fix

app.py

@@ -1,7 +1,7 @@
 import gradio as gr
-from modules.binary_classification import binary_classification_small as binary
-from modules.image_classification import image_classification_small as image
-from modules.multilabel_classification import multi_classification_small as multi
+from modules.binary_classification import binary_classification as binary
+from modules.image_classification import image_classification as image
+from modules.multilabel_classification import multi_classification as multi
 import cv2
 
 image_output = gr.Label()
@@ -14,10 +14,10 @@ def binary_classification(text):
 
 def multi_classification(text):
     if(text != ''):
-        try:
+        # try:
             return multi(text)
-        except:
-            raise gr.Error('Il testo deve essere in inglese!')
+        # except:
+            # raise gr.Error('Il testo deve essere in inglese!')
     else:
         raise gr.Error('Il testo è obbligatorio!')  
     

modules/binary_classification.py

@@ -78,69 +78,7 @@ def predict(model_path, weights_path, tokenizer_path, text, debug) :
 		print('Review: [%s]\nSentiment: %s (%.3f%%)' % (doc, sentiment, percent*100))
 	return doc, sentiment, percent
 
-def binary_classification(request):
-	init()
-	model_custom = False
-	tokenizer_custom = False
-	weights_custom = False
-	try:
-		m = request.files['model']
-		if(utils.allowed_model(m.filename)):
-			model = MODEL + 'upload_' + m.filename
-			m.save(model)
-			model_custom = True
-			print("importato modello " + m.filename)
-		else :
-			raise Exception	
-	except :
-		model = MODEL + 'binary-classification.keras'
-		print("uso modello default")
-	try:
-		w = request.files['weights']
-		if(utils.allowed_model(w.filename)):
-			weights = WEIGHTS + 'upload_' + w.filename
-			w.save(weights)
-			weights_custom = True
-			print("importati pesi " + w.filename)
-		else :
-			raise Exception	
-	except :
-		weights = ''
-		print("uso pesi default")	
-	try:
-		t = request.files['token']
-		if(utils.allowed_tokenizer(t.filename)):
-			tokenizer = TOKEN + 'upload_' + t.filename
-			t.save(tokenizer)
-			tokenizer_custom = True
-			print("importato tokenizer " + t.filename)
-		else :
-			raise Exception	
-	except :
-		tokenizer = TOKEN + 'binary-classification-tokenizer.json'
-		print("uso tokenizer default")
-	try:
-		text = request.form['text']
-		if(text == "") : raise Exception
-	except :
-		return {"error": "Sentence is required"}, 415
-	try:
-		doc, sentiment, percent = predict(model, weights, tokenizer, text, True)
-		response = {
-				"lemma" : doc,
-				"sentiment" : sentiment,
-				"percent" : str(percent * 100)
-		}
-		return jsonify(response)
-	finally:	
-		if(model_custom) :
-			os.remove(model)
-		if(tokenizer_custom) :
-			os.remove(tokenizer)
-		if(weights_custom) :
-			os.remove(weights)
-
-def binary_classification_small(text):
+def binary_classification(text):
 	init()
 	model = MODEL + 'binary-classification.h5'
 	weights = ''

modules/image_classification.py

@@ -23,40 +23,7 @@ def image_predict(image_path, model_path):
     pred_labels = np.argmax(predictions, axis= 1)
     return predictions, pred_labels
 
-def image_classification(request):
-    model_custom = False
-    try:
-        m = request.files['model']
-        if(allowed_model(m.filename)):
-            model = MODEL + 'upload_' + m.filename
-            m.save(model)
-            model_custom = True
-            print("importato modello " + m.filename)
-        else :
-            raise Exception	
-    except :
-        model = MODEL + 'medical-image-classification.keras'
-        print("uso modello default ")
-    try:
-        image = request.files['image']
-        image.save(image.filename)
-    except :
-        return {"error": "Image is required"}, 415
-    try:
-        labels, max_label = image_predict(image.filename, model)
-        response = []
-        for i, label in enumerate(labels[0]):
-            response.append({
-                "classe" : class_names[i],
-                "percent" : "%.8f" % float(label*100)
-            })
-        return jsonify(response)
-    finally:	
-        if(model_custom) :
-            os.remove(model)
-        os.remove(image.filename)	
-
-def image_classification_small(numpy_image):
+def image_classification(numpy_image):
     model = MODEL + 'medical-image-classification.keras' 
     myuuid = uuid.uuid4()
     filename = 'test_' + str(myuuid) + '.png'

modules/multilabel_classification.py

@@ -35,55 +35,7 @@ def predict(model_path, tokenizer_path, sentence):
     predicted_label = class_names[np.argmax(prediction[0])]
     return prediction, predicted_label
 
-def multi_classification(request):
-    model_custom = False
-    tokenizer_custom = False
-    try:
-        m = request.files['model']
-        if(utils.allowed_model(m.filename)):
-            model = MODEL + 'upload_' + m.filename
-            m.save(model)
-            model_custom = True
-            print("importato modello " + m.filename)
-        else :
-            raise Exception	
-    except :
-        model = MODEL + 'multi-classification.keras'
-        print("uso modello default ")
-    try:
-        t = request.files['token']
-        if(utils.allowed_tokenizer(t.filename)):
-            tokenizer = TOKEN + 'upload_' + t.filename
-            t.save(tokenizer)
-            tokenizer_custom = True
-            print("importato tokenizer " + t.filename)
-        else :
-            raise Exception	
-    except :
-        tokenizer = TOKEN + 'multi-classification-tokenizer.json'
-        print("uso tokenizer default")
-    try:
-        text = request.form['text']
-        if(text == "") : raise Exception
-    except :
-        return {"error": "Sentence is required"}, 415
-    try:
-        labels, max_label = predict(model, tokenizer, text)
-        response = []
-        for i, label in enumerate(labels[0]):
-            response.append({
-                "classe" : class_names[i],
-                "percent" : "%.8f" % float(label*100)
-            })
-        return jsonify(response)
-    finally:	
-        if(model_custom) :
-            os.remove(model)
-        if(tokenizer_custom) :
-            os.remove(tokenizer)
-        os.remove('temp.txt')
-
-def multi_classification_small(text):
+def multi_classification(text):
     model = MODEL + 'multi-classification.h5'
     tokenizer = TOKEN + 'multi-classification-tokenizer.json'
     try: