bugfix for epc rels export

Author: valentin-gauthier-geosiris

energyml-utils/.gitignore

@@ -62,6 +62,10 @@ docs/*.md
 *.geojson
 *.vtk
 *.stl
+rc/specs
+rc/**/*.epc
+rc/**/*.h5
+rc/**/*.hdf5
 
 
 # WIP

energyml-utils/example/main_stream.py

@@ -13,7 +13,7 @@
 
 from energyml.utils.introspection import get_obj_uri
 from energyml.utils.constants import EpcExportVersion
-from energyml.utils.epc_stream import read_epc_stream
+from energyml.utils.epc_stream_old import read_epc_stream
 from energyml.utils.epc import (
     Epc,
     create_energyml_object,

energyml-utils/example/main_stream_sample.py

@@ -0,0 +1,192 @@
+import os
+import sys
+import logging
+from energyml.utils.epc_stream import EpcStreamReader, RelsUpdateMode
+from energyml.eml.v2_3.commonv2 import Citation
+from energyml.resqml.v2_2.resqmlv2 import (
+    TriangulatedSetRepresentation,
+    BoundaryFeatureInterpretation,
+    BoundaryFeature,
+    HorizonInterpretation,
+)
+from energyml.utils.introspection import epoch_to_date, epoch
+from energyml.utils.epc import as_dor, gen_uuid, get_obj_identifier
+from energyml.utils.constants import EPCRelsRelationshipType
+
+
+def sample_objects():
+    """Create sample EnergyML objects for testing."""
+    # Create a BoundaryFeature
+    bf = BoundaryFeature(
+        citation=Citation(
+            title="Test Boundary Feature",
+            originator="Test",
+            creation=epoch_to_date(epoch()),
+        ),
+        uuid="25773477-ffee-4cc2-867d-000000000001",
+        object_version="1.0",
+    )
+
+    # Create a BoundaryFeatureInterpretation
+    bfi = BoundaryFeatureInterpretation(
+        citation=Citation(
+            title="Test Boundary Feature Interpretation",
+            originator="Test",
+            creation=epoch_to_date(epoch()),
+        ),
+        uuid="25773477-ffee-4cc2-867d-000000000002",
+        object_version="1.0",
+        interpreted_feature=as_dor(bf),
+    )
+
+    # Create a HorizonInterpretation (independent object)
+    horizon_interp = HorizonInterpretation(
+        citation=Citation(
+            title="Test HorizonInterpretation",
+            originator="Test",
+            creation=epoch_to_date(epoch()),
+        ),
+        interpreted_feature=as_dor(bf),
+        uuid="25773477-ffee-4cc2-867d-000000000003",
+        object_version="1.0",
+        domain="depth",
+    )
+
+    # Create a TriangulatedSetRepresentation
+    trset = TriangulatedSetRepresentation(
+        citation=Citation(
+            title="Test TriangulatedSetRepresentation",
+            originator="Test",
+            creation=epoch_to_date(epoch()),
+        ),
+        uuid="25773477-ffee-4cc2-867d-000000000004",
+        object_version="1.0",
+        represented_object=as_dor(horizon_interp),
+    )
+
+    return {
+        "bf": bf,
+        "bfi": bfi,
+        "trset": trset,
+        "horizon_interp": horizon_interp,
+    }
+
+
+def main(epc_file_path: str):
+    epc = EpcStreamReader(
+        epc_file_path=epc_file_path, enable_parallel_rels=True, rels_update_mode=RelsUpdateMode.UPDATE_AT_MODIFICATION
+    )
+
+    # logging.info(epc.get_statistics())
+
+    for obj in epc.list_objects():
+        logging.info(f"Object: {obj}")
+
+
+def test_create_epc(path: str):
+    # delete file if exists
+    if os.path.exists(path):
+        os.remove(path)
+    logging.info(f"==> Creating new EPC at {path}...")
+    epc = EpcStreamReader(epc_file_path=path, rels_update_mode=RelsUpdateMode.UPDATE_AT_MODIFICATION)
+
+    data = sample_objects()
+
+    logging.info("==> Creating sample objects and adding to EPC...")
+
+    logging.info("==> Adding horizon interpretation")
+    epc.add_object(data["horizon_interp"])
+    logging.info(f"horizon rels : {epc.get_obj_rels(data['horizon_interp'])}")
+
+    logging.info("==> Adding boundary feature")
+    epc.add_object(data["bf"])
