Revert "ENH: Directly open compressed files"

doc/source/examining/loading_data.rst

@@ -36,32 +36,6 @@ any arguments, and it will return a list of the names that can be supplied:
 
 This will return a list of possible filenames; more information can be accessed on the data catalog.
 
-
-.. _loading-archived-data:
-
-Archived Data
--------------
-
-If your data is stored as a (compressed) tar file, you can access the contained
-dataset directly without extracting the tar file.
-This can be achieved using the ``load_archive`` function:
-
-.. code-block:: python
-
-   import yt
-
-   ds = yt.load_archive("IsolatedGalaxy.tar.gz", "IsolatedGalaxy/galaxy0030/galaxy0030")
-
-The first argument is the path to the archive file, the second one is the path to the file to load
-in the archive. Subsequent arguments are passed to ``yt.load``.
-
-The functionality requires the package `ratarmount <https://github.com/mxmlnkn/ratarmount/>`_ to be installed.
-Under the hood, yt will mount the archive as a (read-only) filesystem. Note that this requires the
-entire archive to be read once to compute the location of each file in the archive; subsequent accesses
-will be much faster.
-All archive formats supported by `ratarmount <https://github.com/mxmlnkn/ratarmount>`_ should be loadable, provided
-the dependencies are installed; this includes ``tar``, ``tar.gz`` and tar.bz2`` formats.
-
 .. _loading-amrvac-data:
 
 AMRVAC Data

nose_unit.cfg

@@ -6,5 +6,5 @@ nologcapture=1
 verbosity=2
 where=yt
 with-timer=1
-ignore-files=(test_load_errors.py|test_load_sample.py|test_commons.py|test_ambiguous_fields.py|test_field_access_pytest.py|test_save.py|test_line_annotation_unit.py|test_eps_writer.py|test_registration.py|test_invalid_origin.py|test_outputs_pytest\.py|test_normal_plot_api\.py|test_load_archive\.py)
+ignore-files=(test_load_errors.py|test_load_sample.py|test_commons.py|test_ambiguous_fields.py|test_field_access_pytest.py|test_save.py|test_line_annotation_unit.py|test_eps_writer.py|test_registration.py|test_invalid_origin.py|test_outputs_pytest\.py|test_normal_plot_api\.py)
 exclude-test=yt.frontends.gdf.tests.test_outputs.TestGDF

setup.cfg

@@ -91,7 +91,6 @@ full =
     requests>=2.20.0
     scipy>=1.5.0
     xarray>=0.16.1
-    ratarmount~=0.8.1;platform_system!="Windows"
 mapserver =
     bottle
 minimal =

tests/ci_install.sh

@@ -10,14 +10,13 @@ linux|Linux)
       proj-data \
       proj-bin \
       libgeos-dev \
-      libopenmpi-dev \
-      libfuse2
+      libopenmpi-dev
     ;;
 osx|macOS)
     sudo mkdir -p /usr/local/man
     sudo chown -R "${USER}:admin" /usr/local/man
     brew update
-    HOMEBREW_NO_AUTO_UPDATE=1 brew install hdf5 proj geos open-mpi netcdf ccache osxfuse
+    HOMEBREW_NO_AUTO_UPDATE=1 brew install hdf5 proj geos open-mpi netcdf ccache
     ;;
 esac
 

tests/tests.yaml

@@ -193,7 +193,6 @@ other_tests:
      - "--ignore-files=test_save.py"
      - "--ignore-files=test_registration.py"
      - "--ignore-files=test_invalid_origin.py"
-     - "--ignore-files=test_load_archive\\.py"
      - "--ignore-files=test_outputs_pytest\\.py"
      - "--ignore-files=test_normal_plot_api\\.py"
      - "--exclude-test=yt.frontends.gdf.tests.test_outputs.TestGDF"

yt/__init__.py

@@ -80,7 +80,6 @@
 from yt.loaders import (
     load,
     load_amr_grids,
-    load_archive,
     load_hexahedral_mesh,
     load_octree,
     load_particles,

yt/loaders.py

@@ -2,28 +2,20 @@
 This module gathers all user-facing functions with a `load_` prefix.
 
