Spaces:
Running
Running
| import streamlit as st | |
| import difflib | |
| from sentence_transformers import SentenceTransformer | |
| from sklearn.feature_extraction.text import TfidfVectorizer | |
| from sklearn.metrics.pairwise import cosine_similarity | |
| from xhtml2pdf import pisa | |
| import base64 | |
| import os | |
| from io import BytesIO | |
| # Load SBERT model | |
| sbert_model = SentenceTransformer('all-MiniLM-L6-v2') | |
| def compute_sbert_similarity(text1, text2): | |
| emb1 = sbert_model.encode([text1])[0] | |
| emb2 = sbert_model.encode([text2])[0] | |
| score = cosine_similarity([emb1], [emb2])[0][0] | |
| return score | |
| def compute_tfidf_similarity(text1, text2): | |
| vectorizer = TfidfVectorizer().fit([text1, text2]) | |
| vectors = vectorizer.transform([text1, text2]) | |
| return cosine_similarity(vectors[0:1], vectors[1:2])[0][0] | |
| def html_diff(a, b): | |
| differ = difflib.HtmlDiff() | |
| return differ.make_table(a.splitlines(), b.splitlines(), fromdesc='Original', todesc='Modified', context=True, numlines=2) | |
| def convert_html_to_pdf(source_html): | |
| pdf_file = BytesIO() | |
| pisa_status = pisa.CreatePDF(source_html, dest=pdf_file) | |
| if pisa_status.err: | |
| return None | |
| return pdf_file.getvalue() | |
| def create_download_link(pdf_data, filename="report.pdf"): | |
| b64 = base64.b64encode(pdf_data).decode() | |
| href = f'<a href="data:application/pdf;base64,{b64}" download="{filename}">Download PDF Report</a>' | |
| return href | |
| st.set_page_config(layout="wide") | |
| st.title("Advanced Document Comparison Tool with Semantic and Syntactic Analysis") | |
| col1, col2 = st.columns(2) | |
| with col1: | |
| uploaded_file1 = st.file_uploader("Upload Original Document", type=["txt", "md"]) | |
| with col2: | |
| uploaded_file2 = st.file_uploader("Upload Modified Document", type=["txt", "md"]) | |
| if uploaded_file1 and uploaded_file2: | |
| original_text = uploaded_file1.read().decode("utf-8") | |
| modified_text = uploaded_file2.read().decode("utf-8") | |
| sbert_score = compute_sbert_similarity(original_text, modified_text) | |
| tfidf_score = compute_tfidf_similarity(original_text, modified_text) | |
| html_comparison = html_diff(original_text, modified_text) | |
| st.markdown("### 🔍 Similarity Scores") | |
| st.markdown(f"**SBERT Semantic Similarity:** {sbert_score:.4f}") | |
| st.markdown(f"**TF-IDF Syntactic Similarity:** {tfidf_score:.4f}") | |
| st.markdown("### 📑 Comparison Result") | |
| html_report = f''' | |
| <html> | |
| <head> | |
| <style> | |
| .diff {{ font-family: Courier; border: 1px solid #ccc; overflow-x: scroll; }} | |
| .diff th, .diff td {{ padding: 5px; }} | |
| iframe {{ width: 100%; height: 600px; border: none; }} | |
| </style> | |
| <script> | |
| window.addEventListener("DOMContentLoaded", () => {{ | |
| const iframes = document.querySelectorAll("iframe"); | |
| if (iframes.length === 2) {{ | |
| const syncScroll = (e) => {{ | |
| iframes.forEach((frame) => {{ | |
| if (frame !== e.target) {{ | |
| frame.contentWindow.scrollTo(0, e.target.scrollTop); | |
| }} | |
| }}); | |
| }}; | |
| iframes.forEach((iframe) => {{ | |
| iframe.contentWindow.onscroll = syncScroll; | |
| }}); | |
| }} | |
| }}); | |
| </script> | |
| </head> | |
| <body> | |
| {html_comparison} | |
| </body> | |
| </html> | |
| ''' | |
| st.components.v1.html(html_report, height=700, scrolling=True) | |
| if st.button("Generate PDF Report"): | |
| pdf_bytes = convert_html_to_pdf(html_report) | |
| if pdf_bytes: | |
| st.markdown(create_download_link(pdf_bytes), unsafe_allow_html=True) | |
| else: | |
| st.error("❌ Failed to generate PDF. Check for HTML formatting issues.") | |