feat: initial release of curated portfolio app

README.md

@@ -0,0 +1,20 @@
+---
+title: Version Diffing & Provenance Auditor
+emoji: 📊
+colorFrom: green
+colorTo: cyan
+sdk: gradio
+sdk_version: 4.19.2
+app_file: app.py
+pinned: false
+---
+
+# Version Diffing & Provenance Auditor
+
+A premium computational social science application designed to analyze text version differences, track document provenance, and measure lexical drift across legislative bills, treaties, oral history transcripts, or manuscript revisions.
+
+### Features
+- **Side-by-Side Visual Diffing**: Color-coded HTML grid highlighting insertions (green), modifications (yellow), and deletions (red).
+- **Lexical Drift Analytics**: Computes sequence alignment ratios, Jaccard vocabulary similarity indices, and precise word mutations.
+- **File Upload Support**: Drag-and-drop plain text files or paste text blocks directly.
+- **Headless CPU Optimized**: Pure mathematical comparison algorithms that start in milliseconds.

app.py

@@ -0,0 +1,185 @@
+import gradio as gr
+import difflib
+import re
+
+def clean_and_tokenize(text):
+    # Normalize whitespaces
+    return re.findall(r'\b\w+\b|[^\w\s]', text)
+
+def calculate_drift_metrics(text_a, text_b):
+    words_a = text_a.split()
+    words_b = text_b.split()
+    
+    set_a = set(words_a)
+    set_b = set(words_b)
+    
+    # Jaccard similarity
+    union = len(set_a.union(set_b))
+    intersection = len(set_a.intersection(set_b))
+    jaccard_sim = (intersection / union) * 100 if union > 0 else 100.0
+    
+    # SequenceMatcher
+    matcher = difflib.SequenceMatcher(None, words_a, words_b)
+    ratio = matcher.ratio() * 100
+    
+    # Edits breakdown
+    opcodes = matcher.get_opcodes()
+    insertions = 0
+    deletions = 0
+    replacements = 0
+    
+    for tag, i1, i2, j1, j2 in opcodes:
+        if tag == 'insert':
+            insertions += (j2 - j1)
+        elif tag == 'delete':
+            deletions += (i2 - i1)
+        elif tag == 'replace':
+            replacements += max(i2 - i1, j2 - j1)
+            
+    drift_percent = 100 - ratio
+    
+    return {
+        "jaccard": round(jaccard_sim, 2),
+        "similarity": round(ratio, 2),
+        "drift": round(drift_percent, 2),
+        "insertions": insertions,
+        "deletions": deletions,
+        "replacements": replacements,
+        "words_a": len(words_a),
+        "words_b": len(words_b)
+    }
+
+def generate_word_diff(text_a, text_b):
+    # Clean texts
+    words_a = text_a.splitlines()
+    words_b = text_b.splitlines()
+    
+    # We will generate a beautiful side-by-side or inline HTML diff
+    # difflib.HtmlDiff is nice, but we want a highly premium custom HTML view
+    diff = difflib.HtmlDiff(wrapcolumn=60)
+    html_table = diff.make_table(words_a, words_b, context=True, numlines=3)
+    
+    # Style the table to fit our glowing dark theme
+    styled_html = f"""
+    <style>
+        .diff_table {{ font-family: 'Inter', monospace; font-size: 13px; border-collapse: collapse; width: 100%; color: #e2e8f0; background: #0f172a; border-radius: 8px; overflow: hidden; }}
+        .diff_header {{ background: #1e293b; color: #94a3b8; font-weight: 600; padding: 8px; border-bottom: 2px solid #334155; }}
+        td.diff_header {{ text-align: right; background: #1e293b; color: #64748b; padding: 4px 8px; user-select: none; width: 40px; border-right: 1px solid #334155; }}
+        .diff_next {{ display: none; }}
+        .diff_add {{ background-color: rgba(16, 185, 129, 0.25) !important; color: #34d399 !important; text-decoration: none; }}
+        .diff_chg {{ background-color: rgba(245, 158, 11, 0.25) !important; color: #fbbf24 !important; text-decoration: none; }}
+        .diff_sub {{ background-color: rgba(239, 68, 68, 0.25) !important; color: #f87171 !important; text-decoration: line-through; }}
+        td {{ padding: 4px 12px; word-break: break-all; white-space: pre-wrap; }}
+        tr:hover td {{ background: #1e293b; }}
+    </style>
+    <div style="overflow-x: auto; border-radius: 8px; border: 1px solid #334155;">
+        {html_table}
+    </div>
+    """
+    return styled_html
+
+def handle_compare(text_a, file_a, text_b, file_b):
+    # Retrieve values
+    final_a = text_a
+    if file_a is not None:
+        try:
+            with open(file_a.name, 'r', encoding='utf-8', errors='ignore') as f:
+                final_a = f.read()
+        except Exception as e:
+            final_a = f"Error reading File A: {str(e)}"
+            
+    final_b = text_b
+    if file_b is not None:
+        try:
+            with open(file_b.name, 'r', encoding='utf-8', errors='ignore') as f:
+                final_b = f.read()
+        except Exception as e:
+            final_b = f"Error reading File B: {str(e)}"
+            
+    if not final_a.strip() and not final_b.strip():
+        return "<div style='color: #ef4444; padding: 1rem;'>Please enter or upload text for both versions to compare.</div>", ""
+        
+    # Calculate stats
+    stats = calculate_drift_metrics(final_a, final_b)
+    
+    # HTML stats card
+    stats_html = f"""
+    <div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(180px, 1fr)); gap: 1rem; margin-bottom: 1.5rem;">
