Adding FastapiEndpoint

app.py

@@ -5,11 +5,43 @@ import torch
 import random
 import os
 import numpy as np
+
+
 random.seed(4)
 np.random.seed(4)
 torch.manual_seed(4)
 np.random.seed(4)
 
+
+from fastapi import FastAPI, HTTPException
+import os
+from starlette.middleware.cors import CORSMiddleware
+from pydantic import BaseModel
+import uvicorn
+app = FastAPI()
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+rob_chem_model = ClassificationModel('roberta', 'seyonec/SMILES_tokenized_PubChem_shard00_160k',use_cuda=False ,args={'evaluate_each_epoch':True , 'evaluate_during_training_verbose':True, 'seed':4})
+
+class Query(BaseModel):
+    query :str
+
+@app.post("/ToxicityPrediction")
+async def c(query:Query):
+    try:
+        predictions, raw_outputs = rob_chem_model.predict([str(query.query)])
+        st.write("Received request")
+        return {"prediction":predictions[0]}
+    except Exception as e:
+        raise HTTPException(detail = str(e) , status_code = 500)
+if __name__ == "__main__":
+    uvicorn.run(app, host="0.0.0.0", port=5566)
+
 # Set the Streamlit app title
 st.title("Molecule Toxicity Predictions")
 
@@ -18,7 +50,7 @@ path = 'ToxicityPrediction/Models/transformers/checkpoint-149-epoch-1'
 
 # Load the model from the stage
 #loaded_model = ClassificationModel('roberta', path, use_cuda = False)
-rob_chem_model = ClassificationModel('roberta', 'seyonec/SMILES_tokenized_PubChem_shard00_160k',use_cuda=False ,args={'evaluate_each_epoch':True , 'evaluate_during_training_verbose':True, 'seed':4})
+#rob_chem_model = ClassificationModel('roberta', 'seyonec/SMILES_tokenized_PubChem_shard00_160k',use_cuda=False ,args={'evaluate_each_epoch':True , 'evaluate_during_training_verbose':True, 'seed':4})
 # Predict based on the input
 rob_chem_model.model.eval()
 #target_name= st.text_input('Enter a SMILES string:')
@@ -28,6 +60,7 @@ target_name_list = target_name.splitlines()
 target_name_list = [x.strip() for x in target_name_list]
 predict_toxicity = st.button('Predict Toxicity')
 if predict_toxicity:
-    predictions, raw_outputs = rob_chem_model.predict(target_name_list)
-    df_pred = pd.DataFrame({'Smiles':target_name_list,'predictions': predictions})
-    st.dataframe(df_pred)
+    #predictions, raw_outputs = rob_chem_model.predict(target_name_list)
+    #df_pred = pd.DataFrame({'Smiles':target_name_list,'predictions': predictions})
+    #st.dataframe(df_pred)
+    pass