Test Amber validation with CI

.github/workflows/test_amber.yaml

@@ -0,0 +1,54 @@
+name: Amber conversion validation
+
+on:
+  push:
+    branches:
+      - "main"
+  pull_request:
+    branches:
+      - "main"
+
+defaults:
+  run:
+    shell: bash -l {0}
+
+concurrency:
+  group: "${{ github.workflow }}-${{ github.ref }}"
+  cancel-in-progress: true
+
+jobs:
+  test:
+    name: Test on ${{ matrix.os }}, Python ${{ matrix.python-version }}, OpenMM ${{ matrix.openmm-version }}
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu-latest]
+        python-version: ["3.12"]
+        openmm-version: ["8.3.1"]
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: Setup Conda Environment
+      uses: mamba-org/setup-micromamba@v2
+      with:
+        environment-file: devtools/conda-envs/test_env.yaml
+        create-args: >-
+          python=${{ matrix.python-version }}
+          openmm=${{ matrix.openmm-version }}
+
+    - name: Log into GitHub container registry
+      uses: docker/login-action@v3
+      with:
+        registry: ghcr.io
+        username: ${{ github.actor }}
+        password: ${{ secrets.GITHUB_TOKEN }}
+
+    # NOTE: Currently, testing the conversion itself is not done as the
+    # conversion requires multiple modifications to the internals of a
+    # particular version of ParmEd.  But the output FFXML files do get tested:
+    - name: Test Amber conversion
+      run: |
+        ./test_amber.py cases/* --sander --openmm-amber --openmm-ffxml
+      working-directory: ./amber/test/

amber/biopolymer.yaml

@@ -67,6 +67,15 @@
     A Parameterization of Cholesterol for Mixed Lipid Bilayer Simulation within the Amber Lipid14 Force Field. J Phys Chem B, 2015, 119, 12424-12435.
   Test:
   - lipids
+- Source: leaprc.lipid21
+  CharmmFFXMLFilename: ../openmmforcefields/ffxml/charmm/charmm36_nowaters.xml
+  CharmmLipid2AmberFilename: files/charmmlipid2amber.csv
+  Reference:
+  - >-
+    Dickson, C.J.; Walker, R.C.; Gould, I.R.
+    Lipid21: Complex Lipid Membrane Simulations with AMBER. J. Chem. Theory Comput., 2022, 18, 1726-1736.
+  Test:
+  - lipids
 - Source: leaprc.protein.ff14SB
   Prefix: protein
   Options:
@@ -111,7 +120,7 @@
   CMAPFRCMOD: files/ff19SB/frcmod.ff19SB_only_cmap
   Prefix: protein
   Options:
-    keep_types: ["P"]
+    keep_types: ["P", "OP"]
     override_level:
       HYP: 2
       CHYP: 2
@@ -327,6 +336,7 @@
     override_level:
       HYP: 2
       CHYP: 2
+    keep_types: ["OP"]
   Reference:
   - >-
     Maier, J.A., Martinez, C., Kasavajhala, K., Wickstrom, L., Hauser, K.E., and Simmerling, C. (2015).

amber/convert.sh

@@ -5,6 +5,6 @@ set -e
 python convert_amber_ions.py "${AMBERHOME}"
 python convert_amber.py --input solvents.yaml --verbose
 python convert_amber.py --input biopolymer.yaml --verbose
-python convert_amber.py --input gaff.yaml --verbose
 python convert_amber.py --input glycam/glycan.yaml --verbose
 python convert_amber.py --combination-tests --verbose
+python postprocess_amber.py

amber/convert_amber.py

@@ -1997,18 +1997,13 @@ def modify_glycan_ffxml(input_ffxml_path):
     script = root.find("Script")
     text = script.text
 
-    # Bad NH torsion
-    text_split = text.split("\n")
-    text_NH_removed = [line for line in text_split if "NH" not in line]
-    text = "\n".join(text_NH_removed)
-
-    # Unscaled types
-    pattern = re.compile("unscaled_types = set\\((\\[(.*\n)*?.*?\\])\\)")
-    match = pattern.search(text)
-    types = eval(match.group(1))
-    types = [(replacements[t[0]], replacements[t[1]], replacements[t[2]], replacements[t[3]]) for t in types]
-    types = ",\n    ".join(str(t) for t in types)
-    script.text = pattern.sub(f"unscaled_types = set([{types}])", text)
+    text_lines = []
+    pattern = re.compile(r"(\(\s*')([^']*)('\s*,\s*')([^']*)('\s*,\s*')([^']*)('\s*,\s*')([^']*)('\s*(?:,\s*)?\)\s*:)")
+    for line in text.split("\n"):
+        if "NH" in line:
+            continue
+        text_lines.append(pattern.sub(lambda m: f"{m[1]}{replacements[m[2]]}{m[3]}{replacements[m[4]]}{m[5]}{replacements[m[6]]}{m[7]}{replacements[m[8]]}{m[9]}", line))
+    script.text = "\n".join(text_lines)
 
     # Add initialization script for setting up GlycamTemplateMatcher
     initialization_script = etree.SubElement(root, "InitializationScript")

amber/postprocess_amber.py

@@ -0,0 +1,30 @@
+import io
+import os
+import lxml.etree as etree
+
+FFXML_DIR = "../openmmforcefields/ffxml/amber"
+
+def main():
+    """
+    Apply miscellaneous fixes to generated FFXML files.  It is better to do this
+    separately than to try to add yet another layer of complexity to the script
+    `convert_amber.py`.  Eventually if this infrastructure can be replaced, and
+    such fixes are still required, the approach used in the CHARMM conversion
+    for patching can be employed.
+    """
+
+    # Fix ACE and NME in ff03 and ff03r1:
+    for force_field in ("ff03", "protein.ff03.r1"):
+        path = os.path.join(FFXML_DIR, f"{force_field}.xml")
+        tree = etree.parse(path)
+        find_and_remove_xml_elements(tree.getroot(), tree.findall("./Residues/Residue[@name='ACE']/ExternalBond[@atomName='HH31']"))
+        find_and_remove_xml_elements(tree.getroot(), tree.findall("./Residues/Residue[@name='NME']/ExternalBond[@atomName='CH3']"))
+        tree.write(path)
+
+def find_and_remove_xml_elements(root, target_elements):
+    root[:] = [child for child in root if child not in target_elements]
+    for child in root:
+        find_and_remove_xml_elements(child, target_elements)
+
+if __name__ == "__main__":
+    main()