Update app.py

app.py

@@ -60,19 +60,38 @@ def predict(query):
     return real
 
 def findRealProb(text):
-    chunksOfText = (chunks_of_900(text))
+    data = request.json.get('text')
+    if data is None or len(data) == 0:
+        return jsonify({'error': 'No query provided'}), 400
+    if len(data) > 9400:
+        return jsonify({'error': 'Cannot analyze more than 9400 characters!'}), 400
+    if count_words(data) > 1500:
+        return jsonify({'error': 'Cannot analyze more than 1500 words'}), 400
+    
+    # return {"Real": predict(data)}
+    chunksOfText = (chunks_of_900(data))
     results = []
     for chunk in chunksOfText:
-        output = predict(chunk)
-        results.append([output, len(chunk)])
-    
+        outputv1 = predict(chunk, model, tokenizer)
+        # outputv2 = predict(chunk, modelv2, tokenizerv2)
+        label = "CG"
+        if(outputv1>=0.5):
+            label = "OR"
+        results.append({"Text":chunk, "Label": label, "Confidence":(outputv1)})    
     ans = 0
     cnt = 0
-    for prob, length in results:
+    for result in results:
+        length = len(result["Text"])
+        confidence = result["Confidence"]
         cnt += length
-        ans = ans + prob*length
+        ans = ans + (confidence)*(length)
     realProb = ans/cnt
-    return {"Real": realProb, "Fake": 1-realProb}, results
+    label = "AI"
+    if realProb > 0.7:
+        label = "Human"
+    elif realProb > 0.3 and realProb < 0.7:
+        label = "Might be AI"
+    return jsonify({"Real": realProb, "Fake": 1-realProb, "Label": label,  "Chunks": results})
 
 demo = gr.Interface(
         fn=findRealProb,