Add a workflow to generate the header file from the SMILES templates …

…file (#6)

.github/workflows/update_templates.yml

@@ -0,0 +1,76 @@
+name: Update the header file and image gallery
+on:
+  workflow_dispatch: ~
+  workflow_call: ~
+  push:
+    branches:
+    - main
+    paths:
+      - templates.smi
+
+permissions:
+  contents: write
+
+env:
+  tpl_hdr: 'template_smiles.h'
+  gallery_file: 'gallery.md'
+  img_dir: 'img'
+
+jobs:
+  Check_new_templates:
+    runs-on: ubuntu-latest
+    steps:
+    - name: Checkout repo
+      uses: actions/checkout@v3
+    - name: Set up Python & RDKit
+      uses: mamba-org/setup-micromamba@v1
+      with:
+        micromamba-version: 'latest'
+        environment-file: environment.yml
+        generate-run-shell: true
+        cache-environment: true
+        post-cleanup: 'none'
+    - name: Run basic checks
+      if: success()
+      run: pytest -n=auto checks/test_requisites.py
+      shell: micromamba-shell {0}
+    - name: Run check for duplicates
+      if: success()
+      run: pytest -sv checks/test_duplicates.py
+      shell: micromamba-shell {0}
+    - name: Generate the header file
+      id: header-gen
+      if: success()
+      run: python3 src/header_generation.py
+      shell: micromamba-shell {0}
+    - name: Test the header file
+      if: ${{ steps.header-gen.outputs.header_changed == 'true' && success() }}
+      run: g++ -std=c++17 -I. checks/test_program.cpp -o test && ./test
+      shell: micromamba-shell {0}
+    - name: Update the gallery
+      if: ${{ steps.header-gen.outputs.header_changed == 'true' && success() }}
+      run: python3 src/update_gallery.py
+      shell: micromamba-shell {0}
+
+    - name: Commit and push the new header file
+      if: ${{ steps.header-gen.outputs.header_changed == 'true' && success() }}
+      run: |
+          if ! git diff --quiet --exit-code ${{ env.tpl_hdr }}; then
+              git config user.name "github-actions[bot]"
+              git config user.email "github-actions[bot]@noreply.github.com"
+
+              git checkout main
+              git pull
+
+              git add ${{ env.tpl_hdr }} ${{ env.gallery_file }} ${{ env.img_dir }}
+
+              coauthor=$(git log -1 --pretty=format:'%an <%ae>' -- ${{ env.tpl_hdr }})
+
+              > commitmsg
+              echo -e "[bot] Update molecular templates header\n\n" >> commitmsg
+              echo -e "Co-authored-by: ${coauthor}\n" >> commitmsg
+
+              git commit -F commitmsg
+
+              git push
+          fi

.gitignore

@@ -0,0 +1 @@
+**/__pycache__

checks/test_duplicates.py

@@ -1,10 +1,8 @@
 import pytest
-
+from common_code import load_templates
 from rdkit import Chem
 from rdkit.Chem import RegistrationHash
 
-from common_code import load_templates
-
 
 @pytest.mark.xfail(reason='we know we have some duplicates')
 def test_check_duplicates():
@@ -19,8 +17,8 @@ def test_check_duplicates():
     duplicates = 0
     for i, smiles, cxsmiles in load_templates():
         mol = Chem.MolFromSmiles(cxsmiles)
-        mol_layers = RegistrationHash.GetMolLayers(mol,
-                                                   enable_tautomer_hash_v2=True)
+        mol_layers = RegistrationHash.GetMolLayers(
+            mol, enable_tautomer_hash_v2=True)
         mol_hash = RegistrationHash.GetMolHash(mol_layers)
         if (seen := all_templates.get(mol_hash, None)) is not None:
             seen_idx, seen_smiles = seen

checks/test_program.cpp

@@ -0,0 +1,15 @@
+#include <iostream>
+
+#include "template_smiles.h"
+
+int main()
+{
+    size_t i = 0;
+    for (const auto& cxsmiles : TEMPLATE_SMILES) {
+        std::cout << "Template #" << ++i << ": " << cxsmiles << std::endl;
+    }
+
+    std::cout << "All templates in header listed, check passed." << std::endl;
+
+    return 0;
+}

checks/test_requisites.py

@@ -1,8 +1,6 @@
 import pytest
-
-from rdkit import Chem
-
 from common_code import load_templates
+from rdkit import Chem
 
 
 def generate_params():

environment.yml

@@ -3,6 +3,7 @@ channels:
   - conda-forge
 dependencies:
   - gitpython
+  - gxx
   - python=3.11
   - pytest
   - pytest-xdist

