RMG Installation Update (#820)

Author: alongd

.github/workflows/ci.yml

@@ -71,23 +71,6 @@ jobs:
           echo "::endgroup::"
           julia --version # Check that Julia is installed correctly
 
-      - name: Set RMG-Py + database in PATH and PYTHONPATH
-        shell: micromamba-shell {0}
-        working-directory: ARC
-        run: |
-          # bash devtools/install_rmg.sh
-          echo "RMG_PY_PATH=$(realpath ../RMG-Py)"       >> $GITHUB_ENV
-          echo "RMG_DB_PATH=$(realpath ../RMG-database)" >> $GITHUB_ENV
-          echo "PATH=$(realpath ../RMG-Py):$PATH"        >> $GITHUB_ENV
-          echo "PYTHONPATH=$(realpath ../RMG-Py):$PYTHONPATH" >> $GITHUB_ENV
-
-      - name: Cache RMG-Py source
-        uses: actions/cache@v4
-        with:
-          path: RMG-Py
-          key: ${{ runner.os }}-rmgpy-${{ hashFiles('ARC/devtools/install_rmg.sh') }}
-          restore-keys: ${{ runner.os }}-rmgpy-
-
       - name: Install all extras - CI
         shell: micromamba-shell {0}
         working-directory: ARC

.github/workflows/gh-pages.yml

@@ -20,34 +20,17 @@ jobs:
           path: ARC
           fetch-depth: 1
 
-      # ── pull RMG sources ────────────────────────────────────────
-      - name: Checkout RMG‑Py
-        uses: actions/checkout@v4
-        with:
-          repository: ReactionMechanismGenerator/RMG-Py
-          path: RMG-Py
-          fetch-depth: 1
-
-      - name: Checkout RMG‑database
-        uses: actions/checkout@v4
-        with:
-          repository: ReactionMechanismGenerator/RMG-database
-          path: RMG-database
-          fetch-depth: 1
-
-      # ── build‑only env ───────────────────────────────────────────
+      # ── rmg_env for Arkane/database availability ───────────────
       - name: Set up micromamba (rmg_env)
         uses: mamba-org/setup-micromamba@v2
         with:
           environment-name: rmg_env
-          environment-file: RMG-Py/environment.yml
+          create-args: >-
+            -c rmg -c conda-forge python=3.9 numpy<2 rmg=3.3.0 connie::symmetry
           cache-environment: true
           cache-downloads: true
           generate-run-shell: false
 
-      - name: Build RMG‑Py in rmg_env
-        run: micromamba run -n rmg_env make -C RMG-Py -j"$(nproc)"
-
       # ── docs env (wrapper shell) ────────────────────────────────
       - name: Set up micromamba (arc_env)
         uses: mamba-org/setup-micromamba@v2
@@ -77,9 +60,8 @@ jobs:
       # ── build HTML docs ─────────────────────────────────────────
       - name: Set env vars & Build docs
         run: |
-          export RMG_PATH="$GITHUB_WORKSPACE/RMG-Py"
-          export RMG_DB_PATH="$GITHUB_WORKSPACE/RMG-database"
-          export PYTHONPATH="$GITHUB_WORKSPACE/RMG-Py:$GITHUB_WORKSPACE/ARC:$PYTHONPATH"
+          export RMG_DB_PATH="$MAMBA_ROOT_PREFIX/envs/rmg_env/share/RMG-database"
+          export PYTHONPATH="$GITHUB_WORKSPACE/ARC:$PYTHONPATH"
           export PATH="$GITHUB_WORKSPACE/ARC:$PATH"
           make -C ARC/docs html
         shell: micromamba-shell {0}

.github/workflows/update-cache.yml

@@ -21,42 +21,6 @@ jobs:
       - name: Checkout ARC
         uses: actions/checkout@v4
 
-      # ────────── RMG‑Py ──────────
-      - name: Cache RMG
-        id: cache-rmg
-        uses: actions/cache@v4
-        with:
-          path: RMG-Py
-          key: ${{ runner.os }}-rmg-main
-          restore-keys: |
-            ${{ runner.os }}-rmg-
-      - name: Checkout RMG-Py
-        if: steps.cache-rmg.outputs.cache-hit != 'true'
-        uses: actions/checkout@v4
-        with:
-          repository: ReactionMechanismGenerator/RMG-Py
-          path: RMG-Py
-          ref: 55464c54d1fa61b531e865682df598d33718597d
-          fetch-depth: 1
-
-      # ────────── RMG‑database ──────────
-      - name: Cache RMG-database
-        id: cache-rmg-db
-        uses: actions/cache@v4
-        with:
-          path: RMG-database
-          key: ${{ runner.os }}-rmgdb-main
-          restore-keys: |
-            ${{ runner.os }}-rmgdb-
-      - name: Checkout RMG-database
-        if: steps.cache-rmg-db.outputs.cache-hit != 'true'
-        uses: actions/checkout@v4
-        with:
-          repository: ReactionMechanismGenerator/RMG-database
-          path: RMG-database
-          ref: main
-          fetch-depth: 1
-
       # ────────── AutoTST ──────────
       - name: Cache AutoTST
         id: cache-autotst

