Create app.py

app.py

@@ -0,0 +1,116 @@
+import gradio as gr
+import fitz  # PyMuPDF
+import docx
+import difflib
+from sklearn.feature_extraction.text import TfidfVectorizer
+from sklearn.metrics.pairwise import cosine_similarity
+from sentence_transformers import SentenceTransformer
+import matplotlib.pyplot as plt
+import io
+import textstat
+
+# Model for semantic similarity
+model = SentenceTransformer('all-MiniLM-L6-v2')
+
+def extract_text_from_pdf(file):
+    try:
+        text = ""
+        with fitz.open(stream=file, filetype="pdf") as doc:
+            for page in doc:
+                text += page.get_text()
+        return text.strip()
+    except Exception as e:
+        return ""
+
+def extract_text_from_docx(file):
+    doc = docx.Document(file)
+    return "\n".join([para.text for para in doc.paragraphs if para.text.strip()])
+
+def semantic_similarity(text1, text2):
+    embeddings = model.encode([text1, text2])
+    return float(cosine_similarity([embeddings[0]], [embeddings[1]])[0][0])
+
+def compute_readability_score(text):
+    try:
+        return textstat.flesch_reading_ease(text)
+    except:
+        return 0.0
+
+def compare_handouts(old_file, new_file, lo_file):
+    # Step 1: Extract content
+    old_text = extract_text_from_pdf(old_file)
+    new_text = extract_text_from_pdf(new_file)
+    if not old_text or not new_text:
+        return "❌ Could not extract text from one or both PDFs.", None, None, None, None
+
+    lo_text = extract_text_from_docx(lo_file)
+    lo_list = [lo.strip() for lo in lo_text.split("\n") if lo.strip()]
+    if not lo_list:
+        return "⚠️ No learning outcomes detected.", None, None, None, None
+
+    # Step 2: Semantic LO Matching
+    lo_matches = 0
+    for lo in lo_list:
+        score_old = semantic_similarity(lo, old_text)
+        score_new = semantic_similarity(lo, new_text)
+        if score_new > score_old and score_new > 0.6:
+            lo_matches += 1
+
+    # Step 3: Cosine similarity for content diff
+    tfidf = TfidfVectorizer().fit_transform([old_text, new_text])
+    content_sim = cosine_similarity(tfidf[0:1], tfidf[1:2])[0][0]
+    content_change = round((1 - content_sim) * 100, 2)
+
+    # Step 4: Readability
+    read_old = compute_readability_score(old_text)
+    read_new = compute_readability_score(new_text)
+    read_change = round(read_new - read_old, 2)
+
+    # Step 5: Composite score
+    lo_score = (lo_matches / len(lo_list)) * 100
+    composite_score = round((0.6 * lo_score + 0.3 * content_change + 0.1 * max(read_change, 0)), 2)
+
+    # Step 6: Summary
+    summary = f'''
+📈 Content Change (TF-IDF): {content_change}%
+🧠 LO Alignment: {lo_matches} of {len(lo_list)} matched
+📚 Readability Change: {read_change}
+⭐ Composite Improvement Score: {composite_score}%
+
+🟢 Summary: Based on semantic LO coverage, textual change, and readability,
+the new handout shows {"significant" if composite_score > 50 else "moderate" if composite_score > 25 else "minimal"} improvement.
+'''
+
+    # Step 7: Plot LO Match Bar
+    plt.figure(figsize=(6, 3))
+    plt.bar(["Matched LOs", "Unmatched LOs"], [lo_matches, len(lo_list)-lo_matches], color=["green", "red"])
+    plt.title("Learning Outcome Coverage")
+    plt.ylabel("Number")
+    buf = io.BytesIO()
+    plt.tight_layout()
+    plt.savefig(buf, format='png')
+    plt.close()
+    buf.seek(0)
+
+    return summary, gr.Image(value=buf, format="png"), content_change, lo_score, composite_score
+
+# Gradio interface
+iface = gr.Interface(
+    fn=compare_handouts,
+    inputs=[
+        gr.File(label="Upload OLD Handout (PDF)", type="binary"),
+        gr.File(label="Upload NEW Handout (PDF)", type="binary"),
+        gr.File(label="Upload Learning Outcomes (DOCX)", type="binary")
+    ],
+    outputs=[
+        gr.Textbox(label="📊 Summary"),
+        gr.Image(label="📈 LO Alignment Chart"),
+        gr.Number(label="TF-IDF Content Change (%)"),
+        gr.Number(label="LO Coverage (%)"),
+        gr.Number(label="Composite Score")
+    ],
+    title="📘 Handout Comparison & LO Analysis",
+    description="Upload two versions of a handout and learning outcomes to analyze improvements in content, LO coverage, and readability.",
+)
+
+iface.launch()