Upload 3 files

Dockerfile

@@ -0,0 +1,36 @@
+# ─────────────────────────────────────────────────────────
+# ValuationAI OCR — Docker Image
+# Model: rasmodev/Handwriting_trocr_model
+# ─────────────────────────────────────────────────────────
+FROM python:3.11-slim
+
+RUN apt-get update && apt-get install -y --no-install-recommends \
+        libgl1-mesa-glx libglib2.0-0 libsm6 libxext6 libxrender-dev curl \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN useradd -m -u 1000 appuser
+WORKDIR /app
+
+RUN pip install --no-cache-dir \
+    torch==2.2.2+cpu torchvision==0.17.2+cpu \
+    --extra-index-url https://download.pytorch.org/whl/cpu
+
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY app.py .
+
+ENV HF_HOME=/app/.cache/huggingface
+RUN mkdir -p /app/.cache/huggingface && chown -R appuser:appuser /app
+
+USER appuser
+EXPOSE 7860
+
+HEALTHCHECK --interval=30s --timeout=10s --start-period=90s \
+    CMD curl -f http://localhost:7860/_stcore/health || exit 1
+
+CMD ["streamlit", "run", "app.py", \
+     "--server.port=7860", \
+     "--server.address=0.0.0.0", \
+     "--server.headless=true", \
+     "--browser.gatherUsageStats=false"]

app.py

@@ -0,0 +1,754 @@
+"""
+Valuation Sheet OCR — Production Streamlit App
+Model: rasmodev/Handwriting_trocr_model
+"""
+import io, time, logging
+import streamlit as st
+import pandas as pd
+from PIL import Image
+
+st.set_page_config(
+    page_title="ValuationAI — OCR",
+    page_icon="📋",
+    layout="wide",
+    initial_sidebar_state="collapsed",
+)
+logging.basicConfig(level=logging.INFO)
+
+# ── Design System ─────────────────────────────────────────────────────────────
+st.markdown("""
+<style>
+@import url('https://fonts.googleapis.com/css2?family=Playfair+Display:wght@400;600;700&family=Outfit:wght@300;400;500;600&display=swap');
+
+*, *::before, *::after { box-sizing: border-box; margin: 0; padding: 0; }
+
+html, body, [class*="css"], .stApp {
+    font-family: 'Outfit', sans-serif;
+    background: #0A0F1E;
+    color: #E8EAF0;
+}
+
+/* ── Hide default Streamlit chrome ── */
+#MainMenu, footer, header { visibility: hidden; }
+.block-container { padding: 0 !important; max-width: 100% !important; }
+
+/* ── Page shell ── */
+.page-shell {
+    min-height: 100vh;
+    background:
+        radial-gradient(ellipse 80% 50% at 10% 0%, rgba(99,102,241,0.18) 0%, transparent 60%),
+        radial-gradient(ellipse 60% 40% at 90% 100%, rgba(16,185,129,0.12) 0%, transparent 55%),
+        #0A0F1E;
+    padding: 2.5rem 3rem 4rem;
+}
+
+/* ── Nav bar ── */
+.nav {
+    display: flex;
+    align-items: center;
+    justify-content: space-between;
+    margin-bottom: 3.5rem;
+    padding-bottom: 1.5rem;
+    border-bottom: 1px solid rgba(255,255,255,0.06);
+}
+.nav-brand {
+    display: flex; align-items: center; gap: 0.75rem;
+}
+.nav-dot {
+    width: 10px; height: 10px; border-radius: 50%;
+    background: #6366F1;
+    box-shadow: 0 0 12px rgba(99,102,241,0.8);
+    animation: pulse 2s infinite;
+}
+@keyframes pulse {
+    0%, 100% { opacity: 1; transform: scale(1); }
+    50%       { opacity: 0.6; transform: scale(0.85); }
+}
+.nav-title {
+    font-family: 'Playfair Display', serif;
+    font-size: 1.35rem;
+    font-weight: 700;
+    letter-spacing: -0.02em;
+    color: #fff;
+}
+.nav-subtitle {
+    font-size: 0.72rem;
+    font-weight: 500;
+    letter-spacing: 0.12em;
+    text-transform: uppercase;
+    color: #6B7280;
+}
+.nav-badge {
+    font-size: 0.72rem;
+    background: rgba(99,102,241,0.15);
+    border: 1px solid rgba(99,102,241,0.3);
+    color: #A5B4FC;
