--

Author: valentin-gauthier-geosiris

energyml-utils/.gitignore

@@ -74,4 +74,8 @@ rc/**/*.hdf5
 # WIP
 src/energyml/utils/wip*
 scripts
-rc/camunda
+rc/camunda
+
+
+# code profiling
+*.prof

energyml-utils/example/attic/parsing_improvement_test.py

@@ -34,6 +34,8 @@ def reexport_in_memory_par_read(filepath: str, output_folder: Optional[str] = No
     suffix = "opti" if is_opti else "std"
     if os.environ.get("__ENV__IMPROVEMENT_LXML__", "0") == "1":
         suffix += "_lxml"
+    if os.environ.get("__ENV__IMPROVEMENT__GET_MEMBER__", "0") == "1":
+        suffix += "_get_member"
 
     path_in_memory = filepath.replace(".epc", f"_parsing_imp_xml_{suffix}.epc")
     if output_folder:
@@ -46,17 +48,20 @@ def reexport_in_memory_par_read(filepath: str, output_folder: Optional[str] = No
     epc.export_file(path_in_memory, parallel=True)
 
 
+# ===================================
+
+
 def time_test(f: callable, **kwargs):
-    print(f"⏳ Testing {f.__name__}...")
+    print(f" Testing {f.__name__}...")
     start = time.perf_counter()
     f(**kwargs)
     elapsed_inmem = time.perf_counter() - start
     # results.append(("In-Memory (Epc)", elapsed_inmem))
-    print(f"   ✓ Completed in {elapsed_inmem:.3f}s\n")
+    print(f"    Completed in {elapsed_inmem:.3f}s\n")
     return ("In-Memory (Epc)", elapsed_inmem)
 
 
-if __name__ == "__main__":
+if __name__ == "__main__xmlcontext__":
     logging.basicConfig(level=logging.DEBUG)
 
     os.environ["__ENV__IMPROVEMENT__"] = "0"
@@ -82,3 +87,26 @@ def time_test(f: callable, **kwargs):
         filepath=sys.argv[1] if len(sys.argv) > 1 else "rc/epc/80wells_surf.epc",
         output_folder="results",
     )
+
+if __name__ == "__main__":
+    from energyml.resqml.v2_2.resqmlv2 import TriangulatedSetRepresentation
+
+    print(TriangulatedSetRepresentation.__class__.__module__)
+    print(TriangulatedSetRepresentation.__dataclass_fields__.keys())
+
+    # logging.basicConfig(level=logging.DEBUG)
+
+    # os.environ["__ENV__IMPROVEMENT__GET_MEMBER__"] = "0"
+
+    time_test(
+        reexport_in_memory_par_read,
+        filepath=sys.argv[1] if len(sys.argv) > 1 else "rc/epc/80wells_surf.epc",
+        output_folder="results",
+    )
+
+    # os.environ["__ENV__IMPROVEMENT__GET_MEMBER__"] = "1"
+    # time_test(
+    #     reexport_in_memory_par_read,
+    #     filepath=sys.argv[1] if len(sys.argv) > 1 else "rc/epc/80wells_surf.epc",
+    #     output_folder="results",
+    # )

energyml-utils/pyproject.toml

@@ -83,6 +83,7 @@ black = "^22.3.0"
 pylint = "^2.7.2"
 click = ">=8.1.3, <=8.1.3" # upper version than 8.0.2 fail with black
 pdoc3 = "^0.10.0"
+snakeviz = "^2.1.0" # code perf tests
 pydantic = { version =  "^2.0"}
 energyml-common2-0 = "^1.12.0"
 energyml-common2-1 = "^1.12.0"

energyml-utils/src/energyml/utils/constants.py

@@ -415,8 +415,21 @@ def file_extension_to_mime_type(extension: str) -> Optional[str]:
 # OPTIMIZED UTILITY FUNCTIONS
 # ===================================
 
+_SNAKE_CASE_PATTERNS = [
+    (re.compile(r"(.)([A-Z][a-z]+)"), r"\1_\2"),
+    (re.compile(r"__([A-Z])"), r"_\1"),
+    (re.compile(r"([a-z0-9])([A-Z])"), r"\1_\2"),
+]
+
 
 def snake_case(string: str) -> str:
+    """Transform a string into snake_case (optimized with pre-compiled regexes)"""
+    for pattern, repl in _SNAKE_CASE_PATTERNS:
+        string = pattern.sub(repl, string)
+    return string.lower()
+
+
+def snake_case_old(string: str) -> str:
     """Transform a string into snake_case"""
     string = re.sub("(.)([A-Z][a-z]+)", r"\1_\2", string)
     string = re.sub("__([A-Z])", r"_\1", string)

energyml-utils/src/energyml/utils/introspection.py

@@ -1,5 +1,6 @@
 # Copyright (c) 2023-2024 Geosiris.
 # SPDX-License-Identifier: Apache-2.0
+from functools import lru_cache
 import inspect
 import json
 import logging
@@ -85,6 +86,7 @@ def get_module_classes_from_name(mod_name: str) -> List:
     return get_module_classes(sys.modules[mod_name])
 
 
+@lru_cache(maxsize=None)
 def get_module_classes(mod: ModuleType) -> List:
     return inspect.getmembers(mod, inspect.isclass)