Paraphrasing, Summarizing Added

GrammarSummary/new_model.weights.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e1246b81386e12ce55eb44edc86a97d9d892d7b5adf4222c502fab8b801805e
+size 1673753584

Paraphrase/new_model.weights.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b6ab0eb6c02bd23771aa31e775c848bf3929ef11c34b8f23f788ec4eb953ae0
+size 1673753584

app.py

@@ -1,8 +1,8 @@
 import streamlit as st
-from textSFunctionality import generateText
+from textFunctionality import generateText, modelWG, modelWOG, modelParaphrase
 
 # Set the page configuration and theme once at the top
-st.set_page_config(page_title="Text Summarization", page_icon="⭐")
+st.set_page_config(page_title="Text Summarization and Paraphrasing", page_icon="⭐")
 
 st.write(
     """
@@ -35,19 +35,38 @@ st.write(
 )
 
 def main():
-    st.title('Text Summarization')
+    st.title('Text Summarization And Paraphrasing')
+    st.write("**Summarize Without Grammar Performs Better But Misses Out On Grammar Like Punctuation, Capitalization, etc.**")
 
     # Text area for user input
-    user_input = st.text_area("#### **Enter Text To Summarize**:", height=300)
+    user_input = st.text_area("#### **Enter Text To Summarize or Paraphrase**:", height=300)
 
-    # Button to trigger summarization
-    if st.button("Summarize"):
+    # Button to trigger summarization with grammar
+    if st.button("Summarize With Grammar"):
         if user_input:
-            summary = generateText(user_input)
-            st.write("#### **Summarized Text**:")
+            summary = generateText(user_input, modelWG, 200, False)
+            st.write("#### **Summarized Text (With Grammar)**:")
             st.write(summary)
         else:
-            st.write("Please Enter Some Text To Summarize.")
+            st.write("**Please Enter Some Text To Summarize.**")
+
+    # Button to trigger summarization without grammar
+    if st.button("Summarize Without Grammar"):
+        if user_input:
+            summary = generateText(user_input, modelWOG, 200, True)
+            st.write("#### **Summarized Text (Without Grammar)**:")
+            st.write(summary)
+        else:
+            st.write("**Please Enter Some Text To Summarize.**")
+
+    # Button to trigger paraphrasing
+    if st.button("Paraphrase"):
+        if user_input:
+            paraphrase = generateText(user_input, modelParaphrase, 500, False)
+            st.write("#### **Paraphrased Text**:")
+            st.write(paraphrase)
+        else:
+            st.write("**Please Enter Some Text To Paraphrase.**")
 
 if __name__ == '__main__':
-    main()
+    main()

textFunctionality.py

@@ -0,0 +1,50 @@
+import re
+import os
+import tensorflow as tf
+import keras
+import keras_nlp
+
+# Constants
+MAX_ENCODER_SEQUENCE_LENGTH = 512
+MAX_DECODER_SEQUENCE_LENGTH_FOR_SUMMARY = 128
+MAX_DECODER_SEQUENCE_LENGTH_FOR_PARAPHRASE = 512
+
+
+# Model Files
+MODEL_PATH = r"ModelFiles"
+
+
+# "WithoutGrammarSummary" Model
+WEIGHT_PATH_WO_G = r"WithoutGrammarSummary/new_model.weights.h5"
+WEIGHT_PATH_W_G = r"GrammarSummary/new_model.weights.h5"
+WEIGHT_PATH_PARAPHRASE = r"Paraphrase/new_model.weights.h5"
+ 
+def cleanText(text):
+    text = str(text)
+    text = re.sub(r'[^a-zA-Z0-9\s]', '', text)
+    text = text.lower()
+    return text
+
+# Preprocessor For Summary
+preprocessorForSummary = keras_nlp.models.BartSeq2SeqLMPreprocessor.from_preset(MODEL_PATH, encoder_sequence_length=MAX_ENCODER_SEQUENCE_LENGTH,decoder_sequence_length=MAX_DECODER_SEQUENCE_LENGTH_FOR_SUMMARY,)
+preprocessorForParaphrase = keras_nlp.models.BartSeq2SeqLMPreprocessor.from_preset(MODEL_PATH, encoder_sequence_length=MAX_ENCODER_SEQUENCE_LENGTH,decoder_sequence_length=MAX_DECODER_SEQUENCE_LENGTH_FOR_PARAPHRASE,)
+
+# Load Model
+
+# "WithoutGrammarSummary" Model
+modelWOG = keras_nlp.models.BartSeq2SeqLM.from_preset(MODEL_PATH, preprocessor=preprocessorForSummary)
+modelWOG.load_weights(WEIGHT_PATH_WO_G)
+
+# "WithGrammarSummary" Model
+modelWG = keras_nlp.models.BartSeq2SeqLM.from_preset(MODEL_PATH, preprocessor=preprocessorForSummary)
+modelWG.load_weights(WEIGHT_PATH_W_G)
+
+# "Paraphrase" Model
+modelParaphrase = keras_nlp.models.BartSeq2SeqLM.from_preset(MODEL_PATH, preprocessor=preprocessorForParaphrase)
+modelParaphrase.load_weights(WEIGHT_PATH_W_G)
+
+def generateText(input_text, model, max_length, wo_summary):
+    if wo_summary:
+        input_text = cleanText(input_text)
+    output = model.generate(input_text, max_length=max_length)
+    return output

textSFunctionality.py

@@ -1,27 +0,0 @@
-import re
-import os
-import tensorflow as tf
-import keras
-import keras_nlp
-
-MAX_ENCODER_SEQUENCE_LENGTH = 512
-MAX_DECODER_SEQUENCE_LENGTH = 128
-
-MODEL_PATH = r"TextSummarizationModel"
-WEIGHT_PATH = r"new_model.weights.h5"
-WEIGHT_PATH = os.path.join(MODEL_PATH, WEIGHT_PATH)
-
-def cleanText(text):
-    text = str(text)
-    text = re.sub(r'[^a-zA-Z0-9\s]', '', text)
-    text = text.lower()
-    return text
-
-preprocessor = keras_nlp.models.BartSeq2SeqLMPreprocessor.from_preset(MODEL_PATH, encoder_sequence_length=MAX_ENCODER_SEQUENCE_LENGTH,decoder_sequence_length=MAX_DECODER_SEQUENCE_LENGTH,)
-model = keras_nlp.models.BartSeq2SeqLM.from_preset(MODEL_PATH, preprocessor=preprocessor)
-model.load_weights(WEIGHT_PATH)
-
-def generateText(input_text, model=model, max_length=200):
-    input_text = cleanText(input_text)
-    output = model.generate(input_text, max_length=max_length)
-    return output