+    padding: 0.3rem 0.8rem;
+    border-radius: 20px;
+    font-weight: 500;
+    letter-spacing: 0.04em;
+}
+
+/* ── Hero ── */
+.hero-section {
+    margin-bottom: 3rem;
+}
+.hero-eyebrow {
+    font-size: 0.72rem;
+    font-weight: 600;
+    letter-spacing: 0.18em;
+    text-transform: uppercase;
+    color: #6366F1;
+    margin-bottom: 0.9rem;
+}
+.hero-headline {
+    font-family: 'Playfair Display', serif;
+    font-size: clamp(2.4rem, 4vw, 3.6rem);
+    font-weight: 700;
+    line-height: 1.1;
+    letter-spacing: -0.03em;
+    color: #fff;
+    margin-bottom: 1rem;
+}
+.hero-headline span {
+    background: linear-gradient(135deg, #6366F1 0%, #10B981 100%);
+    -webkit-background-clip: text;
+    -webkit-text-fill-color: transparent;
+    background-clip: text;
+}
+.hero-desc {
+    font-size: 1.05rem;
+    font-weight: 300;
+    color: #9CA3AF;
+    line-height: 1.7;
+    max-width: 520px;
+}
+
+/* ── Two-col layout ── */
+.two-col {
+    display: grid;
+    grid-template-columns: 1fr 1fr;
+    gap: 1.5rem;
+    margin-bottom: 1.5rem;
+}
+
+/* ── Cards ── */
+.glass-card {
+    background: rgba(255,255,255,0.03);
+    border: 1px solid rgba(255,255,255,0.07);
+    border-radius: 20px;
+    padding: 2rem;
+    backdrop-filter: blur(12px);
+    transition: border-color 0.2s;
+}
+.glass-card:hover { border-color: rgba(99,102,241,0.3); }
+
+.card-label {
+    font-size: 0.68rem;
+    font-weight: 600;
+    letter-spacing: 0.14em;
+    text-transform: uppercase;
+    color: #6B7280;
+    margin-bottom: 1.2rem;
+    display: flex; align-items: center; gap: 0.5rem;
+}
+.card-label::before {
+    content: '';
+    display: inline-block;
+    width: 18px; height: 2px;
+    background: #6366F1;
+    border-radius: 2px;
+}
+
+/* ── Upload zone ── */
+.upload-zone {
+    border: 2px dashed rgba(99,102,241,0.35);
+    border-radius: 16px;
+    padding: 3rem 2rem;
+    text-align: center;
+    background: rgba(99,102,241,0.04);
+    transition: all 0.25s;
+    cursor: pointer;
+    margin-bottom: 1rem;
+}
+.upload-zone:hover {
+    border-color: rgba(99,102,241,0.7);
+    background: rgba(99,102,241,0.08);
+}
+.upload-icon { font-size: 2.5rem; margin-bottom: 0.75rem; }
+.upload-main {
+    font-size: 1rem; font-weight: 500; color: #E8EAF0; margin-bottom: 0.4rem;
+}
+.upload-sub { font-size: 0.8rem; color: #6B7280; }
+.file-chip {
+    display: inline-flex; align-items: center; gap: 0.4rem;
+    background: rgba(16,185,129,0.1);
+    border: 1px solid rgba(16,185,129,0.25);
+    color: #6EE7B7;
+    padding: 0.3rem 0.75rem;
+    border-radius: 20px;
+    font-size: 0.75rem;
+    font-weight: 500;
+    margin: 0.2rem;
+}
+
+/* ── Engine pills ── */
+.engine-row { display: flex; gap: 0.6rem; margin-top: 0.5rem; flex-wrap: wrap; }
+.engine-pill {
+    font-size: 0.75rem; font-weight: 500;
+    padding: 0.35rem 0.9rem; border-radius: 20px;
+    border: 1px solid rgba(255,255,255,0.1);
+    color: #9CA3AF; background: transparent; cursor: pointer;
+    transition: all 0.2s;
+}
+.engine-pill.active {
+    background: rgba(99,102,241,0.2);
+    border-color: rgba(99,102,241,0.5);
+    color: #A5B4FC;
+}
+
+/* ── Run button ── */
+.stButton > button {
+    width: 100% !important;
+    background: linear-gradient(135deg, #6366F1 0%, #4F46E5 100%) !important;
+    color: white !important;
+    border: none !important;
+    border-radius: 14px !important;
+    padding: 1rem 2rem !important;
+    font-family: 'Outfit', sans-serif !important;
+    font-size: 0.95rem !important;
+    font-weight: 600 !important;