 """
-import atexit
 import os
 import sys
 import tarfile
-import time
-import types
-import warnings
-from multiprocessing import Pipe, Process
-from multiprocessing.connection import Connection
 from pathlib import Path
-from typing import Dict, List, Optional, Tuple, Union
+from typing import List, Optional, Tuple
 from urllib.parse import urlsplit
 
 import numpy as np
 from more_itertools import always_iterable
 
-from yt.data_objects.static_output import Dataset
 from yt.funcs import levenshtein_distance
 from yt.sample_data.api import lookup_on_disk_data
 from yt.utilities.decompose import decompose_array, get_psize
 from yt.utilities.exceptions import (
-    MountError,
     YTAmbiguousDataType,
     YTIllDefinedAMR,
     YTSimulationNotIdentified,
@@ -36,7 +28,7 @@
     output_type_registry,
     simulation_time_series_registry,
 )
-from yt.utilities.on_demand_imports import _pooch as pooch, _ratarmount as ratarmount
+from yt.utilities.on_demand_imports import _pooch as pooch
 
 # --- Loaders for known data formats ---
 
@@ -1433,129 +1425,3 @@ def load_sample(
         loadable_path = loadable_path.joinpath(load_name, specific_file)
 
     return load(loadable_path, **kwargs)
-
-
-def _mount_helper(
-    archive: str, mountPoint: str, ratarmount_kwa: Dict, conn: Connection
-):
-    try:
-        fuseOperationsObject = ratarmount.TarMount(
-            pathToMount=archive,
-            mountPoint=mountPoint,
-            lazyMounting=True,
-            **ratarmount_kwa,
-        )
-        fuseOperationsObject.use_ns = True
-        conn.send(True)
-    except Exception:
-        conn.send(False)
-        raise
-
-    ratarmount.fuse.FUSE(
-        operations=fuseOperationsObject,
-        mountpoint=mountPoint,
-        foreground=True,
-        nothreads=True,
-    )
-
-
-# --- Loader for tar-based datasets ---
-def load_archive(
-    fn: Union[str, Path],
-    path: str,
-    ratarmount_kwa: Optional[Dict] = None,
-    mount_timeout: float = 1.0,
-    *args,
-    **kwargs,
-) -> Dataset:
-    r"""
-    Load archived data with yt.
-
-    This is a wrapper around :func:`~yt.loaders.load` to include mounting
-    and unmounting the archive as a read-only filesystem and load it.
-
-    Parameters
-    ----------
-
-    fn: str
-        The `filename` of the archive containing the dataset.
-
-    path: str
-        The path to the dataset in the archive.
-
-    ratarmount_kwa: dict, optional
-        Optional parameters to pass to ratarmount to mount the archive.
-
-    mount_timeout: float, optional
-        The timeout to wait for ratarmount to mount the archive. Default is 1s.
-
-    Notes
-    -----
-
-    - The function is experimental and may work or not depending on your setup.
-    - Any additional keyword argument is passed down to :func:`~yt.loaders.load`.
-    - This function requires ratarmount to be installed.
-    - This function does not work on Windows system.
-    """
-
-    warnings.warn(
-        "The 'load_archive' function is still experimental and may be unstable."
-    )
-
-    fn = os.path.expanduser(fn)
-
-    # This will raise FileNotFoundError if the path isn't matched
-    # either in the current dir or yt.config.ytcfg['data_dir_directory']
-    if not fn.startswith("http"):
-        fn = str(lookup_on_disk_data(fn))
-
-    if ratarmount_kwa is None:
-        ratarmount_kwa = {}
-
-    try:
-        tarfile.open(fn)
-    except tarfile.ReadError:
-        raise YTUnidentifiedDataType(fn, *args, **kwargs)
-
-    # Note: the temporary directory will be created by ratarmount
-    tempdir = fn + ".mount"
-    tempdir_base = tempdir
-    i = 0
-    while os.path.exists(tempdir):
-        i += 1
-        tempdir = f"{tempdir_base}.{i}"
-
-    parent_conn, child_conn = Pipe()
-    proc = Process(target=_mount_helper, args=(fn, tempdir, ratarmount_kwa, child_conn))
-    proc.start()
-    if not parent_conn.recv():
-        raise MountError(f"An error occured while mounting {fn} in {tempdir}")
-
-    # Note: the mounting needs to happen in another process which
-    # needs be run in the foreground (otherwise it may
-    # unmount). To prevent a race-condition here, we wait
-    # for the folder to be mounted within a reasonable time.
-    t = 0.0
-    while t < mount_timeout:
-        if os.path.ismount(tempdir):
-            break
-        time.sleep(0.1)
-        t += 0.1
-    else:
-        raise MountError(f"Folder {tempdir} does not appear to be mounted")
-
-    # We need to kill the process at exit (to force unmounting)
-    def umount_callback():
-        proc.terminate()
-
-    atexit.register(umount_callback)
-
-    # Alternatively, can dismount manually
-    def del_callback(self):
-        proc.terminate()
-        atexit.unregister(umount_callback)
-
-    ds = load(os.path.join(tempdir, path), *args, **kwargs)
-    ds.dismount = types.MethodType(del_callback, ds)
-
-    return ds

yt/utilities/exceptions.py

@@ -922,8 +922,3 @@ class GenerationInProgress(Exception):
     def __init__(self, fields):
         self.fields = fields
         super().__init__()
-
-
-class MountError(Exception):
-    def __init__(self, message):
-        self.message = message