add keyphrase extraction

app.py

@@ -3,6 +3,7 @@ import PyPDF2
 from extractive_model import summarize_with_textrank
 from abstractive_model import summarize_with_bart
 from keyword_extraction import extract_keywords
+from keyphrase_extraction import extract_keyphrase
 #from blanc import BlancHelp
 
 # Set page to wide mode
@@ -25,7 +26,7 @@ def main():
 
     # Left column: Radio buttons for summarizer choice
     with col1:
-        radio_options = ['Abstractive', 'Extractive', 'Keyword Extraction']
+        radio_options = ['Abstractive', 'Extractive', 'Keyword Extraction', 'Keyphrase Extraction']
         radio_selection = st.radio("Choose type of summarizer:", radio_options)
 
     # Middle column: Text input and File uploader
@@ -56,11 +57,16 @@ def main():
                 summary = summarize_with_bart(file_content)
                 st.session_state.summary = summary
 
-            # Perform extractive summarization
+            # Perform Keyword Extraction
             if radio_selection == "Keyword Extraction":
                 summary = extract_keywords(file_content)
                 st.session_state.summary = summary
 
+            # Perform Keyphrase Extraction
+            if radio_selection == "Keyphrase Extraction":
+                summary = extract_keyphrase(file_content)
+                st.session_state.summary = summary
+
     # Right column: Displaying text after pressing 'Summarize'
     with col3:
         st.write("Summary:")

keyphrase_extraction.py

@@ -0,0 +1,24 @@
+import spacy
+
+# Load the English language model
+nlp = spacy.load("en_core_web_sm")
+
+# Define a list of obligation words
+obligation_words = ["must", "will", "use", "may", "provides", 'is obliged to', 
+                    'has to', 'needs to', 'is required to',
+                    "shall", "should", "ought to", "required", "obligated", "duty"]
+
+def extract_keyphrase(text):
+    # Parse the input text with SpaCy
+    doc = nlp(text)
+    
+    # Initialize a list to store sentences with obligation words
+    obligation_sentences = []
+    
+    # Iterate through the sentences in the document
+    for sentence in doc.sents:
+        # Check if any of the obligation words appear in the sentence
+        if any(word.text.lower() in obligation_words for word in sentence):
+            obligation_sentences.append(sentence.text)
+    
+    return obligation_sentences