# This file is part of h5py, a Python interface to the HDF5 library. # # http://www.h5py.org # # Copyright 2008-2013 Andrew Collette and contributors # # License: Standard 3-clause BSD; see "license.txt" for full license terms # and contributor agreement. """ Implements high-level support for HDF5 file objects. """ import sys import os from warnings import warn from .compat import filename_decode, filename_encode from .base import phil, with_phil from .group import Group from .. import h5, h5f, h5p, h5i, h5fd, _objects from .. import version mpi = h5.get_config().mpi ros3 = h5.get_config().ros3 direct_vfd = h5.get_config().direct_vfd hdf5_version = version.hdf5_version_tuple[0:3] swmr_support = False if hdf5_version >= h5.get_config().swmr_min_hdf5_version: swmr_support = True libver_dict = {'earliest': h5f.LIBVER_EARLIEST, 'latest': h5f.LIBVER_LATEST} libver_dict_r = dict((y, x) for x, y in libver_dict.items()) if hdf5_version >= (1, 10, 2): libver_dict.update({'v108': h5f.LIBVER_V18, 'v110': h5f.LIBVER_V110}) libver_dict_r.update({h5f.LIBVER_V18: 'v108', h5f.LIBVER_V110: 'v110'}) if hdf5_version >= (1, 11, 4): libver_dict.update({'v112': h5f.LIBVER_V112}) libver_dict_r.update({h5f.LIBVER_V112: 'v112'}) if hdf5_version >= (1, 13, 0): libver_dict.update({'v114': h5f.LIBVER_V114}) libver_dict_r.update({h5f.LIBVER_V114: 'v114'}) def _set_fapl_mpio(plist, **kwargs): """Set file access property list for mpio driver""" if not mpi: raise ValueError("h5py was built without MPI support, can't use mpio driver") import mpi4py.MPI kwargs.setdefault('info', mpi4py.MPI.Info()) plist.set_fapl_mpio(**kwargs) def _set_fapl_fileobj(plist, **kwargs): """Set the Python file object driver in a file access property list""" plist.set_fileobj_driver(h5fd.fileobj_driver, kwargs.get('fileobj')) _drivers = { 'sec2': lambda plist, **kwargs: plist.set_fapl_sec2(**kwargs), 'stdio': lambda plist, **kwargs: plist.set_fapl_stdio(**kwargs), 'core': lambda plist, **kwargs: plist.set_fapl_core(**kwargs), 'family': lambda plist, **kwargs: plist.set_fapl_family( memb_fapl=plist.copy(), **kwargs ), 'mpio': _set_fapl_mpio, 'fileobj': _set_fapl_fileobj, 'split': lambda plist, **kwargs: plist.set_fapl_split(**kwargs), } if ros3: _drivers['ros3'] = lambda plist, **kwargs: plist.set_fapl_ros3(**kwargs) if direct_vfd: _drivers['direct'] = lambda plist, **kwargs: plist.set_fapl_direct(**kwargs) # noqa def register_driver(name, set_fapl): """Register a custom driver. Parameters ---------- name : str The name of the driver. set_fapl : callable[PropFAID, **kwargs] -> NoneType The function to set the fapl to use your custom driver. """ _drivers[name] = set_fapl def unregister_driver(name): """Unregister a custom driver. Parameters ---------- name : str The name of the driver. """ del _drivers[name] def registered_drivers(): """Return a frozenset of the names of all of the registered drivers. """ return frozenset(_drivers) def make_fapl(driver, libver, rdcc_nslots, rdcc_nbytes, rdcc_w0, locking, page_buf_size, min_meta_keep, min_raw_keep, alignment_threshold, alignment_interval, meta_block_size, **kwds): """ Set up a file access property list """ plist = h5p.create(h5p.FILE_ACCESS) if libver is not None: if libver in libver_dict: low = libver_dict[libver] high = h5f.LIBVER_LATEST else: low, high = (libver_dict[x] for x in libver) else: # we default to earliest low, high = h5f.LIBVER_EARLIEST, h5f.LIBVER_LATEST plist.set_libver_bounds(low, high) plist.set_alignment(alignment_threshold, alignment_interval) cache_settings = list(plist.get_cache()) if rdcc_nslots is not None: cache_settings[1] = rdcc_nslots if rdcc_nbytes is not None: cache_settings[2] = rdcc_nbytes if rdcc_w0 is not None: cache_settings[3] = rdcc_w0 