chore: add h5py utils to agebd package

packages/agebd/pyproject.toml

@@ -5,6 +5,7 @@ description = "SLS HLA Framework"
 requires-python = "==3.10.*"
 
 dependencies = [
+    "h5py>=3.16.0",
     "typer>=0.23.2",
 ]
 

packages/agebd/src/agebd/h5.py

@@ -0,0 +1,113 @@
+import os
+from time import strftime
+
+import numpy as np
+from h5py import File as h5pyFile
+from h5py._hl.dataset import Dataset
+from h5py._hl.group import Group
+
+
+def h5save(filename, datadict, timestamp=True):
+    """save dataset to hdf5 format (for load see print(h5load.__doc__))
+    input:
+        - desired (path/)filename as string
+        - dictionary of data
+    return:
+        - saves data to "(path/)timestamp_filename.h5"
+        - complete (path/)filename is returned
+    usage-example:
+                datadict = {'dataset1' : {'x': array(...), 'y': array(...)},
+                            'dataset2' : {'x': array(...), 'y': array(...), 'yerr': array(...)},
+                            'parameter1' : 1.337,
+                            'list1' : [1, 2, 'c']}
+                h5save(filename, True. datadict)
+    """
+
+    def dict2h5(datadict, h5id):
+        for key, val in datadict.items():
+            if isinstance(key, bytes):
+                key = key.decode().replace("/", "|")
+            else:
+                key = key.replace("/", "|")
+
+            if isinstance(val, (list, tuple, str, bytes, int, float, np.ndarray)):
+                try:
+                    h5id.create_dataset(key, data=val)
+                except:
+                    print(
+                        "Data of type {:} ({:}) is not yet supported, sorry for that!".format(
+                            type(val), key
+                        )
+                    )
+            elif isinstance(val, (dict)):
+                hdf5_subid = h5id.create_group(key)
+                dict2h5(val, hdf5_subid)
+            else:
+                print(
+                    "Data of type {:} ({:}) is not yet supported, sorry for that!".format(
+                        type(val), key
+                    )
+                )
+                raise Exception("Datatype is not yet supported, sorry for that!")
+        return
+
+    if timestamp:
+        path = "/".join(filename.split("/")[:-1] + [""])
+        filename = strftime("%Y%m%d%H%M%S") + "_" + filename.split("/")[-1]
+        filename = path + filename
+    if filename[-3:] != ".h5":
+        filename += ".h5"
+
+    hdf5_fid = h5pyFile(filename, "w")
+
+    metadata = {
+        "Date": strftime("%d.%m.%Y %H:%M:%S"),
+        "Uname": str(os.uname()),
+        "User": os.getlogin(),
+    }
+    hdf5_subid = hdf5_fid.create_group("metadata")
+    dict2h5(metadata, hdf5_subid)
+
+    dict2h5(datadict, hdf5_fid)
+
+    hdf5_fid.close()
+    return filename
+
+
+def h5load(filename):
+    """h5load(filename, verbose)
+    input:
+        - filename (as string) of h5save savedfile
+        - desired verbosity
+    return:
+        - dictionary of saved data
+    ALTERNATIVE:
+        if the dataset is too large for memory it is also possible to work with it on disk:
+        >>> import h5py
+        >>> data = h5py.File(filename, 'r')
+    """
+
+    def h52dict(h5id, datadict):
+        for key, val in h5id.items():
+            if isinstance(val, (Dataset)):
+                datadict[key] = h5id[key][()]
+            elif isinstance(val, (Group)):
+                datadict[key] = {}
+                h52dict(h5id[key], datadict[key])
+            else:
+                print(
+                    "Data of type {:} ({:}) is not yet supported, sorry for that!".format(
+                        type(val), key
+                    )
+                )
+                raise Exception("Datatype is not yet supported, sorry for that!")
+        return
+
+    if filename[-3:] != ".h5":
+        filename += ".h5"
+
+    data = {}
+    hdf5_fid = h5pyFile(filename, "r")
+    h52dict(hdf5_fid, data)
+    hdf5_fid.close()
+    return data