+    logging.info(f"boundary feature rels : {epc.get_obj_rels(data['bf'])}")
+
+    logging.info("==> Adding boundary feature interpretation")
+    epc.add_object(data["bfi"])
+    logging.info("==> Adding triangulated set representation")
+    epc.add_object(data["trset"])
+
+    # Debug: Print all metadata identifiers
+    logging.info(f"==> All metadata identifiers: {list(epc._metadata_mgr._metadata.keys())}")
+
+    logging.info("==> All objects added. Closing EPC to write to disk.")
+
+    horizon_id = get_obj_identifier(data["horizon_interp"])
+    logging.info(f"==> Horizon identifier: {horizon_id}")
+    logging.info(f"==> Horizon in metadata: {horizon_id in epc._metadata_mgr._metadata}")
+
+    # Debug: Test _id_from_uri_or_identifier
+    resolved_id = epc._id_from_uri_or_identifier(data["horizon_interp"])
+    logging.info(f"==> Resolved ID from object: {resolved_id}")
+    logging.info(
+        f"==> Resolved ID in metadata: {resolved_id in epc._metadata_mgr._metadata if resolved_id else 'ID is None'}"
+    )
+
+    horizon_rels = epc.get_obj_rels(data["horizon_interp"])
+    assert (
+        len(horizon_rels) == 2
+    ), f"Expected 2 relationships in horizon rels since both bfi and trset should refer to horizon as interpreted feature {horizon_rels}"
+    epc.close()
+
+    epc_reopen = EpcStreamReader(epc_file_path=path, rels_update_mode=RelsUpdateMode.UPDATE_AT_MODIFICATION)
+
+    horizon_rels = epc_reopen.get_obj_rels(data["horizon_interp"])
+    assert (
+        len(horizon_rels) == 2
+    ), f"Expected 2 relationships in horizon rels since both bfi and trset should refer to horizon as interpreted feature {horizon_rels}"
+
+    logging.info("==> Reopened EPC, listing objects:")
+    for obj in epc_reopen.list_objects():
+        logging.info(f"Object: {obj}")
+        obj_rels = epc_reopen.get_obj_rels(obj)
+        logging.info(f"\tObject rels: {obj_rels}")
+        dest_rels = [r for r in obj_rels if r.type_value == str(EPCRelsRelationshipType.DESTINATION_OBJECT)]
+        logging.info(f"\tObject DESTINATION rels: {dest_rels}")
+
+    # remove trset to check if horizon has no more source rels
+    epc_reopen.remove_object(data["trset"])
+
+    horizon_rels_after_removal = epc_reopen.get_obj_rels(data["horizon_interp"])
+    logging.info(f"Horizon interpretation rels after removing trset: {horizon_rels_after_removal}")
+    source_rels_after_removal = [
+        r for r in horizon_rels_after_removal if r.type_value == str(EPCRelsRelationshipType.SOURCE_OBJECT)
+    ]
+    logging.info(f"Horizon interpretation SOURCE rels after removing trset: {source_rels_after_removal}")
+    assert (
+        len(source_rels_after_removal) == 0
+    ), "Expected no SOURCE relationships in horizon rels after removing trset since trset was the only destination referring to horizon"
+
+    assert (
+        len(horizon_rels_after_removal) == 1
+    ), "Expected 1 relationship in horizon rels after removing trset since bfi should still refer to horizon as interpreted feature"
+
+    epc_reopen.close()
+
+
+def test_create_epc_v2(path: str):
+
+    if os.path.exists(path):
+        os.remove(path)
+    logging.info(f"==> Creating new EPC at {path}...")
+    epc = EpcStreamReader(epc_file_path=path, rels_update_mode=RelsUpdateMode.UPDATE_AT_MODIFICATION)
+
+    data = sample_objects()
+
+    epc.add_object(data["bf"])
+    # epc.add_object(data["bfi"])
+    epc.add_object(data["horizon_interp"])
+    epc.add_object(data["trset"])
+
+    hi_rels = epc.get_obj_rels(data["horizon_interp"])
+
+    logging.info(f"Horizon interpretation rels: {hi_rels}")
+
+
+if __name__ == "__main__":
+    logging.basicConfig(level=logging.DEBUG)
+
+    # main((sys.argv[1] if len(sys.argv) > 1 else None) or "wip/80wells_surf.epc")
+
+    # test_create_epc("wip/test_create.epc")
+    test_create_epc_v2("wip/test_create.epc")

