Add ./gui to ./pypef (only API usage and QThread block until now) and support for Py3.12

Author: niklases

gui/PyPEFGUIQtWindow.py pypef/gui/PyPEFGUIQtWindow.pygui/PyPEFGUIQtWindow.py renamed to pypef/gui/PyPEFGUIQtWindow.py

@@ -8,8 +8,8 @@
 import os
 from PySide6 import QtCore, QtWidgets
 from PySide6.QtCore import QSize, Signal, QThread
-pypef_root = os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))
-sys.path.append(pypef_root)
+#pypef_root = os.path.abspath(os.path.join(os.path.dirname(__file__), '..', '..'))
+#sys.path.append(pypef_root)
 from pypef import __version__
 from pypef.main import __doc__, run_main, logger, formatter
 from pypef.utils.helpers import get_device, get_vram, get_torch_version, get_gpu_info
@@ -18,7 +18,7 @@
 logger.setLevel(logging.INFO)
 
 
-EXEC_API_OR_CLI = ['cli', 'api'][0]
+EXEC_API_OR_CLI = ['cli', 'api'][1]
 
 
 print(sys.executable)
@@ -939,9 +939,13 @@ def process_finished(self):
         self.version_text.setText("Finished...")
 
 
-if __name__ == "__main__":
+def run_app():
     app = QtWidgets.QApplication([])
     widget = MainWindow()
     widget.resize(800, 600)
     widget.show()
     sys.exit(app.exec())
+
+
+if __name__ == "__main__":
+    run_app()

pyproject.toml

@@ -10,7 +10,7 @@ authors = [
     { name = "Alexander-Maurice Illig" }
 ]
 description = "A tool for performing data-driven protein engineering by building machine learning models from protein variant fitness data."
-requires-python = ">=3.10, <3.13"
+requires-python = ">=3.10"
 license = { text = "CC BY-SA 4.0" }
 keywords = ["Pythonic Protein Engineering Framework"]
 classifiers = [
@@ -19,6 +19,7 @@ classifiers = [
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
     "Topic :: Scientific/Engineering :: Bio-Informatics",
     "Topic :: Scientific/Engineering :: Artificial Intelligence"
 ]
@@ -39,6 +40,8 @@ dependencies = [
     "torch",
     "torch-geometric"
 ]
+[project.optional-dependencies]
+gui = ["pyside6"]
 
 [tool.setuptools.dynamic]
 version = {attr = "pypef.__version__"}
@@ -49,6 +52,8 @@ Homepage = "https://github.com/niklases/PyPEF"
 
 [project.scripts]
 pypef = "pypef.main:run_main"
+pypef-gui = "pypef.gui.PyPEFGUIQtWindow:run_app"
+
 
 [tool.setuptools.packages.find]
 where = ["."]

scripts/ProteinGym_runs/run_performance_tests_proteingym_hybrid_dca_llm_random_modulo_continuous_splits.py

@@ -180,16 +180,15 @@ def compute_performances(mut_data, mut_sep=':', start_i: int = 0, already_tested
                 continue 
 
             ns_y_test = [len(variants)]
-            temp_perfs = []
+            ds = DatasetSplitter(csv_file=csv_path, n_cv=5)
             for train_indices, test_indices in zip(
                 ds.random_splits_train_indices_combined, ds.random_splits_test_indices_combined
             ):
                 try:
-                    ds = DatasetSplitter(csv_file=csv_path, n_cv=5)
-                    x_dca_train, x_dca_test = x_dca[train_indices], x_dca[test_indices]
-                    x_llm_train_prosst, x_llm_test_prosst = x_prosst[train_indices], x_prosst[test_indices]
-                    x_llm_train_esm, x_llm_test_esm = x_esm[train_indices], x_esm[test_indices]
-                    y_train, y_test = fitnesses[train_indices], fitnesses[test_indices]
+                    x_dca_train, x_dca_test = np.asarray(x_dca)[train_indices], np.asarray(x_dca)[test_indices]
+                    x_llm_train_prosst, x_llm_test_prosst = np.asarray(x_prosst)[train_indices], np.asarray(x_prosst)[test_indices]
+                    x_llm_train_esm, x_llm_test_esm = np.asarray(x_esm)[train_indices], np.asarray(x_esm)[test_indices]
+                    y_train, y_test = np.asarray(fitnesses)[train_indices], np.asarray(fitnesses)[test_indices]
                     prosst_lora_model_2 = copy.deepcopy(prosst_lora_model)
                     prosst_optimizer = torch.optim.Adam(prosst_lora_model_2.parameters(), lr=0.0001)
                     esm_lora_model_2 = copy.deepcopy(esm_lora_model)
@@ -244,7 +243,7 @@ def compute_performances(mut_data, mut_sep=':', start_i: int = 0, already_tested
                     }
                 }
                 print(f'Train: {len(np.array(y_train))} --> Test: {len(np.array(y_test))}')
-                if len(y_test) <= 50:
+                if len(y_test) <= 20: # TODO: 50
                     print(f"Only {len(fitnesses)} in total, splitting the data "
                           f"in N_Train = {len(y_train)} and N_Test = {len(y_test)} "
                           f"results in N_Test <= 50 variants - not getting "
@@ -539,7 +538,7 @@ def plot_csv_data(csv, plot_name):
     if not JUST_PLOT_RESULTS:
         compute_performances(
             mut_data=combined_mut_data, 
-            start_i=start_i, 
+            start_i=11, 
             already_tested_is=already_tested_is
         )