+    letter-spacing: 0.02em !important;
+    cursor: pointer !important;
+    transition: all 0.2s !important;
+    box-shadow: 0 4px 24px rgba(99,102,241,0.35) !important;
+    margin-top: 1rem !important;
+}
+.stButton > button:hover {
+    transform: translateY(-1px) !important;
+    box-shadow: 0 8px 32px rgba(99,102,241,0.5) !important;
+}
+.stButton > button:disabled {
+    background: rgba(255,255,255,0.05) !important;
+    color: #4B5563 !important;
+    box-shadow: none !important;
+    transform: none !important;
+    cursor: not-allowed !important;
+}
+
+/* ── Progress ── */
+.stProgress > div > div > div {
+    background: linear-gradient(90deg, #6366F1, #10B981) !important;
+    border-radius: 4px !important;
+}
+.stProgress > div > div {
+    background: rgba(255,255,255,0.06) !important;
+    border-radius: 4px !important;
+    height: 6px !important;
+}
+
+/* ── Stat cards ── */
+.stat-row {
+    display: grid;
+    grid-template-columns: repeat(4, 1fr);
+    gap: 1rem;
+    margin-bottom: 1.5rem;
+}
+.stat-card {
+    background: rgba(255,255,255,0.03);
+    border: 1px solid rgba(255,255,255,0.07);
+    border-radius: 16px;
+    padding: 1.4rem 1.6rem;
+    position: relative;
+    overflow: hidden;
+}
+.stat-card::before {
+    content: '';
+    position: absolute;
+    top: 0; left: 0; right: 0;
+    height: 2px;
+    background: linear-gradient(90deg, #6366F1, #10B981);
+}
+.stat-value {
+    font-family: 'Playfair Display', serif;
+    font-size: 2.2rem;
+    font-weight: 700;
+    color: #fff;
+    line-height: 1;
+    margin-bottom: 0.3rem;
+}
+.stat-label {
+    font-size: 0.72rem;
+    font-weight: 500;
+    letter-spacing: 0.1em;
+    text-transform: uppercase;
+    color: #6B7280;
+}
+
+/* ── Dataframe ── */
+.stDataFrame {
+    border-radius: 16px !important;
+    overflow: hidden !important;
+    border: 1px solid rgba(255,255,255,0.07) !important;
+}
+.stDataFrame iframe { border-radius: 16px !important; }
+
+/* ── Download button ── */
+div[data-testid="stDownloadButton"] button {
+    background: rgba(16,185,129,0.12) !important;
+    border: 1px solid rgba(16,185,129,0.3) !important;
+    color: #6EE7B7 !important;
+    border-radius: 12px !important;
+    padding: 0.75rem 1.5rem !important;
+    font-family: 'Outfit', sans-serif !important;
+    font-weight: 600 !important;
+    font-size: 0.9rem !important;
+    box-shadow: none !important;
+    width: auto !important;
+    transition: all 0.2s !important;
+}
+div[data-testid="stDownloadButton"] button:hover {
+    background: rgba(16,185,129,0.22) !important;
+    border-color: rgba(16,185,129,0.5) !important;
+    transform: translateY(-1px) !important;
+}
+
+/* ── File uploader override ── */
+[data-testid="stFileUploader"] {
+    background: transparent !important;
+}
+[data-testid="stFileUploader"] section {
+    background: rgba(99,102,241,0.04) !important;
+    border: 2px dashed rgba(99,102,241,0.35) !important;
+    border-radius: 16px !important;
+    padding: 2rem !important;
+    transition: all 0.25s !important;
+}
+[data-testid="stFileUploader"] section:hover {
+    border-color: rgba(99,102,241,0.7) !important;
+    background: rgba(99,102,241,0.08) !important;
+}
+[data-testid="stFileUploader"] label { color: #9CA3AF !important; font-size: 0.9rem !important; }
+
+/* ── Selectbox ── */
+[data-testid="stSelectbox"] > div > div {
+    background: rgba(255,255,255,0.04) !important;
+    border: 1px solid rgba(255,255,255,0.1) !important;
+    border-radius: 12px !important;
+    color: #E8EAF0 !important;
+}
+
+/* ── Divider ── */
