Fix formatting and inconsistencies between CSV and SMI

.github/workflows/check_abbreviations.yml

@@ -15,3 +15,8 @@ jobs:
       - uses: actions/checkout@v3
       - name: Run duplicate check
         run: ./scripts/check_duplicates_abbreviations.sh
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v3
+      - name: Check CSV/SMI sync
+        run: uv run lint.py

abbreviations.smi

@@ -1,14 +1,14 @@
-Abbreviation SMILES
-C6H4F *c1ccc(F)cc1
-Mes *c1c(C)cc(C)cc1(C)
-S S
-N N
-NH2 N
-–OCH2O– *OCO*
-s-Bu *C(C)CC
-i-Pr *C(C)C
-2,6-(CH3)2C6H3 *C1=C(C)C=CC=C1(C)
-PMB *OCC1=CC=C(C=C1)OC
+Abbreviation	SMILES
+C6H4F	*c1ccc(F)cc1
+Mes	*c1c(C)cc(C)cc1(C)
+S	S
+N	N
+NH2	N
+–OCH2O–	*OCO*
+s-Bu	*C(C)CC
+i-Pr	*C(C)C
+2,6-(CH3)2C6H3	*C1=C(C)C=CC=C1(C)
+PMB	*OCC1=CC=C(C=C1)OC
 Bpin	CC1(C)OB([*])OC1(C)C
 PMP	COC1=CC=C([*])C=C1
 OPP	O=P(OP(O[*])(O)=O)(O)O
@@ -268,6 +268,6 @@ Ad	[*]C12CC3CC(C2)CC(C3)C1
 ADMB	O=C(O[*])C(C)(C)CCOC(C)=O
 Adoc	O=C([*])OC12CC3CC(C2)CC(C3)C1
 Adpoc	O=C([*])OC(C)(C12CC3CC(C2)CC(C3)C1)C
-Alloc O=C([*])OCC=C
+Alloc	O=C([*])OCC=C
 AOC	O=C([*])OCC=C
-Bns O=S(CC1=CC=CC=C1)([O*])=O
+Bns	O=S(CC1=CC=CC=C1)([O*])=O

lint.py

@@ -0,0 +1,24 @@
+# /// script
+# requires-python = ">=3.10"
+# dependencies = [
+#     "pandas",
+# ]
+# ///
+
+from pathlib import Path
+
+import pandas as pd
+
+HERE = Path(__file__).parent.resolve()
+
+
+def main():
+    smi_path = HERE.joinpath("abbreviations.smi")
+    df2 = pd.read_csv(smi_path, sep="\t")
+
+    if smi_path.read_text() != df2.to_csv(sep='\t', index=False):
+        raise ValueError(f"{smi_path} is not formatted properly.")
+
+
+if __name__ == '__main__':
+    main()