plist.set_cache(*cache_settings) if page_buf_size: plist.set_page_buffer_size(int(page_buf_size), int(min_meta_keep), int(min_raw_keep)) if meta_block_size is not None: plist.set_meta_block_size(int(meta_block_size)) if locking is not None: if hdf5_version < (1, 12, 1) and (hdf5_version[:2] != (1, 10) or hdf5_version[2] < 7): raise ValueError( "HDF5 version >= 1.12.1 or 1.10.x >= 1.10.7 required for file locking.") if locking in ("false", False): plist.set_file_locking(False, ignore_when_disabled=False) elif locking in ("true", True): plist.set_file_locking(True, ignore_when_disabled=False) elif locking == "best-effort": plist.set_file_locking(True, ignore_when_disabled=True) else: raise ValueError(f"Unsupported locking value: {locking}") if driver is None or (driver == 'windows' and sys.platform == 'win32'): # Prevent swallowing unused key arguments if kwds: msg = "'{key}' is an invalid keyword argument for this function" \ .format(key=next(iter(kwds))) raise TypeError(msg) return plist try: set_fapl = _drivers[driver] except KeyError: raise ValueError('Unknown driver type "%s"' % driver) else: set_fapl(plist, **kwds) return plist def make_fcpl(track_order=False, fs_strategy=None, fs_persist=False, fs_threshold=1, fs_page_size=None): """ Set up a file creation property list """ if track_order or fs_strategy: plist = h5p.create(h5p.FILE_CREATE) if track_order: plist.set_link_creation_order( h5p.CRT_ORDER_TRACKED | h5p.CRT_ORDER_INDEXED) plist.set_attr_creation_order( h5p.CRT_ORDER_TRACKED | h5p.CRT_ORDER_INDEXED) if fs_strategy: strategies = { 'fsm': h5f.FSPACE_STRATEGY_FSM_AGGR, 'page': h5f.FSPACE_STRATEGY_PAGE, 'aggregate': h5f.FSPACE_STRATEGY_AGGR, 'none': h5f.FSPACE_STRATEGY_NONE } fs_strat_num = strategies.get(fs_strategy, -1) if fs_strat_num == -1: raise ValueError("Invalid file space strategy type") plist.set_file_space_strategy(fs_strat_num, fs_persist, fs_threshold) if fs_page_size and fs_strategy == 'page': plist.set_file_space_page_size(int(fs_page_size)) else: plist = None return plist def make_fid(name, mode, userblock_size, fapl, fcpl=None, swmr=False): """ Get a new FileID by opening or creating a file. Also validates mode argument.""" if userblock_size is not None: if mode in ('r', 'r+'): raise ValueError("User block may only be specified " "when creating a file") try: userblock_size = int(userblock_size) except (TypeError, ValueError): raise ValueError("User block size must be an integer") if fcpl is None: fcpl = h5p.create(h5p.FILE_CREATE) fcpl.set_userblock(userblock_size) if mode == 'r': flags = h5f.ACC_RDONLY if swmr and swmr_support: flags |= h5f.ACC_SWMR_READ fid = h5f.open(name, flags, fapl=fapl) elif mode == 'r+': fid = h5f.open(name, h5f.ACC_RDWR, fapl=fapl) elif mode in ['w-', 'x']: fid = h5f.create(name, h5f.ACC_EXCL, fapl=fapl, fcpl=fcpl) elif mode == 'w': fid = h5f.create(name, h5f.ACC_TRUNC, fapl=fapl, fcpl=fcpl) elif mode == 'a': # Open in append mode (read/write). # If that fails, create a new file only if it won't clobber an # existing one (ACC_EXCL) try: fid = h5f.open(name, h5f.ACC_RDWR, fapl=fapl) # Not all drivers raise FileNotFoundError (commented those that do not) except FileNotFoundError if fapl.get_driver() in ( h5fd.SEC2, h5fd.DIRECT if direct_vfd else -1, # h5fd.STDIO, # h5fd.CORE, h5fd.FAMILY, h5fd.WINDOWS, # h5fd.MPIO, # h5fd.MPIPOSIX, h5fd.fileobj_driver, h5fd.ROS3D if ros3 else -1, ) else OSError: fid = h5f.create(name, h5f.ACC_EXCL, fapl=fapl, fcpl=fcpl) else: raise ValueError("Invalid mode; must be one of r, r+, w, w-, x, a") try: if userblock_size is not None: existing_fcpl = fid.get_create_plist() if existing_fcpl.get_userblock() != userblock_size: raise ValueError("Requested userblock size (%d) does not match that of existing file (%d)" % (userblock_size, existing_fcpl.get_userblock())) except Exception as e: fid.close() raise e return fid class File(Group): """ Represents an HDF5 file. """ @property def attrs(self): """ Attributes attached to this object """ # hdf5 complains that a file identifier is an invalid location for an # attribute. Instead of self, pass the root group to AttributeManager: from . import attrs with phil: return attrs.AttributeManager(self['/']) @property @with_phil def filename(self): """File name on disk""" return filename_decode(h5f.get_name(self.id)) @property @with_phil def driver(self): """Low-level HDF5 file driver used to open file""" drivers = {h5fd.SEC2: 'sec2', h5fd.STDIO: 'stdio', h5fd.CORE: 'core', h5fd.FAMILY: 'family', h5fd.WINDOWS: 'windows', h5fd.MPIO: 'mpio', h5fd.MPIPOSIX: 'mpiposix', h5fd.fileobj_driver: 'fileobj'} if ros3: drivers[h5fd.ROS3D] = 'ros3' if direct_vfd: drivers[h5fd.DIRECT] = 'direct' return drivers.get(self.id.get_access_plist().get_driver(), 'unknown') @property @with_phil def mode(self): """ Python mode used to open file """ write_intent = h5f.ACC_RDWR if swmr_support: write_intent |= h5f.ACC_SWMR_WRITE return 'r+' if self.id.get_intent() & write_intent else 'r' @property @with_phil def libver(self): """File format version bounds (2-tuple: low, high)""" bounds = self.id.get_access_plist().get_libver_bounds() return tuple(libver_dict_r[x] for x in bounds) @property @with_phil def userblock_size(self): """ User block size (in bytes) """ fcpl = self.id.get_create_plist() return fcpl.get_userblock() @property @with_phil def meta_block_size(self): """ Meta block size (in bytes) """ fapl = self.id.get_access_plist() return fapl.get_meta_block_size() if mpi and hdf5_version >= (1, 8, 9): @property @with_phil def atomic(self): """ Set/get MPI-IO atomic mode """ return self.id.get_mpi_atomicity() @atomic.setter @with_phil def atomic(self, value): # pylint: disable=missing-docstring self.id.set_mpi_atomicity(value) @property @with_phil def swmr_mode(self): """ Controls single-writer multiple-reader mode """ return swmr_support and bool(self.id.get_intent() & (h5f.ACC_SWMR_READ | h5f.ACC_SWMR_WRITE)) @swmr_mode.setter @with_phil def swmr_mode(self, value): # pylint: disable=missing-docstring if swmr_support: if value: self.id.start_swmr_write() else: raise ValueError("It is not possible to forcibly switch SWMR mode off.") else: raise RuntimeError('SWMR support is not available in HDF5 version {}.{}.{}.'.format(*hdf5_version)) def __init__(self, name, mode='r', driver=None, libver=None, userblock_size=None, swmr=False, rdcc_nslots=None, rdcc_nbytes=None, rdcc_w0=None, track_order=None, fs_strategy=None, fs_persist=False, fs_threshold=1, fs_page_size=None, page_buf_size=None, min_meta_keep=0, min_raw_keep=0, locking=None, alignment_threshold=1, alignment_interval=1, meta_block_size=None, **kwds): """Create a new file object. See the h5py user guide for a detailed explanation of the options. name Name of the file on disk, or file-like object. Note: for files created with the 'core' driver, HDF5 still requires this be non-empty. mode r Readonly, file must exist (default) r+ Read/write, file must exist w Create file, truncate if exists w- or x Create file, fail if exists a Read/write if exists, create otherwise driver Name of the driver to use. Legal values are None (default, recommended), 'core', 'sec2', 'direct', 'stdio', 'mpio', 'ros3'. libver Library version bounds. Supported values: 'earliest', 'v108', 'v110', 'v112' and 'latest'. The 'v108', 'v110' and 'v112' options can only be specified with the HDF5 1.10.2 library or later. userblock_size Desired size of user block. Only allowed when creating a new file (mode w, w- or x). swmr Open the file in SWMR read mode. Only