h5json 2.0.0__py3-none-any.whl

h5json/__init__.py

@@ -0,0 +1,41 @@
+##############################################################################
+# Copyright by The HDF Group.                                                #
+# All rights reserved.                                                       #
+#                                                                            #
+# This file is part of H5Serv (HDF5 REST Server) Service, Libraries and      #
+# Utilities.  The full HDF5 REST Server copyright notice, including          #
+# terms governing use, modification, and redistribution, is contained in     #
+# the file COPYING, which can be found at the root of the source code        #
+# distribution tree.  If you do not have access to this file, you may        #
+# request a copy from help@hdfgroup.org.                                     #
+##############################################################################
+
+
+"""
+This is the h5json package, a mapping between HDF5 objects and JSON
+"""
+
+from __future__ import absolute_import
+
+from importlib.metadata import PackageNotFoundError as _PackageNotFoundError
+from importlib.metadata import version as _pkg_version
+
+from .hdf5dtype import getTypeItem
+from .hdf5dtype import getTypeResponse
+from .hdf5dtype import getItemSize
+from .hdf5dtype import createDataType
+from .objid import createObjId
+from .objid import getCollectionForId
+from .objid import isObjId
+from .objid import isS3ObjKey
+from .objid import getS3Key
+from .objid import getObjId
+from .objid import isSchema2Id
+from .objid import isRootObjId
+from .hdf5db import Hdf5db
+
+try:
+    __version__ = _pkg_version("h5json")
+except _PackageNotFoundError:
+    # running from a source tree that has not been installed
+    __version__ = "0.0.0.dev0"

h5json/apiversion.py

@@ -0,0 +1,15 @@
+##############################################################################
+# Copyright by The HDF Group.                                                #
+# All rights reserved.                                                       #
+#                                                                            #
+# This file is part of H5Serv (HDF5 REST Server) Service, Libraries and      #
+# Utilities.  The full HDF5 REST Server copyright notice, including          #
+# terms governing use, modification, and redistribution, is contained in     #
+# the file COPYING, which can be found at the root of the source code        #
+# distribution tree.  If you do not have access to this file, you may        #
+# request a copy from help@hdfgroup.org.                                     #
+##############################################################################
+
+# IMPORTANT: HDF5/JSON apiVersion key value is set here. Update only when the
+# HDF5/JSON spec changes.
+_apiver = "1.1.1"

h5json/apps/__init__.py

@@ -0,0 +1,11 @@
+##############################################################################
+# Copyright by The HDF Group.                                                #
+# All rights reserved.                                                       #
+#                                                                            #
+# This file is part of H5Serv (HDF5 REST Server) Service, Libraries and      #
+# Utilities.  The full HDF5 REST Server copyright notice, including          #
+# terms governing use, modification, and redistribution, is contained in     #
+# the file COPYING, which can be found at the root of the source code        #
+# distribution tree.  If you do not have access to this file, you may        #
+# request a copy from help@hdfgroup.org.                                     #
+##############################################################################

