change

.gitignore

@@ -0,0 +1 @@
+__pycache__

generate.py

@@ -47,6 +47,7 @@ class GenerateRunner():
         self.exist_flag = Path(f'{self.save_path}/generated_molecules.csv').exists()
         self.overwrite = opt.overwrite
         self.dev_no = opt.dev_no
+        self.device = torch.device('cpu')
         global LOG
         LOG = ul.get_logger(name="generate",
                             log_path=os.path.join(self.save_path, 'generate.log'))
@@ -59,6 +60,17 @@ class GenerateRunner():
         self.vocab = vocab
         self.tokenizer = mv.SMILESTokenizer()
 
+        # 加载模型
+        file_name = os.path.join(opt.model_path, f'model_{opt.epoch}.pt')
+        if opt.model_choice == 'transformer':
+            self.model = EncoderDecoder.load_from_file(file_name)
+            self.model.to(self.device)
+            self.model.eval()
+        elif opt.model_choice == 'seq2seq':
+            self.model = Model.load_from_file(file_name, evaluation_mode=True)
+            self.model.network.encoder.to(self.device)
+            self.model.network.decoder.to(self.device)
+
     def initialize_dataloader(self, opt, vocab, test_file):
         """
         Initialize dataloader
@@ -84,7 +96,10 @@ class GenerateRunner():
         # torch.cuda.set_device(1)
         # current_device = torch.cuda.current_device()
         # print("当前使用的 CUDA 设备编号是:", current_device)
-        device = torch.device(f'cuda:{self.dev_no}')
+        # device = torch.device(f'cuda:{self.dev_no}')
+        device = torch.device('cpu')
+        print(f"-------device:---------")
+        print(device)
         # 构造loader
         dataloader_test = self.initialize_dataloader(opt, self.vocab, opt.test_file_name)
 
@@ -273,12 +288,15 @@ def run_main():
     parser = argparse.ArgumentParser(
         description='generate.py',
         formatter_class=argparse.ArgumentDefaultsHelpFormatter)
-
+    print("--------parser-------------")
+    print(parser)
     opts.generate_opts(parser)
     opt = parser.parse_args()
     opt.test_file_name = prepare_input(opt)
-
+    print("opt输出如下")
+    print(opt)
     runner = GenerateRunner(opt)
+    print()
     runner.generate(opt)
 
 

main.py

@@ -69,19 +69,24 @@ def calculate_descriptors(smiles):
 def run_generate_runner(const_smiles, var_smiles, main_cls, minor_cls, delta_value, num_samples):
     # 初始化生成器的配置选项
     opt = {
-        'model_choice': 'transformer',
-        'model_path': '$(pwd)/raw_pretrain_frag/checkpoint',
-        'vocab_path': '$(pwd)',
-        'epoch': 20,
-        # 'save_directory': '$(pwd)/demo_gen',
-        # 'data_path': '/home/yichao/zhilian/GenAICode/CLModel_v2_zl',
-        # 'test_file_name': 'test_100',
-        'batch_size': num_samples
+        'batch_size': num_samples,
+        'data_path' :  './',
+        'decode_type' :  'multinomial',
+        'dev_no' :  0,
+        'epoch' :  20,
+        'model_choice' :  'transformer',
+        'model_path' :  './raw_pretrain_frag/checkpoint',
+        'num_samples' :  50,
+        'overwrite' :  True,
+        'save_directory' :  './demo_gen',
+        'test_file_name' :  'test_cut',
+        'vocab_path' :  './'
     }
 
     # 将 opt 字典转换为 Options 对象
     opt = Options(**opt)
-
+    print("--------------opt---------------")
+    print(opt)
     runner = GenerateRunner(opt)
 
     # 创建数据
@@ -134,7 +139,7 @@ def run_generate_runner(const_smiles, var_smiles, main_cls, minor_cls, delta_val
     return result
 