+.section-divider {
+    height: 1px;
+    background: linear-gradient(90deg, transparent, rgba(255,255,255,0.08), transparent);
+    margin: 2.5rem 0;
+}
+
+/* ── Result header ── */
+.result-header {
+    display: flex; align-items: center; justify-content: space-between;
+    margin-bottom: 1.5rem;
+}
+.result-title {
+    font-family: 'Playfair Display', serif;
+    font-size: 1.5rem;
+    font-weight: 600;
+    color: #fff;
+}
+.result-tag {
+    font-size: 0.72rem;
+    font-weight: 600;
+    letter-spacing: 0.1em;
+    text-transform: uppercase;
+    color: #10B981;
+    background: rgba(16,185,129,0.1);
+    border: 1px solid rgba(16,185,129,0.2);
+    padding: 0.3rem 0.8rem;
+    border-radius: 20px;
+}
+
+/* ── Toast-like success ── */
+.success-banner {
+    background: rgba(16,185,129,0.1);
+    border: 1px solid rgba(16,185,129,0.25);
+    border-radius: 12px;
+    padding: 1rem 1.5rem;
+    display: flex; align-items: center; gap: 0.75rem;
+    margin-bottom: 1.5rem;
+    font-size: 0.9rem;
+    color: #6EE7B7;
+    font-weight: 500;
+}
+
+/* Spinner color */
+.stSpinner > div { border-top-color: #6366F1 !important; }
+</style>
+""", unsafe_allow_html=True)
+
+
+# ── Model loading ─────────────────────────────────────────────────────────────
+@st.cache_resource(show_spinner="Initialising model…")
+def load_model():
+    import torch
+    from transformers import TrOCRProcessor, VisionEncoderDecoderModel
+    MODEL_ID  = "rasmodev/Handwriting_trocr_model"
+    processor = TrOCRProcessor.from_pretrained(MODEL_ID)
+    model     = VisionEncoderDecoderModel.from_pretrained(MODEL_ID)
+    device    = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+    model.to(device).eval()
+    return processor, model, device
+
+@st.cache_resource(show_spinner="Loading EasyOCR…")
+def load_easyocr():
+    import easyocr
+    return easyocr.Reader(["en"], gpu=False)
+
+
+# ── OCR ───────────────────────────────────────────────────────────────────────
+def run_finetuned(img):
+    import torch
+    processor, model, device = load_model()
+    pv = processor(images=img.convert("RGB"), return_tensors="pt").pixel_values.to(device)
+    with torch.no_grad():
+        gen = model.generate(pv, max_new_tokens=64)
+    return processor.batch_decode(gen, skip_special_tokens=True)[0].strip()
+
+def run_easyocr(img):
+    import numpy as np
+    results = load_easyocr().readtext(np.array(img.convert("RGB")), detail=0, paragraph=True)
+    return "\n".join(results).strip()
+
+def ocr_image(img, engine):
+    try:
+        return run_finetuned(img) if engine == "finetuned" else run_easyocr(img)
+    except Exception as e:
+        try: return run_easyocr(img)
+        except: return ""
+
+
+# ── PDF handler ───────────────────────────────────────────────────────────────
+def pdf_to_images(file_bytes, dpi=200, max_pages=None):
+    import fitz
+    doc   = fitz.open(stream=file_bytes, filetype="pdf")
+    mat   = fitz.Matrix(dpi/72, dpi/72)
+    limit = min(len(doc), max_pages) if max_pages else len(doc)
+    pages = []
+    for i in range(limit):
+        pix = doc[i].get_pixmap(matrix=mat, alpha=False)
+        pages.append((i+1, Image.open(io.BytesIO(pix.tobytes("png"))).convert("RGB")))
+    doc.close()
+    return pages
+
+
+# ── Field extraction ──────────────────────────────────────────────────────────
+def extract_fields(text, filename):
+    import re
+    def find(label):
+        m = re.search(label + r'[:\s]+([^\|\n]{1,80})', text, re.IGNORECASE)
+        return m.group(1).strip() if m else ""
+    amounts = re.findall(r'AMT:\s*([\d,]+)', text)
