Updates RaSP to throw an error on invalid inputs (#259)

* Updates the error message to make sure the sequence is in alphabet in RaSP

* Adds a regression test, changes the error to a ValueError

* Updates version

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "poli"
-version = "1.0.0.dev8"
+version = "1.0.0.dev9"
 description = "poli, a library of discrete objective functions"
 readme = "README.md"
 authors = [{name="Miguel González-Duque", email="[email protected]"}, {name="Simon Bartels"}]
@@ -53,7 +53,7 @@ profile = "black"
 exclude = ["src/poli/core/util/proteins/rasp/inner_rasp", "src/poli/objective_repository/gfp_cbas"]
 
 [tool.bumpversion]
-current_version = "1.0.0.dev8"
+current_version = "1.0.0.dev9"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = poli
-version = "1.0.0.dev8"
+version = "1.0.0.dev9"
 author_email = [email protected]
 description = Protein Objectives Library
 long_description = file: README.md

src/poli/__init__.py

@@ -1,6 +1,6 @@
 """poli, a library for discrete black-box objective functions."""
 
-__version__ = "1.0.0.dev8"
+__version__ = "1.0.0.dev9"
 from .core.util.isolation.instancing import instance_function_as_isolated_process
 
 # from .core import get_problems

src/poli/core/util/proteins/rasp/rasp_interface.py

@@ -108,6 +108,28 @@ def __init__(
         self.download_cavity_and_downstream_models(
             verbose=verbose, verify_integrity_of_download=verify_integrity_of_download
         )
+        self.alphabet = [
+            "A",
+            "C",
+            "D",
+            "E",
+            "F",
+            "G",
+            "H",
+            "I",
+            "K",
+            "L",
+            "M",
+            "N",
+            "P",
+            "Q",
+            "R",
+            "S",
+            "T",
+            "V",
+            "W",
+            "Y",
+        ]
 
     def get_and_compile_reduce(self):
         """
@@ -513,28 +535,7 @@ def create_df_structure(
         df_structure_no_mt["pdbid"] = res_info["pdb_id"]
         df_structure_no_mt["chainid"] = res_info["chain_id"]
         df_structure_no_mt["variant"] = res_info["wt_AA"] + res_info["pos"] + "X"
-        aa_list = [
-            "A",
-            "C",
-            "D",
-            "E",
-            "F",
-            "G",
-            "H",
-            "I",
-            "K",
-            "L",
-            "M",
-            "N",
-            "P",
-            "Q",
-            "R",
-            "S",
-            "T",
-            "V",
-            "W",
-            "Y",
-        ]
+        aa_list = self.alphabet
 
         # Considering ALL single mutations in ALL sites.
         wildtype_residue_string = "".join(

src/poli/objective_repository/rasp/isolated_function.py

@@ -313,6 +313,13 @@ def parse_pdb_as_residue_strings(self, pdb_file: Path) -> List[str]:
     def _compute_mutant_residue_string_ddg(
         self, mutant_residue_string: str
     ) -> np.ndarray:
+        for i, char in enumerate(mutant_residue_string):
+            if char not in self.rasp_interface.alphabet:
+                raise ValueError(
+                    f"Invalid residue {char} at position {i} in the mutant "
+                    "residue string. Please make sure that all residues are "
+                    f"in the alphabet: {self.rasp_interface.alphabet}."
+                )
         try:
             (
                 closest_wildtype_pdb_file,

src/poli/tests/registry/proteins/test_rasp.py

@@ -179,3 +179,20 @@ def test_rasp_penalization_works_on_multiple_inputs():
     combination = np.vstack([problem.x0, problematic_x])
     y = f(combination)
     assert y[-1] == -100.0
+
+
+@pytest.mark.poli__rasp
+def test_rasp_fails_on_invalid_amino_acids():
+    problem = objective_factory.create(
+        name="rasp",
+        wildtype_pdb_path=THIS_DIR / "3ned.pdb",
+    )
+    f, x0 = problem.black_box, problem.x0
+
+    # This is an unfeasible mutation, since joining
+    # all the strings would result in a sequence
+    # that is _not_ the same length as the wildtype.
+    problematic_x = x0.copy()
+    problematic_x[0][0] = "B"
+    with pytest.raises(ValueError):
+        f(problematic_x)