if isinstance(part, (bytes, str)) or hasattr(part, "to_bytes"):
                 size = len(part)
             elif len(part) == 1:
                 size = None
@@ -1100,10 +1111,33 @@ class ReferenceFileSystem(AsyncFileSystem):
             self.dircache[par].append({"name": path, "type": "file", "size": size})
 
     def _open(self, path, mode="rb", block_size=None, cache_options=None, **kwargs):
-        data = self.cat_file(path)  # load whole chunk into memory
-        return io.BytesIO(data)
+        part_or_url, start0, end0 = self._cat_common(path)
+        # This logic is kept outside `ReferenceFile` to avoid unnecessary redirection.
+        # That does mean `_cat_common` gets called twice if it eventually reaches `ReferenceFile`.
+        if isinstance(part_or_url, bytes):
+            return io.BytesIO(part_or_url[start0:end0])
+
+        protocol, _ = split_protocol(part_or_url)
+        if start0 is None and end0 is None:
+            return self.fss[protocol]._open(
+                part_or_url,
+                mode,
+                block_size=block_size,
+                cache_options=cache_options,
+                **kwargs,
+            )
+
+        return ReferenceFile(
+            self,
+            path,
+            mode,
+            block_size=block_size,
+            cache_options=cache_options,
+            **kwargs,
+        )
 
     def ls(self, path, detail=True, **kwargs):
+        logger.debug("list %s", path)
         path = self._strip_protocol(path)
         if isinstance(self.references, LazyReferenceMapper):
             try:
@@ -1214,3 +1248,58 @@ class ReferenceFileSystem(AsyncFileSystem):
                 out[k] = v
         with fsspec.open(url, "wb", **storage_options) as f:
             f.write(json.dumps({"version": 1, "refs": out}).encode())
+
+
+class ReferenceFile(AbstractBufferedFile):
+    def __init__(
+        self,
+        fs,
+        path,
+        mode="rb",
+        block_size="default",
+        autocommit=True,
+        cache_type="readahead",
+        cache_options=None,
+        size=None,
+        **kwargs,
+    ):
+        super().__init__(
+            fs,
+            path,
+            mode=mode,
+            block_size=block_size,
+            autocommit=autocommit,
+            size=size,
+            cache_type=cache_type,
+            cache_options=cache_options,
+            **kwargs,
+        )
+        part_or_url, self.start, self.end = self.fs._cat_common(self.path)
+        protocol, _ = split_protocol(part_or_url)
+        self.src_fs = self.fs.fss[protocol]
+        self.src_path = part_or_url
+        self._f = None
+
+    @property
+    def f(self):
+        if self._f is None or self._f.closed:
+            self._f = self.src_fs._open(
+                self.src_path,
+                mode=self.mode,
+                block_size=self.blocksize,
+                autocommit=self.autocommit,
+                cache_type="none",
+                **self.kwargs,
+            )
+        return self._f
+
+    def close(self):
+        if self._f is not None:
+            self._f.close()
+        return super().close()
+
+    def _fetch_range(self, start, end):
+        start = start + self.start
+        end = min(end + self.start, self.end)
+        self.f.seek(start)
+        return self.f.read(end - start)

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",
@@ -218,9 +221,9 @@ known_implementations = {
     "zip": {"class": "fsspec.implementations.zip.ZipFileSystem"},
 }
 
-assert list(known_implementations) == sorted(
-    known_implementations
-), "Not in alphabetical order"
+assert list(known_implementations) == sorted(known_implementations), (
+    "Not in alphabetical order"
+)
 
 
 def get_filesystem_class(protocol):

fsspec/spec.py

@@ -382,7 +382,7 @@ class AbstractFileSystem(metaclass=_Cached):
             pass
 
     def walk(self, path, maxdepth=None, topdown=True, on_error="omit", **kwargs):
-        """Return all files belows path
+        """Return all files under the given path.
 
         List all files, recursing into subdirectories; output is iterator-style,
         like ``os.walk()``. For a simple list of files, ``find()`` is available.
@@ -2131,7 +2131,7 @@ class AbstractBufferedFile(io.IOBase):
         return b"".join(out)
 
     def readline(self):
-        """Read until first occurrence of newline character
+        """Read until and including the first occurrence of newline character
 
         Note that, because of character encoding, this is not necessarily a
         true line ending.
@@ -2148,7 +2148,7 @@ class AbstractBufferedFile(io.IOBase):
         return self
 
     def readlines(self):
-        """Return all data, split by the newline character"""
+        """Return all data, split by the newline character, including the newline character"""
         data = self.read()
         lines = data.split(b"\n")
         out = [l + b"\n" for l in lines[:-1]]

{fsspec-2024.12.0.dist-info → fsspec-2025.3.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fsspec
-Version: 2024.12.0
+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/
@@ -131,7 +131,6 @@ Requires-Dist: pytest-rerunfailures; extra == 'test'
 Requires-Dist: requests; extra == 'test'
 Provides-Extra: test-downstream
 Requires-Dist: aiobotocore<3.0.0,>=2.5.4; extra == 'test-downstream'
-Requires-Dist: dask-expr; extra == 'test-downstream'
 Requires-Dist: dask[dataframe,test]; extra == 'test-downstream'
 Requires-Dist: moto[server]<5,>4; extra == 'test-downstream'
 Requires-Dist: pytest-timeout; extra == 'test-downstream'

{fsspec-2024.12.0.dist-info → fsspec-2025.3.0.dist-info}/RECORD

@@ -1,13 +1,13 @@
 fsspec/__init__.py,sha256=l9MJaNNV2d4wKpCtMvXDr55n92DkdrAayGy3F9ICjzk,1998
-fsspec/_version.py,sha256=wQ2VhCCZZrkDgAic2RGrr4PbmFQdFL4PTIWVnK3r9tM,419
-fsspec/archive.py,sha256=S__DzfZj-urAN3tp2W6jJ6YDiXG1fAl7FjvWUN73qIE,2386
+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
 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=A2r3PiZd17192sGHLwWNFbK8RFiDA7gSbfboIJ07wTY,11471
-fsspec/spec.py,sha256=d_NY5YVuwV7YCRduKkaR_z8B9GUna4-H9mOinymEMFY,75971
+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=cXfSkF2AaboInIIA_6jmB796RP_BXd8u08loPAHQsxQ,2864
+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=a0eNjLxyfFK7pbEa52U8K-PhNHukzdGVx1eLcVniaXY,15092
-fsspec/implementations/dirfs.py,sha256=ymakitNNQ07tW76EShyw3rC9RvIDHl4gtuOhE_h1vUg,12032
+fsspec/implementations/dbfs.py,sha256=XwpotuS_ncz3XK1dkUteww9GnTja7HoY91c0m4GUfwI,15092
+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=E-XYtnsHpE1e4x2io1ILOioGsWpCLDH1bqYN3QPGUJI,45930
+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-2024.12.0.dist-info/METADATA,sha256=bZRSxMVT7oHOUheJQEDMsChBFuby1pTg8WUbnLYATrc,11801
-fsspec-2024.12.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-fsspec-2024.12.0.dist-info/licenses/LICENSE,sha256=LcNUls5TpzB5FcAIqESq1T53K0mzTN0ARFBnaRQH7JQ,1513
-fsspec-2024.12.0.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,,