.gitignore

@@ -67,3 +67,6 @@ build/*
 
 *.log
 *.xml
+
+# AI Agent files
+AGENTS.md

Makefile

@@ -7,7 +7,7 @@
 DEVTOOLS_DIR := devtools
 
 .PHONY: all help clean test test-unittests test-functional test-all \
-        install-all install-ci install-pyrdl install-rmgdb install-autotst install-gcn \
+        install-all install-ci install-pyrdl install-rmg install-rmgdb install-autotst install-gcn \
         install-gcn-cpu install-kinbot install-sella install-xtb install-torchani install-ob \
         lite check-env compile
 
@@ -28,6 +28,7 @@ help:
 	@echo "  install          Install all external dependencies"
 	@echo "  install-ci       Install all external dependencies for CI (no clean)"
 	@echo "  install-pyrdl    Install PyRDL"
+	@echo "  install-rmg      Install RMG (default: packaged, override via RMG_ARGS=...)"
 	@echo "  install-rmgdb    Install RMG-database"
 	@echo "  install-autotst  Install AutoTST"
 	@echo "  install-gcn      Install TS-GCN (GPU)"
@@ -75,6 +76,9 @@ install-lite:
 install-pyrdl:
 	bash $(DEVTOOLS_DIR)/install_pyrdl.sh
 
+install-rmg:
+	bash $(DEVTOOLS_DIR)/install_rmg.sh $(RMG_ARGS)
+
 install-rmgdb:
 	bash $(DEVTOOLS_DIR)/install_rmgdb.sh
 

arc/family/family.py

@@ -23,6 +23,23 @@
 logger = get_logger()
 
 
+def get_rmg_db_subpath(*parts: str, must_exist: bool = False) -> str:
+    """Return a path under the RMG database, handling both source and packaged layouts."""
+    if RMG_DB_PATH is None:
+        if must_exist:
+            raise FileNotFoundError('RMG_DB_PATH is not set; cannot locate the RMG database.')
+        return os.path.join('input', *parts)
+    candidates = [
+        os.path.join(RMG_DB_PATH, 'input', *parts),
+        os.path.join(RMG_DB_PATH, *parts),
+    ]
+    if must_exist:
+        for candidate in candidates:
+            if os.path.exists(candidate):
+                return candidate
+    return candidates[0]
+
+
 class ReactionFamily(object):
     """
     A class for representing a reaction family.
@@ -72,7 +89,7 @@ def get_groups_file_as_lines(self, consider_arc_families: bool = True) -> List[s
         Returns:
             List[str]: The groups file as a list of lines.
         """
-        groups_path = os.path.join(RMG_DB_PATH, 'input', 'kinetics', 'families', self.label, 'groups.py')
+        groups_path = get_rmg_db_subpath('kinetics', 'families', self.label, 'groups.py', must_exist=True)
         if not os.path.isfile(groups_path):
             if consider_arc_families:
                 groups_path = os.path.join(ARC_FAMILIES_PATH, f'{self.label}.py')
@@ -605,7 +622,7 @@ def get_rmg_recommended_family_sets() -> Dict[str, str]:
         Dict[str, str]: The recommended RMG family sets.
     """
     family_sets = dict()
-    recommended_path = os.path.join(RMG_DB_PATH, 'input', 'kinetics', 'families', 'recommended.py')
+    recommended_path = get_rmg_db_subpath('kinetics', 'families', 'recommended.py', must_exist=True)
     if not os.path.isfile(recommended_path):
         raise FileNotFoundError(f'Could not find the recommended RMG families file at {recommended_path}')
     with open(recommended_path, 'r') as f:

arc/family/family_test.py

@@ -12,6 +12,7 @@
 from arc.family.family import (ReactionFamily,
                                ARC_FAMILIES_PATH,
                                RMG_DB_PATH,
+                               get_rmg_db_subpath,
                                add_labels_to_molecule,
                                check_product_isomorphism,
                                descent_complex_group,
@@ -50,8 +51,10 @@ def setUpClass(cls):
 
     def test_rmgdb_path(self):
         """Test finding the RMG-database path"""
-        self.assertIn('RMG-database', RMG_DB_PATH)
         self.assertTrue(os.path.isdir(RMG_DB_PATH))
+        self.assertTrue(any(token in RMG_DB_PATH for token in ('RMG-database', 'rmgdatabase', 'rmg_database')))
+        self.assertTrue(os.path.isdir(os.path.join(RMG_DB_PATH, 'input'))
+                        or os.path.isdir(os.path.join(RMG_DB_PATH, 'kinetics')))
 