h5json/apps/h5tojson.py

@@ -0,0 +1,70 @@
+##############################################################################
+# Copyright by The HDF Group.                                                #
+# All rights reserved.                                                       #
+#                                                                            #
+# This file is part of H5Serv (HDF5 REST Server) Service, Libraries and      #
+# Utilities.  The full HDF5 REST Server copyright notice, including          #
+# terms governing use, modification, and redistribution, is contained in     #
+# the file COPYING, which can be found at the root of the source code        #
+# distribution tree.  If you do not have access to this file, you may        #
+# request a copy from help@hdfgroup.org.                                     #
+##############################################################################
+import sys
+import os.path as op
+import logging
+
+from h5json import Hdf5db
+from h5json.jsonstore.h5json_plugin import H5JsonPlugin
+from h5json.h5pystore.h5py_plugin import H5pyPlugin
+
+
+def main():
+    if len(sys.argv) < 2 or sys.argv[1] in ("-h", "--help"):
+        print(f"usage: {sys.argv[0]} [-h] [--nodata] [--data-limit n] <hdf5_file>")
+        sys.exit(0)
+
+    data_limit = None
+    filename = None
+    for i in range(1, len(sys.argv)):
+        if sys.argv[i] == "--nodata":
+            data_limit = 0
+        elif sys.argv[i] == "--data-limit":
+            i += 1
+            if i >= len(sys.argv):
+                sys.exit("Error: --data-limit requires a numeric argument")
+            try:
+                data_limit = int(sys.argv[i])
+            except ValueError:
+                sys.exit("Error: --data-limit requires a numeric argument")
+        else:
+            filename = sys.argv[i]
+
+    # create logger
+    logfname = "h5tojson.log"
+    loglevel = logging.DEBUG
+    logging.basicConfig(filename=logfname, format='%(levelname)s %(asctime)s %(message)s', level=loglevel)
+    log = logging.getLogger()
+
+    # check that the input file exists
+    if not op.isfile(filename):
+        sys.exit(f"Cannot find file: {filename}")
+
+    log.info(f"h5tojson {filename}")
+
+    # read_only=True: open the source file in h5py mode='r' - src_db never
+    # creates/modifies anything, and read_only guarantees that even if it
+    # somehow did, nothing could actually be written back to the source file
+    src_db = Hdf5db(plugin=H5pyPlugin(filename, read_only=True, app_logger=log), app_logger=log)
+    src_db.open()  # read HDF5 data into src_db
+
+    dst_db = Hdf5db(plugin=H5JsonPlugin(None, data_limit=data_limit, app_logger=log), app_logger=log)
+    dst_db.open()
+
+    src_db.copy(dst_db)  # write src_db's content into dst_db
+
+    dst_db.close()  # triggers write to json file (stdout, since filepath is None)
+    src_db.close()
+
+
+if __name__ == "__main__":
+    main()

h5json/apps/jsontoh5.py

@@ -0,0 +1,65 @@
+##############################################################################
+# Copyright by The HDF Group.                                                #
+# All rights reserved.                                                       #
+#                                                                            #
+# This file is part of H5Serv (HDF5 REST Server) Service, Libraries and      #
+# Utilities.  The full HDF5 REST Server copyright notice, including          #
+# terms governing use, modification, and redistribution, is contained in     #
+# the file COPYING, which can be found at the root of the source code        #
+# distribution tree.  If you do not have access to this file, you may        #
+# request a copy from help@hdfgroup.org.                                     #
+##############################################################################
+import sys
+import os.path as op
+import logging
+
+from h5json import Hdf5db
+from h5json.h5pystore.h5py_plugin import H5pyPlugin
+from h5json.jsonstore.h5json_plugin import H5JsonPlugin
+
+
+def main():
+    if len(sys.argv) < 3 or sys.argv[1] in ("-h", "--help"):
+        print(f"usage: {sys.argv[0]} [-h] [--nodata] <json_file> <h5_file>")
+        sys.exit(0)
+
+    no_data = False
+    json_filename = None
+    hdf5_filename = None
+    for i in range(1, len(sys.argv)):
+        if sys.argv[i] == "--nodata":
+            no_data = True
+        elif not json_filename:
+            json_filename = sys.argv[i]
+        else:
+            hdf5_filename = sys.argv[i]
+
+    # create logger
+    logfname = "jsontoh5.log"
+    loglevel = logging.DEBUG
+    logging.basicConfig(filename=logfname, format='%(levelname)s %(asctime)s %(message)s', level=loglevel)
+    log = logging.getLogger()
+
+    # check that the input file exists
+    if not op.isfile(json_filename):
+        sys.exit(f"Cannot find file: {json_filename}")
+
+    log.info(f"jsontoh5 {json_filename} to {hdf5_filename}")
+
+    # read_only=True: src_db never creates/modifies anything, and read_only
+    # guarantees flush() can never write back to json_filename even if it
+    # somehow did (append alone would still permit a write)
+    src_db = Hdf5db(plugin=H5JsonPlugin(json_filename, read_only=True, app_logger=log), app_logger=log)
+    src_db.open()  # read json data into src_db
+
+    dst_db = Hdf5db(plugin=H5pyPlugin(hdf5_filename, no_data=no_data, app_logger=log), app_logger=log)
+    dst_db.open()
+
+    src_db.copy(dst_db)  # write everything src_db read to the output file
+
+    dst_db.close()
+    src_db.close()
+
+
+if __name__ == "__main__":
+    main()