src/header_generation.py

@@ -0,0 +1,83 @@
+import filecmp
+import os
+import shutil
+import tempfile
+
+from rdkit import Chem
+
+HEADER_FILE = 'template_smiles.h'
+TEMPLATE_FILE = 'templates.smi'
+
+HEADER_TEXT = """//
+//  Copyright (C) 2023 Schrödinger, LLC
+//
+//   @@ All Rights Reserved @@
+//  This file is part of the RDKit.
+//  The contents are covered by the terms of the BSD license
+//  which is included in the file license.txt, found at the root
+//  of the RDKit source tree.
+//
+// THIS FILE IS AUTOMATICALLY GENERATED. It contains templates used
+// in 2D coordinate generation. If you want to contribute to these
+// templates, please refer to instructions in:
+// https://github.com/rdkit/molecular_templates/blob/main/README.md
+//
+
+#include <string>
+#include <vector>
+
+// clang-format off
+const std::vector<std::string> TEMPLATE_SMILES = {
+"""
+
+
+def clean_smiles(template_smiles):
+    """
+    Translate all atoms into dummy atoms so that templates are not atom-specific.
+    """
+    template = Chem.MolFromSmiles(template_smiles)
+    for atom in template.GetAtoms():
+        atom.SetAtomicNum(0)
+
+    # TO_DO: replace bonds with query bonds
+
+    return Chem.MolToCXSmiles(template)
+
+
+def generate_header(generated_header_path):
+    with open(generated_header_path, 'w') as f_out:
+        f_out.write(HEADER_TEXT)
+        with open(TEMPLATE_FILE, 'r') as f_in:
+            for line in f_in:
+                if not (cxsmiles := line.strip()):
+                    continue
+
+                # TO_DO: Clean smiles to make them atom-type and bond-type agnostic
+                # cxsmiles = clean_smiles(cxsmiles)
+
+                f_out.write(f'    "{cxsmiles}",\n')
+        f_out.write('};\n// clang-format on\n')
+    print(f"Successfully generated {generated_header_path}")
+
+
+def check_header_changed(header_file_path):
+    if header_changed := not filecmp.cmp(header_file_path, HEADER_FILE):
+        if gh_output := os.environ.get('GITHUB_OUTPUT', ''):
+            with open(gh_output, 'a') as f:
+                f.write(f'header_changed={str(header_changed).lower()}')
+    print(f'Header file has {"" if header_changed else "not "}changed')
+    return header_changed
+
+
+def main():
+    with tempfile.TemporaryDirectory() as tmpdir:
+        generated_header_path = os.path.join(tmpdir, HEADER_FILE)
+        generate_header(generated_header_path)
+
+        if check_header_changed(generated_header_path):
+            shutil.copy(generated_header_path, HEADER_FILE)
+            print(f'Updated {HEADER_FILE} with the generated header')
+
+
+if __name__ == '__main__':
+    main()

src/img_generation.py

@@ -3,7 +3,7 @@
 
 import git
 from rdkit import Chem
-from rdkit.Chem.Draw import MolToImage
+from rdkit.Chem.Draw import MolDraw2DSVG, MolToImage
 
 from imgur_upload import upload_img
 
@@ -42,7 +42,7 @@ def get_new_templates():
             yield line_nums[cxsmiles], cxsmiles
 
 
-def draw_mol(cxsmiles, idx, output_dir='.'):
+def draw_png(cxsmiles, idx, output_dir='.'):
     smiles = cxsmiles.split()[0]
     legend = f"#{idx} {smiles}"
     mol = Chem.MolFromSmiles(cxsmiles)
@@ -57,6 +57,21 @@ def draw_mol(cxsmiles, idx, output_dir='.'):
     return fpath, legend
 
 
+def draw_svg(cxsmiles, fname, legend):
+    mol = Chem.MolFromSmiles(cxsmiles)
+    if not mol.GetNumConformers():
+        raise ValueError("SMILES must include coordinates")
+
+    print(f'Creating SVG image {legend}: {fname}')
+
+    drawer = MolDraw2DSVG(img_width, img_height)
+    drawer.DrawMolecule(mol, legend=legend)
+    drawer.FinishDrawing()
+
+    with open(fname, 'w') as f:
+        f.write(drawer.GetDrawingText())
+
+
 def export_image_urls(template_imgs):
     if gh_output := os.environ.get('GITHUB_OUTPUT', ''):
         markdown_imgs = [f'![{title}]({url})' for url, title in template_imgs]
@@ -68,7 +83,7 @@ def main():
     template_imgs = []
     with tempfile.TemporaryDirectory() as tmpdir:
         for idx, cxsmiles in get_new_templates():
-            fpath, title = draw_mol(cxsmiles, idx, tmpdir)
+            fpath, title = draw_png(cxsmiles, idx, tmpdir)
             img_url, title = upload_img(fpath, title)
             template_imgs.append((img_url, title))
 

src/imgur_upload.py

@@ -9,7 +9,7 @@
 UPLOAD_RETRIES = 5
 
 IMGUR_CLIENT_ID = os.environ.get('IMGUR_CLIENT_ID', None)
-IMGUR_TOKEN = json.loads(os.environ.get('IMGUR_TOKEN', ''))
+IMGUR_TOKEN = json.loads(os.environ.get('IMGUR_TOKEN', '[]'))
 
 
 def upload_img(fpath, title):