used when mode = 'r'. rdcc_nbytes Total size of the dataset chunk cache in bytes. The default size is 1024**2 (1 MiB) per dataset. Applies to all datasets unless individually changed. rdcc_w0 The chunk preemption policy for all datasets. This must be between 0 and 1 inclusive and indicates the weighting according to which chunks which have been fully read or written are penalized when determining which chunks to flush from cache. A value of 0 means fully read or written chunks are treated no differently than other chunks (the preemption is strictly LRU) while a value of 1 means fully read or written chunks are always preempted before other chunks. If your application only reads or writes data once, this can be safely set to 1. Otherwise, this should be set lower depending on how often you re-read or re-write the same data. The default value is 0.75. Applies to all datasets unless individually changed. rdcc_nslots The number of chunk slots in the raw data chunk cache for this file. Increasing this value reduces the number of cache collisions, but slightly increases the memory used. Due to the hashing strategy, this value should ideally be a prime number. As a rule of thumb, this value should be at least 10 times the number of chunks that can fit in rdcc_nbytes bytes. For maximum performance, this value should be set approximately 100 times that number of chunks. The default value is 521. Applies to all datasets unless individually changed. track_order Track dataset/group/attribute creation order under root group if True. If None use global default h5.get_config().track_order. fs_strategy The file space handling strategy to be used. Only allowed when creating a new file (mode w, w- or x). Defined as: "fsm" FSM, Aggregators, VFD "page" Paged FSM, VFD "aggregate" Aggregators, VFD "none" VFD If None use HDF5 defaults. fs_page_size File space page size in bytes. Only used when fs_strategy="page". If None use the HDF5 default (4096 bytes). fs_persist A boolean value to indicate whether free space should be persistent or not. Only allowed when creating a new file. The default value is False. fs_threshold The smallest free-space section size that the free space manager will track. Only allowed when creating a new file. The default value is 1. page_buf_size Page buffer size in bytes. Only allowed for HDF5 files created with fs_strategy="page". Must be a power of two value and greater or equal than the file space page size when creating the file. It is not used by default. min_meta_keep Minimum percentage of metadata to keep in the page buffer before allowing pages containing metadata to be evicted. Applicable only if page_buf_size is set. Default value is zero. min_raw_keep Minimum percentage of raw data to keep in the page buffer before allowing pages containing raw data to be evicted. Applicable only if page_buf_size is set. Default value is zero. locking The file locking behavior. Defined as: - False (or "false") -- Disable file locking - True (or "true") -- Enable file locking - "best-effort" -- Enable file locking but ignore some errors - None -- Use HDF5 defaults .. warning:: The HDF5_USE_FILE_LOCKING environment variable can override this parameter. Only available with HDF5 >= 1.12.1 or 1.10.x >= 1.10.7. alignment_threshold Together with ``alignment_interval``, this property ensures that any file object greater than or equal in size to the alignement threshold (in bytes) will be aligned on an address which is a multiple of alignment interval. alignment_interval This property should be used in conjunction with ``alignment_threshold``. See the description above. For more details, see https://portal.hdfgroup.org/display/HDF5/H5P_SET_ALIGNMENT meta_block_size Set the current minimum