 
-@app.get("/fragmentize/", response_model=FragmentResponse)
+@app.get("/fragmentize", response_model=FragmentResponse)
 async def fragmentize(smiles: str = Query(..., description="SMILES string of the molecule")):
     try:
         fragment_df = fragmentize_molecule(smiles)
@@ -147,6 +152,7 @@ async def fragmentize(smiles: str = Query(..., description="SMILES string of the
 async def generate_molecules(request: GenerateRequest):
     try:
         # 调用 SMILES 生成逻辑
+        print("123123")
         result = run_generate_runner(request.constSmiles, request.varSmiles, request.mainCls, request.minorCls, request.deltaValue, request.num)
         return result
     except Exception as e:

test_cut.csv

@@ -0,0 +1,11 @@
+cpd1SMILES,cpd2SMILES,constantSMILES,fromVarSMILES,toVarSMILES,Delta_Value,main_cls,minor_cls,value_type,target_name
+Cc1c(CCC(=O)NC(Cc2ccccc2)C(=O)O)c(=O)oc2cc3oc4c(c3cc12)CCCC4,Cc1c(CCC(=O)Nc2ccc(C(=O)O)cc2)c(=O)oc2cc3oc4c(c3cc12)CCCC4,[*:1]C(=O)O.[*:2]NC(=O)CCc1c(C)c2cc3c4c(oc3cc2oc1=O)CCCC4,[*:1]C([*:2])Cc1ccccc1,[*:1]c1ccc([*:2])cc1,"(0.5, 1.0]",activity,EC50,seq,Nuclear factor NF-kappa-B p105 subunit
+Cc1oc2c(C)c3oc(=O)c(CCC(=O)Nc4cccc(C(=O)O)c4)c(C)c3cc2c1C,Cc1oc2c(C)c3oc(=O)c(CC(=O)Nc4cccc(C(=O)O)c4)c(C)c3cc2c1C,[*:1]C(=O)Nc1cccc(C(=O)O)c1.[*:2]c1c(C)c2cc3c(C)c(C)oc3c(C)c2oc1=O,[*:1]CC[*:2],[*:1]C[*:2],"(-1.0, -0.5]",activity,EC50,seq,Nuclear factor NF-kappa-B p105 subunit
+Cc1nc(Nc2cccc(-c3cccc(-c4nc5cc(CN6CC[C@@H](C(=O)O)C6)cc(C#N)c5o4)c3C)c2C)c2ncc(CN3CC[C@](C)(O)C3)cc2n1,Cc1nc(Nc2cccc(-c3cccc(-c4nc5cc(CN6CCC(C(=O)O)CC6)cc(C#N)c5o4)c3C)c2C)c2ncc(CN3CC[C@](C)(O)C3)cc2n1,[*:1]C(=O)O.[*:2]Cc1cc(C#N)c2oc(-c3cccc(-c4cccc(Nc5nc(C)nc6cc(CN7CC[C@](C)(O)C7)cnc56)c4C)c3C)nc2c1,[*:1][C@@H]1CCN([*:2])C1,[*:1]C1CCN([*:2])CC1,"(-0.5, 0.0]",activity,EC50,seq,Nuclear factor NF-kappa-B p105 subunit
+Cc1c(CCC(=O)Nc2ccc(C(=O)O)cc2)c(=O)oc2cc3occ(C(C)(C)C)c3cc12,Cc1c(CCC(=O)Nc2cccc(C(=O)O)c2)c(=O)oc2cc3occ(C(C)(C)C)c3cc12,[*:2]C(=O)O.[*:1]CCc1c(C)c2cc3c(C(C)(C)C)coc3cc2oc1=O,[*:1]C(=O)Nc1ccc([*:2])cc1,[*:1]C(=O)Nc1cccc([*:2])c1,"(-0.5, 0.0]",activity,EC50,seq,Nuclear factor NF-kappa-B p105 subunit