+    dates   = re.search(r'DATE:\s*(\d{4}-\d{2}-\d{2})', text)
+    return {
+        "File":           filename,
+        "Plot Number":    find("PLOT"),
+        "Location":       find("LOC"),
+        "Area":           find("AREA"),
+        "Amount (KES)":   int(amounts[0].replace(",","")) if amounts else None,
+        "Date":           dates.group(1) if dates else None,
+        "VOS":            find("VOS"),
+    }
+
+
+# ── Excel export ──────────────────────────────────────────────────────────────
+def to_excel(records):
+    from openpyxl import load_workbook
+    from openpyxl.styles import Font, PatternFill, Alignment, Border, Side
+    from openpyxl.utils import get_column_letter
+    buf = io.BytesIO()
+    df  = pd.DataFrame(records)
+    df.to_excel(buf, index=False, sheet_name="Valuation Data")
+    buf.seek(0)
+    wb  = load_workbook(buf)
+    ws  = wb.active
+    thin = Side(style="thin", color="E5E7EB")
+    for col_idx, cell in enumerate(ws[1], 1):
+        cell.font      = Font(name="Calibri", bold=True, color="FFFFFF", size=11)
+        cell.fill      = PatternFill("solid", start_color="1E1B4B")
+        cell.alignment = Alignment(horizontal="center", vertical="center")
+        cell.border    = Border(bottom=Side(style="medium", color="6366F1"))
+        ws.column_dimensions[get_column_letter(col_idx)].width = 22
+    ws.row_dimensions[1].height = 32
+    for row in ws.iter_rows(min_row=2, max_row=ws.max_row):
+        for i, cell in enumerate(row):
+            cell.alignment = Alignment(vertical="center", wrap_text=True)
+            cell.border    = Border(
+                left=Side(style="thin", color="F3F4F6"),
+                right=Side(style="thin", color="F3F4F6"),
+                top=Side(style="thin", color="F3F4F6"),
+                bottom=Side(style="thin", color="F3F4F6"),
+            )
+            if cell.row % 2 == 0:
+                cell.fill = PatternFill("solid", start_color="F8F7FF")
+    ws.freeze_panes = "A2"
+    out = io.BytesIO()
+    wb.save(out)
+    return out.getvalue()
+
+
+# ═════════════════════════════════════════════════════════════════════════════
+# SESSION STATE
+# ═════════════════════════════════════════════════════════════════════════════
+for k, v in [("records", []), ("errors", []), ("excel_bytes", None), ("processed", False)]:
+    if k not in st.session_state:
+        st.session_state[k] = v
+
+
+# ═════════════════════════════════════════════════════════════════════════════
+# PAGE SHELL OPEN
+# ═════════════════════════════════════════════════════════════════════════════
+st.markdown('<div class="page-shell">', unsafe_allow_html=True)
+
+# ── Nav ───────────────────────────────────────────────────────────────────────
+st.markdown("""
+<div class="nav">
+  <div class="nav-brand">
+    <div class="nav-dot"></div>
+    <div>
+      <div class="nav-title">ValuationAI</div>
+      <div class="nav-subtitle">Handwriting Recognition System</div>
+    </div>
+  </div>
+  <div class="nav-badge">rasmodev/Handwriting_trocr_model</div>
+</div>
+""", unsafe_allow_html=True)
+
+# ── Hero ──────────────────────────────────────────────────────────────────────
+st.markdown("""
+<div class="hero-section">
+  <div class="hero-eyebrow">AI-Powered Document Intelligence</div>
+  <div class="hero-headline">
+    Extract data from<br><span>handwritten valuations</span>
+  </div>
+  <div class="hero-desc">
+    Upload scanned valuation sheet PDFs and get a clean, structured Excel export
+    in seconds — powered by a fine-tuned TrOCR model.