size, in bytes, of new metadata block allocations. See https://portal.hdfgroup.org/display/HDF5/H5P_SET_META_BLOCK_SIZE Additional keywords Passed on to the selected file driver. """ if (fs_strategy or page_buf_size) and hdf5_version < (1, 10, 1): raise ValueError("HDF5 version 1.10.1 or greater required for file space strategy or page buffering support.") if swmr and not swmr_support: raise ValueError("The SWMR feature is not available in this version of the HDF5 library") if driver == 'ros3': if ros3: from urllib.parse import urlparse url = urlparse(name) if url.scheme == 's3': aws_region = kwds.get('aws_region', b'').decode('ascii') if len(aws_region) == 0: raise ValueError('AWS region required for s3:// location') name = f'https://s3.{aws_region}.amazonaws.com/{url.netloc}{url.path}' elif url.scheme not in ('https', 'http'): raise ValueError(f'{name}: S3 location must begin with ' 'either "https://", "http://", or "s3://"') else: raise ValueError( "h5py was built without ROS3 support, can't use ros3 driver") if locking is not None and hdf5_version < (1, 12, 1) and ( hdf5_version[:2] != (1, 10) or hdf5_version[2] < 7): raise ValueError("HDF5 version >= 1.12.1 or 1.10.x >= 1.10.7 required for file locking options.") if isinstance(name, _objects.ObjectID): if fs_strategy: raise ValueError("Unable to set file space strategy of an existing file") with phil: fid = h5i.get_file_id(name) else: if hasattr(name, 'read') and hasattr(name, 'seek'): if driver not in (None, 'fileobj'): raise ValueError("Driver must be 'fileobj' for file-like object if specified.") driver = 'fileobj' if kwds.get('fileobj', name) != name: raise ValueError("Invalid value of 'fileobj' argument; " "must equal to file-like object if specified.") kwds.update(fileobj=name) name = repr(name).encode('ASCII', 'replace') else: name = filename_encode(name) if track_order is None: track_order = h5.get_config().track_order if fs_strategy and mode not in ('w', 'w-', 'x'): raise ValueError("Unable to set file space strategy of an existing file") if swmr and mode != 'r': warn( "swmr=True only affects read ('r') mode. For swmr write " "mode, set f.swmr_mode = True after opening the file.", stacklevel=2, ) with phil: fapl = make_fapl(driver, libver, rdcc_nslots, rdcc_nbytes, rdcc_w0, locking, page_buf_size, min_meta_keep, min_raw_keep, alignment_threshold=alignment_threshold, alignment_interval=alignment_interval, meta_block_size=meta_block_size, **kwds) fcpl = make_fcpl(track_order=track_order, fs_strategy=fs_strategy, fs_persist=fs_persist, fs_threshold=fs_threshold, fs_page_size=fs_page_size) fid = make_fid(name, mode, userblock_size, fapl, fcpl, swmr=swmr) if isinstance(libver, tuple): self._libver = libver else: self._libver = (libver, 'latest') super().__init__(fid) def close(self): """ Close the file. All open objects become invalid """ with phil: # Check that the file is still open, otherwise skip if self.id.valid: # We have to explicitly murder all open objects related to the file # Close file-resident objects first, then the files. # Otherwise we get errors in MPI mode. self.id._close_open_objects(h5f.OBJ_LOCAL | ~h5f.OBJ_FILE) self.id._close_open_objects(h5f.OBJ_LOCAL | h5f.OBJ_FILE) self.id.close() _objects.nonlocal_close() def flush(self): """ Tell the HDF5 library to flush its buffers. """ with phil: h5f.flush(self.id) @with_phil def __enter__(self): return self @with_phil def __exit__(self, *args): if self.id: self.close() @with_phil def __repr__(self): if not self.id: r = '' else: # Filename has to be forced to Unicode if it comes back bytes # Mode is always a "native" string filename = self.filename if isinstance(filename, bytes): # Can't decode fname filename = filename.decode('utf8', 'replace') r = f'' return r