+Cc1c(CCC(=O)Nc2cccc(C(=O)O)c2)c(=O)oc2cc3occ(C(C)(C)C)c3cc12,Cc1c(CCC(=O)Nc2ccccc2C(=O)O)c(=O)oc2cc3occ(C(C)(C)C)c3cc12,[*:2]C(=O)O.[*:1]CCc1c(C)c2cc3c(C(C)(C)C)coc3cc2oc1=O,[*:1]C(=O)Nc1cccc([*:2])c1,[*:1]C(=O)Nc1ccccc1[*:2],"(-0.5, 0.0]",activity,EC50,seq,Nuclear factor NF-kappa-B p105 subunit
+Cc1c(CCC(=O)Nc2ccc(CC(=O)O)cc2)c(=O)oc2cc3occ(C(C)(C)C)c3cc12,Cc1c(CCC(=O)Nc2cccc(C(=O)O)c2)c(=O)oc2cc3occ(C(C)(C)C)c3cc12,[*:1]C(=O)O.[*:2]NC(=O)CCc1c(C)c2cc3c(C(C)(C)C)coc3cc2oc1=O,[*:1]Cc1ccc([*:2])cc1,[*:1]c1cccc([*:2])c1,"(0.0, 0.5]",activity,EC50,seq,Nuclear factor NF-kappa-B p105 subunit
+Cc1c(CC(=O)Nc2ccc(C(N)=O)cc2)c(=O)oc2c(C)c3oc4c(c3cc12)CCCC4,Cc1c(CC(=O)Nc2ccc(O)cc2)c(=O)oc2c(C)c3oc4c(c3cc12)CCCC4,[*:1]NC(=O)Cc1c(C)c2cc3c4c(oc3c(C)c2oc1=O)CCCC4,[*:1]c1ccc(C(N)=O)cc1,[*:1]c1ccc(O)cc1,"(-0.5, 0.0]",activity,EC50,seq,Nuclear factor NF-kappa-B p105 subunit
+Cc1nc(Nc2cccc(-c3cccc(-c4nc5cc(CN6CCC(C(=O)O)CC6)cc(C#N)c5o4)c3C)c2C)c2ncc(CN3CC[C@@H](O)C3)cc2n1,Cc1c(Nc2nc(CO)nc3cc(CN4CC[C@@H](O)C4)cnc23)cccc1-c1cccc(-c2nc3cc(CN4CCC(C(=O)O)CC4)cc(C#N)c3o2)c1C,[*:1]Cc1nc(Nc2cccc(-c3cccc(-c4nc5cc(CN6CCC(C(=O)O)CC6)cc(C#N)c5o4)c3C)c2C)c2ncc(CN3CC[C@@H](O)C3)cc2n1,[*:1][H],[*:1]O,"(-0.5, 0.0]",activity,EC50,seq,Nuclear factor NF-kappa-B p105 subunit
+Cc1c(CCC(=O)Nc2ccccc2C(=O)O)c(=O)oc2cc3occ(C(C)(C)C)c3cc12,Cc1c(CCC(=O)Nc2cccc(C(=O)O)c2)c(=O)oc2cc3occ(C(C)(C)C)c3cc12,[*:2]C(=O)O.[*:1]CCc1c(C)c2cc3c(C(C)(C)C)coc3cc2oc1=O,[*:1]C(=O)Nc1ccccc1[*:2],[*:1]C(=O)Nc1cccc([*:2])c1,"(-0.5, 0.0]",activity,EC50,seq,Nuclear factor NF-kappa-B p105 subunit
+Cc1c(Nc2nc(C(F)F)nc3cc(CN4CC[C@@H](O)C4)cnc23)cccc1-c1cccc(-c2nc3cc(CN4CC[C@@](C)(C(=O)O)C4)cc(C#N)c3o2)c1C,Cc1c(Nc2nc(C(F)F)nc3cc(CN4CC[C@@H](O)C4)cnc23)cccc1-c1cccc(-c2nc3cc(CN4CCC(C)(C(=O)O)CC4)cc(C#N)c3o2)c1C,[*:2]C(=O)O.[*:1]Cc1cc(C#N)c2oc(-c3cccc(-c4cccc(Nc5nc(C(F)F)nc6cc(CN7CC[C@@H](O)C7)cnc56)c4C)c3C)nc2c1,[*:1]N1CC[C@]([*:2])(C)C1,[*:1]N1CCC([*:2])(C)CC1,"(-0.5, 0.0]",activity,EC50,seq,Nuclear factor NF-kappa-B p105 subunit