+  </div>
+</div>
+""", unsafe_allow_html=True)
+
+# ═════════════════════════════════════════════════════════════════════════════
+# UPLOAD + SETTINGS
+# ═════════════════════════════════════════════════════════════════════════════
+col_upload, col_settings = st.columns([3, 2], gap="large")
+
+with col_upload:
+    st.markdown('<div class="glass-card">', unsafe_allow_html=True)
+    st.markdown('<div class="card-label">Upload Documents</div>', unsafe_allow_html=True)
+    uploaded_files = st.file_uploader(
+        "Drop PDF files here",
+        type=["pdf", "png", "jpg", "jpeg", "tiff", "tif", "bmp"],
+        accept_multiple_files=True,
+        label_visibility="collapsed",
+    )
+    if uploaded_files:
+        chips = "".join(
+            f'<span class="file-chip">📄 {f.name[:28]}{"…" if len(f.name)>28 else ""}</span>'
+            for f in uploaded_files
+        )
+        st.markdown(f'<div style="margin-top:1rem">{chips}</div>', unsafe_allow_html=True)
+    st.markdown('</div>', unsafe_allow_html=True)
+
+with col_settings:
+    st.markdown('<div class="glass-card">', unsafe_allow_html=True)
+    st.markdown('<div class="card-label">Processing Settings</div>', unsafe_allow_html=True)
+
+    engine_choice = st.selectbox(
+        "OCR Engine",
+        ["🟢  Fine-tuned Model  (Recommended)", "🔵  EasyOCR  (Fallback)"],
+        label_visibility="visible",
+    )
+    engine = "finetuned" if "Fine-tuned" in engine_choice else "easyocr"
+
+    dpi = st.select_slider(
+        "Scan Resolution (DPI)",
+        options=[150, 200, 250, 300],
+        value=200,
+    )
+
+    max_p = st.number_input("Max pages per PDF  (0 = all)", min_value=0, max_value=100, value=0)
+    max_pages = None if max_p == 0 else int(max_p)
+
+    run_btn = st.button(
+        "⚡  Run OCR & Extract",
+        disabled=not uploaded_files,
+        use_container_width=True,
+    )
+    st.markdown('</div>', unsafe_allow_html=True)
+
+
+# ═════════════════════════════════════════════════════════════════════════════
+# PROCESSING
+# ═════════════════════════════════════════════════════════════════════════════
+if run_btn and uploaded_files:
+    st.session_state.records  = []
+    st.session_state.errors   = []
+    st.session_state.excel_bytes = None
+    st.session_state.processed   = False
+
+    st.markdown('<div class="section-divider"></div>', unsafe_allow_html=True)
+
+    bar    = st.progress(0.0)
+    status = st.empty()
+    t0     = time.time()
+    total  = len(uploaded_files)
+
+    for fi, uf in enumerate(uploaded_files):
+        fname      = uf.name
+        file_bytes = uf.read()
+        bar.progress((fi) / total, text=f"Reading  {fname}…")
+
+        try:
+            ext = fname.lower().rsplit(".", 1)[-1]
+            if ext == "pdf":
+                pages = pdf_to_images(file_bytes, dpi=dpi, max_pages=max_pages)
+            else:
+                img   = Image.open(io.BytesIO(file_bytes)).convert("RGB")
+                pages = [(1, img)]
+
+            for page_no, img in pages:
+                status.markdown(
+                    f'<div style="font-size:0.82rem;color:#9CA3AF;margin-bottom:0.5rem">'
+                    f'Processing  <strong style="color:#A5B4FC">{fname}</strong>'
+                    f'  —  page {page_no} of {len(pages)}</div>',
+                    unsafe_allow_html=True,
+                )
+                text   = ocr_image(img, engine)
+                record = extract_fields(text, fname)
+                st.session_state.records.append(record)
+
+        except Exception as e:
+            st.session_state.errors.append(f"{fname}: {e}")
+
+        bar.progress((fi + 1) / total)
+
+    elapsed = time.time() - t0
+    bar.empty()
+    status.empty()
+
+    if st.session_state.records:
+        st.session_state.excel_bytes = to_excel(st.session_state.records)
+    st.session_state.processed = True
+
+    st.success(f"✅  Processed {len(st.session_state.records)} page(s) from {total} file(s) in {elapsed:.1f}s")
+
+
+# ═════════════════════════════════════════════════════════════════════════════
+# RESULTS
+# ═════════════════════════════════════════════════════════════════════════════
+if st.session_state.processed and st.session_state.records:
+    records = st.session_state.records
+    df      = pd.DataFrame(records)
+
+    st.markdown('<div class="section-divider"></div>', unsafe_allow_html=True)
+
