Update app.py

app.py

@@ -28,6 +28,7 @@ import os
 from tensorflow.keras.applications.resnet50 import ResNet50,preprocess_input, decode_predictions
 from tensorflow.keras.preprocessing import image
 from sklearn.feature_extraction.text import TfidfVectorizer
+from sklearn.metrics.pairwise import cosine_similarity
 
 
 
@@ -433,24 +434,29 @@ def inference(img, lang):
     #im.save('result.jpg')
     return bounds
 
-def compute_tfidf_embeddings(documents1, documents2):
-    # Combine both lists of words into a single list
-    combined_documents = documents1 + documents2
+def compute_tfidf_embeddings(words_list1, words_list2):
+    # Combine the words from both lists
+    combined_words = words_list1 + words_list2
 
     # Initialize the TF-IDF vectorizer
     vectorizer = TfidfVectorizer()
 
-    # Fit the vectorizer on the combined documents
-    vectorizer.fit(combined_documents)
+    # Compute the TF-IDF matrix
+    tfidf_matrix = vectorizer.fit_transform(combined_words)
 
-    # Transform the documents to TF-IDF embeddings
-    embeddings1 = vectorizer.transform(documents1)
-    embeddings2 = vectorizer.transform(documents2)
+    # Split the matrix into separate parts for the two lists
+    tfidf_matrix_list1 = tfidf_matrix[:len(words_list1)]
+    tfidf_matrix_list2 = tfidf_matrix[len(words_list1):]
 
-    return embeddings1, embeddings2
+    return tfidf_matrix_list1, tfidf_matrix_list2
+
+def compute_cosine_similarity(tfidf_matrix_list1, tfidf_matrix_list2):
+    # Compute the cosine similarity between the two TF-IDF matrices
+    similarity_matrix = cosine_similarity(tfidf_matrix_list1, tfidf_matrix_list2)
+
+    return similarity_matrix
+    
 
-import requests
-import base64
 def extract_eval(Ideal_Answer_Text,Ideal_Answer_Diagram,Submitted_Answer_Text,Submitted_Answer_Diagram):
     # print(image1)
     # ideal_text=extract(Ideal_Answer_Text)
@@ -470,16 +476,18 @@ def extract_eval(Ideal_Answer_Text,Ideal_Answer_Diagram,Submitted_Answer_Text,Su
     print(str(int(float(str(diagram_embed_sim_score).split("[")[2].split("]")[0])*10.0)))
     diagram_1_text=inference(Ideal_Answer_Diagram,['en'])
     diagram_2_text=inference(Submitted_Answer_Diagram,['en'])
-    print(diagram_1_text)
-    print(diagram_2_text)
-    diagram_1_text_embed,diagram_2_text_embed=compute_tfidf_embeddings(diagram_1_text,diagram_2_text)
-    diagram_text_similarity=util.pytorch_cos_sim(diagram_1_text_embed,diagram_2_text_embed)
+    tfidf_matrix_list1, tfidf_matrix_list2 = compute_tfidf_embeddings(diagram_1_text, diagram_2_text)
+    similarity_matrix = compute_cosine_similarity(tfidf_matrix_list1, tfidf_matrix_list2)
     print("Diagram Text Embedding Similarity Score \n")
-    print(str(int(float(str(diagram_text_similarity).split("[")[2].split("]")[0])*10.0)))
+    print(similarity_matrix)
+
     
     
     
 
+print(similarity_matrix)
+    
+
 iface = gr.Interface(fn=extract_eval, 
                      inputs=["image","image","image","image"],
                      outputs=gr.outputs.Textbox(),)