docs(app.py): :memo: just add documentation

app.py

@@ -12,13 +12,28 @@ tokenizer = AutoTokenizer.from_pretrained("indolem/indobert-base-uncased")
 model = AutoModel.from_pretrained("indolem/indobert-base-uncased")
 
 # Mapping dictionaries for labels
-priority_score_mapping = {1: "low", 2: "medium", 3: "high"}
-problem_domain_mapping = {0: "operational", 1: "tech"}
+priority_score_mapping = {1: "LOW", 2: "MEDIUM", 3: "HIGH"}
+problem_domain_mapping = {0: "OPERATIONAL", 1: "TECHNICAL"}
 
 # Load the trained Random Forest models
 best_classifier1 = joblib.load('best_classifier1_optimized.pkl')
 best_classifier2 = joblib.load('best_classifier2_optimized.pkl')
 
+markdown_text = '''
+## Label Description
+
+### Priority Score
+* **Low** label, means that the temporary/corrective solution can solve the problem. A permanent solution will be provided later because the impact on the business can still be handled.
+* **Medium** label, means that there's a need to determine the time constraint to solve the problem. If it remains too long, it will impact the business side.
+* **High** label, means that the problem is urgent and must be solved immediately.
+
+### Problem Domain
+* **Operational** label, means that the scope of the problem is on the business or daily operational.
+* **Technical** label, means that the scope of the problem is on the technical (technology) side like the mobile/web application.
+'''
+
+description="Write the feedback about the capsule hotel that you've ever visited or stayed there."
+
 # Function to perform predictions
 def predict(text):
     # Convert the sentences into input features
@@ -59,4 +74,4 @@ def predict(text):
 
 
 # Create a Gradio interface
-gr.Interface(fn=predict, inputs="text", outputs="text", title="Simple Risk Classifier Demo").launch(debug=True)
+gr.Interface(fn=predict, inputs="text", outputs="text", title="Simple Risk Classifier Demo (Case Study: Capsule Hotel)", description=description, article=markdown_text).launch(debug=True)