Update app.py

app.py

@@ -1,79 +1,42 @@
 from transformers import pipeline
-import gradio as gr
-
-# Load model and tokenizer
-classifier = pipeline(
-    "text-classification",
-    model="WarTitan2077/Number-Classifier",
-    tokenizer="WarTitan2077/Number-Classifier",
-    top_k=None
-)
-
-# Map Hugging Face's LABEL_X to human-readable labels
-label_map = {
-    "LABEL_0": "Symbolic1",
-    "LABEL_1": "Numeric1",
-    "LABEL_2": "Natural1",
-    "LABEL_3": "Integer1",
-    "LABEL_4": "Rational1",
-    "LABEL_5": "Irrational1",
-    "LABEL_6": "Real1",
-    "LABEL_7": "Prime1",
-    "LABEL_8": "Composite1",
-    "LABEL_9": "Symbolic2",
-    "LABEL_10": "Numeric2",
-    "LABEL_11": "Natural2",
-    "LABEL_12": "Integer2",
-    "LABEL_13": "Rational2",
-    "LABEL_14": "Irrational2",
-    "LABEL_15": "Real2",
-    "LABEL_16": "Prime2",
-    "LABEL_17": "Composite2",
-    "LABEL_18": "Symbolic3",
-    "LABEL_19": "Numeric3",
-    "LABEL_20": "Natural3",
-    "LABEL_21": "Integer3",
-    "LABEL_22": "Rational3",
-    "LABEL_23": "Irrational3",
-    "LABEL_24": "Real3",
-    "LABEL_25": "Prime3",
-    "LABEL_26": "Composite3"
-}
-
-def classify_numbers(input1, input2, input3):
-    inputs = [input1, input2, input3]
-    predictions = {}
-
-    for i, value in enumerate(inputs):
-        result = classifier(value)[0]  # full list of label scores
-        number_index = str(i + 1)
-
-        input_result = {}
-
-        for item in result:
-            label = label_map[item["label"]]
-            if label.endswith(number_index):
-                input_result[label] = round(item["score"], 3)
-
-        # ✅ Sort the results by score (highest to lowest)
-        input_result = dict(
-            sorted(input_result.items(), key=lambda x: x[1], reverse=True)
-        )
-
-        predictions[f"Input{number_index}"] = input_result
-
-    return predictions
-
-
-
-
-# Gradio UI
-iface = gr.Interface(
-    fn=classify_numbers,
-    inputs=["text", "text", "text"],
-    outputs="json",
-    title="Number Classifier",
-    description="Enter three numbers or expressions to classify them (e.g., 2, pi, 3.5)"
-)
-
-iface.launch()
+from fastapi import FastAPI, Request
+from pydantic import BaseModel
+import uvicorn
+
+# Load classifier
+classifier = pipeline("text-classification", model="WarTitan2077/Number-Classifier", tokenizer="WarTitan2077/Number-Classifier", top_k=None)
+
+# Define labels (same as during training)
+labels = ["Symbolic", "Numeric", "Natural", "Integer", "Rational", "Irrational", "Real", "Prime", "Composite"]
+
+# Map label IDs back to names
+label_map = {f"LABEL_{i}": label for i, label in enumerate(labels)}
+
+# Input schema
+class Inputs(BaseModel):
+    Input1: str
+    Input2: str
+    Input3: str
+
+# App
+app = FastAPI()
+
+@app.post("/predict")
+async def predict(data: Inputs):
+    inputs = {
+        "Input1": data.Input1,
+        "Input2": data.Input2,
+        "Input3": data.Input3
+    }
+
+    response = {}
+
+    for key, value in inputs.items():
+        if value.strip():  # only classify non-empty inputs
+            result = classifier(value)
+            preds = {label_map[item["label"]]: round(item["score"], 3) for item in result[0]}
+            response[key] = preds
+        else:
+            response[key] = {}
+
+    return response