energyml-utils/example/validate_epc_example.py

@@ -0,0 +1,99 @@
+#!/usr/bin/env python3
+# Copyright (c) 2023-2024 Geosiris.
+# SPDX-License-Identifier: Apache-2.0
+"""
+Example script demonstrating EPC validation.
+
+This script shows how to validate EPC files and generate reports.
+"""
+
+import sys
+from pathlib import Path
+
+from energyml.utils.epc_validator import validate_epc_file
+
+
+def validate_single_file(epc_path: str) -> None:
+    """Validate a single EPC file and print results."""
+    print(f"\n{'=' * 70}")
+    print(f"Validating: {epc_path}")
+    print(f"{'=' * 70}\n")
+
+    try:
+        result = validate_epc_file(epc_path, strict=True, check_relationships=True)
+
+        print(result)
+
+        if result.is_valid:
+            print("\n✓ Validation PASSED!")
+        else:
+            print("\n✗ Validation FAILED!")
+            sys.exit(1)
+
+    except Exception as e:
+        print(f"\n✗ Error during validation: {e}")
+        sys.exit(1)
+
+
+def validate_directory(directory: str) -> None:
+    """Validate all EPC files in a directory."""
+    print(f"\n{'=' * 70}")
+    print(f"Validating all EPC files in: {directory}")
+    print(f"{'=' * 70}\n")
+
+    epc_files = list(Path(directory).glob("**/*.epc"))
+
+    if not epc_files:
+        print(f"No EPC files found in {directory}")
+        return
+
+    print(f"Found {len(epc_files)} EPC file(s)\n")
+
+    results = {}
+    for epc_file in epc_files:
+        print(f"Validating {epc_file.name}...", end=" ")
+        result = validate_epc_file(str(epc_file))
+
+        if result.is_valid:
+            print("✓ PASSED")
+        else:
+            print("✗ FAILED")
+            for error in result.errors[:3]:  # Show first 3 errors
+                print(f"    - {error}")
+            if len(result.errors) > 3:
+                print(f"    ... and {len(result.errors) - 3} more errors")
+
+        results[epc_file.name] = result
+
+    # Summary
+    print(f"\n{'=' * 70}")
+    print("SUMMARY")
+    print(f"{'=' * 70}")
+    passed = sum(1 for r in results.values() if r.is_valid)
+    failed = len(results) - passed
+    print(f"Total files: {len(results)}")
+    print(f"Passed: {passed}")
+    print(f"Failed: {failed}")
+
+
+def main():
+    """Main entry point."""
+    if len(sys.argv) < 2:
+        print("Usage:")
+        print(f"  {sys.argv[0]} <epc_file>        # Validate a single file")
+        print(f"  {sys.argv[0]} <directory>       # Validate all EPC files in directory")
+        sys.exit(1)
+
+    path = sys.argv[1]
+
+    if Path(path).is_file():
+        validate_single_file(path)
+    elif Path(path).is_dir():
+        validate_directory(path)
+    else:
+        print(f"Error: '{path}' is neither a file nor a directory")
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()