downgradin h5py version

Author: valentin-gauthier-geosiris

energyml-utils/example/attic/perf_tests.py

@@ -1,11 +1,25 @@
 # Benchmark de performance pour get_obj_uuid
 import time
 import re
-from types import SimpleNamespace
 
 UUID_RGX: re.Pattern = re.compile(r"[Uu]u?id|UUID")
 
 
+# Version dot
+def get_obj_uuid_pointe(obj):
+    try:
+        return obj.uuid
+    except AttributeError:
+        try:
+            return obj.uid
+        except AttributeError:
+            if isinstance(obj, dict):
+                for k in obj.keys():
+                    if UUID_RGX.match(k):
+                        return obj[k]
+    return None
+
+
 # Version originale
 def get_obj_uuid_original(obj):
     try:
@@ -54,5 +68,12 @@ def __init__(self, uuid):
     assert get_obj_uuid_fast(obj) == obj.uuid
 elapsed_fast = time.perf_counter() - start
 
+# Test version pointe
+start = time.perf_counter()
+for obj in objs:
+    assert get_obj_uuid_pointe(obj) == obj.uuid
+elapsed_point = time.perf_counter() - start
+
 print(f"Original version: {elapsed_original:.6f} s for {N} calls")
 print(f"Optimized version: {elapsed_fast:.6f} s for {N} calls")
+print(f"Point version: {elapsed_point:.6f} s for {N} calls")

energyml-utils/pyproject.toml

@@ -67,7 +67,7 @@ segy = ["segyio"]
 python = "^3.9"
 xsdata = {version = "^24.0", extras = ["cli", "lxml"]}
 energyml-opc = "^1.12.0"
-h5py = { version = "^3.7.0", optional = false }
+h5py = { version = "^3.11.0", optional = false }
 numpy = { version = "^1.16.6", optional = false }
 pyarrow = { version = "^14.0.1", optional = true }
 pandas = { version = "^1.1.0", optional = true }

energyml-utils/src/energyml/utils/data/datasets_io.py

@@ -1024,6 +1024,9 @@ def list_arrays(self, source: Union[BytesIO, str, Any]) -> List[str]:
         def can_handle_file(self, file_path: str) -> bool:
             return os.path.splitext(file_path)[1].lower() in [".h5", ".hdf5"]  # dat for Galaxy compatibility
 
+    # Alias so the public name is always importable
+    HDF5ArrayHandler = MockHDF5ArrayHandler
+
 
 # Parquet Handler
 if __PARQUET_MODULE_EXISTS__:
@@ -1214,6 +1217,9 @@ def list_arrays(self, source: Union[BytesIO, str, Any]) -> List[str]:
         def can_handle_file(self, file_path: str) -> bool:
             return os.path.splitext(file_path)[1].lower() in [".parquet", ".pq"]
 
+    # Alias so the public name is always importable
+    ParquetArrayHandler = MockParquetArrayHandler
+
 
 # CSV Handler
 if __CSV_MODULE_EXISTS__:
@@ -1563,6 +1569,9 @@ def can_handle_file(self, file_path: str) -> bool:
             ext = os.path.splitext(file_path)[1].lower()
             return ext == ".las"
 
+    # Alias so the public name is always importable
+    LASArrayHandler = MockLASArrayHandler
+
 
 # SEG-Y Handler
 if __SEGYIO_MODULE_EXISTS__:
@@ -1792,3 +1801,6 @@ def can_handle_file(self, file_path: str) -> bool:
             """Check if this handler can process the file."""
             ext = os.path.splitext(file_path)[1].lower()
             return ext in [".sgy", ".segy"]
+
+    # Alias so the public name is always importable
+    SEGYArrayHandler = MockSEGYArrayHandler