+        <div style="background: #1e293b; border: 1px solid #334155; border-radius: 8px; padding: 1rem; text-align: center;">
+            <div style="font-size: 0.8rem; color: #94a3b8; font-weight: 500; text-transform: uppercase; letter-spacing: 0.05em;">Document Similarity</div>
+            <div style="font-size: 2rem; font-weight: 700; color: #10b981; margin-top: 0.25rem;">{stats['similarity']}%</div>
+        </div>
+        <div style="background: #1e293b; border: 1px solid #334155; border-radius: 8px; padding: 1rem; text-align: center;">
+            <div style="font-size: 0.8rem; color: #94a3b8; font-weight: 500; text-transform: uppercase; letter-spacing: 0.05em;">Lexical Drift Spread</div>
+            <div style="font-size: 2rem; font-weight: 700; color: #ef4444; margin-top: 0.25rem;">{stats['drift']}%</div>
+        </div>
+        <div style="background: #1e293b; border: 1px solid #334155; border-radius: 8px; padding: 1rem; text-align: center;">
+            <div style="font-size: 0.8rem; color: #94a3b8; font-weight: 500; text-transform: uppercase; letter-spacing: 0.05em;">Word Additions</div>
+            <div style="font-size: 2rem; font-weight: 700; color: #3b82f6; margin-top: 0.25rem;">+{stats['insertions']}</div>
+        </div>
+        <div style="background: #1e293b; border: 1px solid #334155; border-radius: 8px; padding: 1rem; text-align: center;">
+            <div style="font-size: 0.8rem; color: #94a3b8; font-weight: 500; text-transform: uppercase; letter-spacing: 0.05em;">Word Deletions</div>
+            <div style="font-size: 2rem; font-weight: 700; color: #f43f5e; margin-top: 0.25rem;">-{stats['deletions']}</div>
+        </div>
+    </div>
+    <div style="background: #1e293b; border: 1px solid #334155; border-radius: 8px; padding: 1rem; margin-bottom: 1.5rem; color: #cbd5e1;">
+        <h4 style="margin: 0 0 0.5rem 0; color: #e2e8f0;">Provenance Analysis Summary</h4>
+        <p style="margin: 0; font-size: 0.95rem; line-height: 1.5;">
+            Version A contains <strong>{stats['words_a']}</strong> words, and Version B contains <strong>{stats['words_b']}</strong> words. 
+            There are <strong>{stats['replacements']}</strong> words modified in place. 
+            The vocabulary overlap index (Jaccard similarity) is <strong>{stats['jaccard']}%</strong>.
+            This indicates a <strong>{"high" if stats['drift'] > 30 else "moderate" if stats['drift'] > 10 else "minimal"} degree of document mutation</strong> over successive drafts.
+        </p>
+    </div>
+    """
+    
+    # Generate interactive table
+    diff_table = generate_word_diff(final_a, final_b)
+    
+    return stats_html, diff_table
+
+# Setup CSS styling for premium look
+custom_css = """
+body, .gradio-container { background-color: #0b0f19 !important; font-family: 'Inter', sans-serif !important; }
+h1, h2, h3 { color: #f8fafc !important; }
+.gr-button-primary { background: linear-gradient(135deg, #059669 0%, #10b981 100%) !important; border: none !important; }
+.gr-button-primary:hover { opacity: 0.9 !important; }
+"""
+
+# App Interface
+with gr.Blocks(theme=gr.themes.Soft(primary_hue="emerald", secondary_hue="slate"), css=custom_css) as demo:
+    gr.HTML("""
+    <div style="text-align: center; margin-bottom: 2rem; border-bottom: 1px solid #1e293b; padding-bottom: 1.5rem;">
+        <h1 style="font-size: 2.2rem; font-weight: 800; background: linear-gradient(135deg, #10b981 0%, #3b82f6 100%); -webkit-background-clip: text; -webkit-text-fill-color: transparent;">📊 Version Diffing & Provenance Auditor</h1>
+        <p style="color: #94a3b8; font-size: 1.1rem; margin-top: 0.5rem;">Audit the evolution of policies, legislative bills, manuscripts, or transcripts with statistical drift analytics.</p>
+    </div>
+    """)
+    
+    with gr.Row():
+        with gr.Column(scale=1):
+            gr.Markdown("### 📜 Original Version (Draft A)")
+            text_input_a = gr.Textbox(label="Paste Text A", placeholder="Enter original draft here...", lines=10)
+            file_input_a = gr.File(label="Or Upload Text File A", file_types=[".txt"])
+            
+        with gr.Column(scale=1):
+            gr.Markdown("### 📝 New Version (Draft B)")
+            text_input_b = gr.Textbox(label="Paste Text B", placeholder="Enter revised draft here...", lines=10)
+            file_input_b = gr.File(label="Or Upload Text File B", file_types=[".txt"])
+            
+    compare_btn = gr.Button("🔍 Audit Document Drift & Generate Diff Table", variant="primary")
+    
+    gr.Markdown("### 📈 Analytical Provenance Dashboard & Drift Metrics")
+    stats_output = gr.HTML(value="<div style='color: #64748b;'>Submit documents above to calculate version changes.</div>")
+    
+    gr.Markdown("### 🔍 Color-Coded Structural Difference View")
+    diff_output = gr.HTML(value="<div style='color: #64748b;'>Visual diff will render here. Green represents additions, red represents deletions, and yellow represents replacements.</div>")
+    
+    # Wiring events
+    compare_btn.click(
+        fn=handle_compare,
+        inputs=[text_input_a, file_input_a, text_input_b, file_input_b],
+        outputs=[stats_output, diff_output]
+    )
+
+if __name__ == "__main__":
+    demo.launch()

requirements.txt

@@ -0,0 +1,3 @@
+gradio>=4.0.0
+pandas
+jinja2