Update app.py

app.py

@@ -1,19 +1,15 @@
 import pandas as pd
 import numpy as np
 import streamlit as st
-
 from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline
 
+# Initialize transformers
 model_name = "deepset/roberta-base-squad2"
-
-# a) Get predictions
 nlp = pipeline('question-answering', model=model_name, tokenizer=model_name)
-
-# b) Load model & tokenizer
 model = AutoModelForQuestionAnswering.from_pretrained(model_name)
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 
-
+# Define suspicious words and questions
 suspicious_words = [
     "robbery", "crime", "exchange", "extortion", "threat", "suspicious", "fraud", "laundering",
     "illegal", "contraband", "smuggling", "burglary", "assault", "hijacking", "kidnapping", "ransom",
@@ -29,47 +25,36 @@ suspicious_words = [
     "saboteur", "suicide", "discreet", "hide", "action", "profile", "alert", "vigilant", "clandestine", "riot", "arms", "deal"
 ]
 
+questions = ["What event is going to take place?", "Where is it going to happen", "What time is it going to happen?"]
 
-q = ["","",""]
-a = ["","",""]
-
-
-q[0] = "What event is going to take place?"
-q[1] = "Where is it going to happen"
-q[2] = "What time is it going to happen?"
-
-
-QA_input = [{} for i in range(3)]
-res = [{} for i in range(3)]
+# Initialize Streamlit app
+st.title("Crime Detection App")
 
+# Load data
 df = pd.read_excel('senti.xlsx')
-
 parsed_column = df['sentences'].to_list()
 
-print(parsed_column)
+# Process sentences and store results
+output_data = {'Crime Detected': [], 'Location Detected': [], 'Time Detected': []}
+
 for sentence in parsed_column:
-    for i in range(3):
-      QA_input[i] = {
-        'question': q[i],
-        'context': sentence
-      }
-      res[i] = nlp(QA_input[i])
-      a[i] = res[i]['answer']
-    
-    a1 = a[0].lower()
-    a1s = set(a1.split())
-    sus = set(suspicious_words)
-    cw = a1s.intersection(sus)
-    
-    if len(cw) != 0:
-      st.write("The crime detected is: ",a[0])
-      if len(a[1]) != 0:
-        st.write("The location of crime detected is: ",a[1])
-      elif len(a[1]) == 0:
-        st.write("No location detected")
-      if len(a[2]) != 0:
-        st.write("The time of crime detected is: ",a[2])
-      elif len(a[2]) == 0:
-        st.write("No time detected")
-    elif len(cw) == 0:
-      st.write("No crime detected")
+    answers = nlp(questions, sentence)
+    cw = set(answers[0]['answer'].lower().split()) & set(suspicious_words)
+
+    if cw:
+        output_data['Crime Detected'].append(answers[0]['answer'])
+        output_data['Location Detected'].append(answers[1]['answer'] if answers[1]['answer'] else 'No location detected')
+        output_data['Time Detected'].append(answers[2]['answer'] if answers[2]['answer'] else 'No time detected')
+    else:
+        output_data['Crime Detected'].append('No crime detected')
+        output_data['Location Detected'].append('No location detected')
+        output_data['Time Detected'].append('No time detected')
+
+# Convert data to DataFrame
+output_df = pd.DataFrame(output_data)
+
+# Display results
+st.write(output_df)
+
+# Download button for Excel file
+st.download_button(label="Download Excel", data=output_df.to_excel(), file_name='crime_data_output.xlsx', mime='application/vnd.openxmlformats-officedocument.spreadsheetml.sheet')