     def test_arc_families_path(self):
         """Test finding the ARC families folder path"""
@@ -920,14 +923,14 @@ def test_get_product_num(self):
     def test_get_reactant_groups_from_template(self):
         """Test getting reactant groups from a template"""
         fam_1 = ReactionFamily('6_membered_central_C-C_shift')
-        groups_path = os.path.join(RMG_DB_PATH, 'input', 'kinetics', 'families', fam_1.label, 'groups.py')
+        groups_path = get_rmg_db_subpath('kinetics', 'families', fam_1.label, 'groups.py', must_exist=True)
         with open(groups_path, 'r') as f:
             groups = f.readlines()
         reactants_1 = get_reactant_groups_from_template(groups)
         self.assertEqual(reactants_1, [['1_5_unsaturated_hexane']])
 
         fam_2 = ReactionFamily('H_Abstraction')
-        groups_path = os.path.join(RMG_DB_PATH, 'input', 'kinetics', 'families', fam_2.label, 'groups.py')
+        groups_path = get_rmg_db_subpath('kinetics', 'families', fam_2.label, 'groups.py', must_exist=True)
         with open(groups_path, 'r') as f:
             groups = f.readlines()
         reactants_2 = get_reactant_groups_from_template(groups)
@@ -936,7 +939,7 @@ def test_get_reactant_groups_from_template(self):
         self.assertEqual(reactants_2, expected_reactants)
 
         fam_3 = ReactionFamily('1,2-Birad_to_alkene')
-        groups_path = os.path.join(RMG_DB_PATH, 'input', 'kinetics', 'families', fam_3.label, 'groups.py')
+        groups_path = get_rmg_db_subpath('kinetics', 'families', fam_3.label, 'groups.py', must_exist=True)
         with open(groups_path, 'r') as f:
             groups = f.readlines()
         reactants_3 = get_reactant_groups_from_template(groups)
@@ -988,7 +991,7 @@ def test_get_recipe_actions(self):
                                    ['LOSE_RADICAL', '*3', '1']])
 
         fam_1 = ReactionFamily('6_membered_central_C-C_shift')
-        groups_path = os.path.join(RMG_DB_PATH, 'input', 'kinetics', 'families', fam_1.label, 'groups.py')
+        groups_path = get_rmg_db_subpath('kinetics', 'families', fam_1.label, 'groups.py', must_exist=True)
         with open(groups_path, 'r') as f:
             groups = f.readlines()
         actions = get_recipe_actions(groups)

arc/processor.py

@@ -8,6 +8,7 @@
 
 import arc.plotter as plotter
 from arc.common import ARC_PATH, get_logger, read_yaml_file, save_yaml_file
+from arc.imports import settings
 from arc.level import Level
 from arc.job.local import execute_command
 from arc.statmech.factory import statmech_factory
@@ -246,11 +247,15 @@ def compare_thermo(species_for_thermo_lib: list,
     species_thermo_path = os.path.join(output_directory, 'RMG_thermo.yml')
     save_yaml_file(path=species_thermo_path,
                    content=[{'label': spc.label, 'adjlist': spc.mol.copy(deep=True).to_adjacency_list()} for spc in species_for_thermo_lib])
+    env_name = settings.get('RMG_ENV_NAME', 'rmg_env')
+    rmg_db_path = settings.get('RMG_DB_PATH') or ""
     commands = ['bash -lc "set -euo pipefail; '
+                f'export RMG_DB_PATH=\\"{rmg_db_path}\\"; '
+                f'export RMG_DATABASE=\\"{rmg_db_path}\\"; '
                 'if command -v micromamba >/dev/null 2>&1; then '
-                f'    micromamba run -n rmg_env python {THERMO_SCRIPT_PATH} {species_thermo_path}; '
+                f'    micromamba run -n {env_name} python {THERMO_SCRIPT_PATH} {species_thermo_path}; '
                 'elif command -v conda >/dev/null 2>&1 || command -v mamba >/dev/null 2>&1; then '
-                f'    conda run -n rmg_env python {THERMO_SCRIPT_PATH} {species_thermo_path}; '
+                f'    conda run -n {env_name} python {THERMO_SCRIPT_PATH} {species_thermo_path}; '
                 'else '
                 '    echo \'❌ Micromamba/Mamba/Conda required\' >&2; exit 1; '
                 'fi"',
@@ -300,7 +305,8 @@ def compare_rates(rxns_for_kinetics_lib: list,
                              'family': rxn.family,
                              } for rxn in rxns_for_kinetics_lib],
                    )
-    env_name = 'rmg_env'
+    env_name = settings.get('RMG_ENV_NAME', 'rmg_env')
+    rmg_db_path = settings.get('RMG_DB_PATH') or ""
     shell_script = f"""if command -v micromamba &> /dev/null; then
     eval "$(micromamba shell hook --shell=bash)"
     micromamba activate {env_name}
@@ -313,6 +319,8 @@ def compare_rates(rxns_for_kinetics_lib: list,
 else
     exit 1
 fi
+export RMG_DB_PATH="{rmg_db_path}"
+export RMG_DATABASE="{rmg_db_path}"
 python {KINETICS_SCRIPT_PATH} {reactions_kinetics_path}   > >(tee -a stdout.log) 2> >(tee -a stderr.log >&2)
 """
     o, e = execute_command(command=shell_script,