h5json/apps/validator.py

@@ -0,0 +1,101 @@
+##############################################################################
+# Copyright by The HDF Group.                                                #
+# All rights reserved.                                                       #
+#                                                                            #
+# This file is part of h5json. The full copyright notice, including          #
+# terms governing use, modification, and redistribution, is contained in     #
+# the file COPYING, which can be found at the root of the source code        #
+# distribution tree. If you do not have access to this file, you may         #
+# request a copy from help@hdfgroup.org.                                     #
+##############################################################################
+import sys
+import argparse
+from pathlib import Path
+import json
+import jsonschema
+from h5json import schema
+try:
+    import importlib_resources as ilr
+except ImportError:
+    import importlib.resources as ilr
+
+
+def prepare_validator() -> jsonschema.Draft202012Validator:
+    """Return a configured jsonschema.Draft202012Validator instance."""
+    with ilr.open_text(schema, "hdf5.schema.json") as f:
+        h5schema = json.load(f)
+
+    schema_store = dict()
+    schema_components = [
+        "attribute.schema.json",
+        "filters.schema.json",
+        "group.schema.json",
+        "datatypes.schema.json",
+        "dataspaces.schema.json",
+        "dataset.schema.json",
+    ]
+    for sc in schema_components:
+        with ilr.open_text(schema, sc) as f:
+            temp = json.load(f)
+        schema_store[temp["$id"]] = temp
+    resolver = jsonschema.RefResolver(h5schema["$id"], h5schema, store=schema_store)
+    return jsonschema.Draft202012Validator(h5schema, resolver=resolver)
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(
+        description="HDF5/JSON validator",
+        epilog="Copyright 2021 The HDF Group",
+        formatter_class=argparse.ArgumentDefaultsHelpFormatter,
+    )
+    parser.add_argument(
+        "jsonloc",
+        nargs="+",
+        help="JSON location (files or folders)",
+        metavar="JSON_LOC",
+        type=Path,
+    )
+    parser.add_argument(
+        "--stop",
+        "-s",
+        action="store_true",
+        help="Stop after first HDF5/JSON file failed validation",
+    )
+    args = parser.parse_args()
+
+    # Find all JSON files for validation...
+    json_files = list()
+    for p in args.jsonloc:
+        if p.is_file():
+            json_files.append(p)
+        elif p.is_dir():
+            json_files.extend([f for f in p.glob("*.json")])
+    if not json_files:
+        sys.exit("No JSON files for validation found.")
+
+    validator = prepare_validator()
+
+    # Validate HDF5/JSON files...
+    valid_errors = False
+    for h5j in json_files:
+        print(f"Validating {str(h5j)} ... ", end="")
+        try:
+            with h5j.open() as f:
+                inst = json.load(f)
+            validator.validate(inst)
+            print("pass")
+        except jsonschema.exceptions.ValidationError:
+            print("FAIL")
+            valid_errors = True
+            inst_name = str(h5j)
+            print(f"HDF5/JSON validation failed for {inst_name}", file=sys.stderr)
+            for err in validator.iter_errors(inst):
+                print(f"{inst_name} ---> {err}", file=sys.stderr)
+            if args.stop:
+                sys.exit("HDF5/JSON validation failed.")
+    if valid_errors:
+        sys.exit("HDF5/JSON validation failed.")
+
+
+if __name__ == "__main__":
+    main()