Adds more access to initial solutions in ehrlich-holo (#279)

* Adds more access to initial solutions of ehrlich holo

* Adds dist to gitignore

* Updates config to specifiy python version

.gitignore

@@ -6,6 +6,7 @@ src/poli.egg-info/
 temp
 build
 .tox
+dist/
 
 # Ignore the registered objectives
 src/poli/registered_objectives/*.sh

setup.cfg

@@ -1,8 +1,9 @@
 [metadata]
 name = poli
 version = "1.0.0.dev11"
+author = Miguel González-Duque
 author_email = [email protected]
-description = Protein Objectives Library
+description = A library of discrete objective functions
 long_description = file: README.md
 long_description_content_type = text/markdown
 classifiers =
@@ -14,7 +15,7 @@ classifiers =
 package_dir =
     =src
 packages = find:
-python_requires = >=3.8
+python_requires = >=3.9
 include_package_data = True
 
 [options.packages.find]

src/poli/objective_repository/ehrlich_holo/isolated_function.py

@@ -65,16 +65,18 @@ def __init__(
     def __call__(self, x: np.ndarray, context: None) -> np.ndarray:
         # First, we transform the strings into integers using the alphabet
         batch_size = x.shape[0]
+        if len(x.shape) > 1:
+            # Flattening each element
+            x = np.array(["".join(x_i) for x_i in x])
         x_ = np.array([[self.alphabet.index(c) for c in s] for s in x.flatten()])
 
         values = self.inner_ehrlich(torch.from_numpy(x_)).numpy(force=True)
         values[values == -np.inf] = self.return_value_on_unfeasible
 
         return values.reshape(batch_size, 1)
 
-    @property
-    def initial_solution(self):
-        return self.inner_ehrlich.initial_solution()
+    def initial_solution(self, n_samples: int = 1):
+        return self.inner_ehrlich.initial_solution(n=n_samples)
 
     @property
     def optimal_solution(self):

src/poli/objective_repository/ehrlich_holo/register.py

@@ -147,12 +147,24 @@ def __init__(
             evaluation_budget=evaluation_budget,
         )
 
-    def initial_solution(self) -> np.ndarray:
+    def initial_solution(self, n_samples: int = 1) -> np.ndarray:
         # This is a sequence of ints.
-        initial_solution_as_ints = self.inner_function.initial_solution
+        initial_solution_as_ints = self.inner_function.initial_solution(
+            n_samples=n_samples
+        )
 
         # We convert it to a sequence of strings.
-        return np.array(["".join([self.alphabet[i] for i in initial_solution_as_ints])])
+        if n_samples == 1:
+            return np.array(
+                ["".join([self.alphabet[i] for i in initial_solution_as_ints])]
+            )
+        else:
+            return np.array(
+                [
+                    "".join([self.alphabet[i] for i in z_i])
+                    for z_i in initial_solution_as_ints
+                ]
+            )
 
     def random_solution(self) -> np.ndarray:
         random_solution_as_ints = self.inner_function.random_solution