+    # ── Stats ──────────────────────────────────────────────────────────────
+    n_plots   = df["Plot Number"].notna().sum() if "Plot Number" in df else 0
+    n_amounts = df["Amount (KES)"].notna().sum() if "Amount (KES)" in df else 0
+    n_dates   = df["Date"].notna().sum() if "Date" in df else 0
+
+    st.markdown(f"""
+    <div class="stat-row">
+      <div class="stat-card">
+        <div class="stat-value">{len(records)}</div>
+        <div class="stat-label">Pages Processed</div>
+      </div>
+      <div class="stat-card">
+        <div class="stat-value">{n_plots}</div>
+        <div class="stat-label">Plot Numbers</div>
+      </div>
+      <div class="stat-card">
+        <div class="stat-value">{n_amounts}</div>
+        <div class="stat-label">Amounts Extracted</div>
+      </div>
+      <div class="stat-card">
+        <div class="stat-value">{n_dates}</div>
+        <div class="stat-label">Dates Captured</div>
+      </div>
+    </div>
+    """, unsafe_allow_html=True)
+
+    # ── Results header + download ──────────────────────────────────────────
+    r_col1, r_col2 = st.columns([3, 1])
+    with r_col1:
+        st.markdown("""
+        <div class="result-header">
+          <div class="result-title">Extracted Data</div>
+          <div class="result-tag">✓ Ready to export</div>
+        </div>
+        """, unsafe_allow_html=True)
+    with r_col2:
+        if st.session_state.excel_bytes:
+            st.download_button(
+                label="⬇  Download Excel",
+                data=st.session_state.excel_bytes,
+                file_name="valuation_extracted.xlsx",
+                mime="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
+            )
+
+    # ── Dataframe ──────────────────────────────────────────────────────────
+    st.dataframe(
+        df,
+        use_container_width=True,
+        height=min(80 + len(df) * 38, 520),
+        hide_index=True,
+    )
+
+    if st.session_state.errors:
+        st.markdown('<div class="section-divider"></div>', unsafe_allow_html=True)
+        with st.expander(f"⚠  {len(st.session_state.errors)} file(s) had errors"):
+            for e in st.session_state.errors:
+                st.markdown(
+                    f'<div style="font-size:0.85rem;color:#F87171;padding:0.3rem 0">{e}</div>',
+                    unsafe_allow_html=True,
+                )
+
+
+# ═════════════════════════════════════════════════════════════════════════════
+# EMPTY STATE
+# ═════════════════════════════════════════════════════════════════════════════
+if not st.session_state.processed and not uploaded_files:
+    st.markdown('<div class="section-divider"></div>', unsafe_allow_html=True)
+    st.markdown("""
+    <div style="display:grid;grid-template-columns:repeat(3,1fr);gap:1.5rem;margin-top:1rem">
+
+      <div class="glass-card" style="text-align:center;padding:2.5rem 1.5rem">
+        <div style="font-size:2rem;margin-bottom:1rem">📄</div>
+        <div style="font-weight:600;color:#E8EAF0;margin-bottom:0.5rem">Upload PDFs</div>
+        <div style="font-size:0.82rem;color:#6B7280;line-height:1.6">
+          Drag in one or multiple scanned valuation sheet PDFs
+        </div>
+      </div>
+
+      <div class="glass-card" style="text-align:center;padding:2.5rem 1.5rem">
+        <div style="font-size:2rem;margin-bottom:1rem">🤖</div>
+        <div style="font-weight:600;color:#E8EAF0;margin-bottom:0.5rem">AI Reads Fields</div>
+        <div style="font-size:0.82rem;color:#6B7280;line-height:1.6">
+          Fine-tuned TrOCR extracts plot numbers, amounts, dates and locations
+        </div>
+      </div>
+
+      <div class="glass-card" style="text-align:center;padding:2.5rem 1.5rem">
+        <div style="font-size:2rem;margin-bottom:1rem">📊</div>
+        <div style="font-weight:600;color:#E8EAF0;margin-bottom:0.5rem">Download Excel</div>
+        <div style="font-size:0.82rem;color:#6B7280;line-height:1.6">
+          Get a clean, formatted spreadsheet ready for data entry or analysis
+        </div>
+      </div>
+
+    </div>
+    """, unsafe_allow_html=True)
+
+
+# PAGE SHELL CLOSE
+st.markdown('</div>', unsafe_allow_html=True)

requirements.txt

@@ -0,0 +1,14 @@
+streamlit>=1.32.0
+transformers>=4.40.0
+torch>=2.2.0
+torchvision>=0.17.0
+Pillow>=10.0.0
+PyMuPDF>=1.23.0
+pandas>=2.0.0
+openpyxl>=3.1.0
+easyocr>=1.7.0
+opencv-python-headless>=4.9.0
+numpy>=1.24.0
+sentencepiece>=0.2.0
+accelerate>=0.28.0
+huggingface_hub>=0.20.0