import io import os import re from datetime import datetime from collections import Counter import pandas as pd import pytz import streamlit as st from PIL import Image from reportlab.pdfgen import canvas from reportlab.lib.utils import ImageReader # --- App Configuration ---------------------------------- st.set_page_config( page_title="Image → PDF Comic Layout", layout="wide", initial_sidebar_state="expanded", ) st.title("🖼️ Image → PDF • Scan, Reorder & Caption Generator") st.markdown( "Scan docs or upload images, filter by orientation, reorder, then generate a captioned PDF matching each image’s dimensions." ) # --- Session State for Snapshots ----------------------- st.session_state.setdefault('snapshots', []) # --- Sidebar: Page Settings ----------------------------- st.sidebar.header("1️⃣ Page Aspect Ratio & Size") ratio_map = { "4:3 (Landscape)": (4, 3), "16:9 (Landscape)": (16, 9), "1:1 (Square)": (1, 1), "2:3 (Portrait)": (2, 3), "9:16 (Portrait)": (9, 16), } ratio_choice = st.sidebar.selectbox( "Preset Ratio", list(ratio_map.keys()) + ["Custom…"] ) if ratio_choice != "Custom…": rw, rh = ratio_map[ratio_choice] else: rw = st.sidebar.number_input("Custom Width Ratio", min_value=1, value=4) rh = st.sidebar.number_input("Custom Height Ratio", min_value=1, value=3) BASE_WIDTH_PT = st.sidebar.slider( "Base Page Width (pt)", min_value=400, max_value=1200, value=800, step=100 ) page_width = BASE_WIDTH_PT page_height = int(BASE_WIDTH_PT * (rh / rw)) st.sidebar.markdown(f"**Preview page size:** {page_width}×{page_height} pt") # --- Main: Document Scan & Image Upload ---------------- st.header("2️⃣ Document Scan & Image Upload") # Camera scan cam = st.camera_input("📸 Scan Document") if cam: central = pytz.timezone("US/Central") now = datetime.now(central) prefix = now.strftime("%Y-%m%d-%I%M%p") + "-" + now.strftime("%a").upper() scan_name = f"{prefix}-scan.png" with open(scan_name, "wb") as f: f.write(cam.getvalue()) st.image(Image.open(scan_name), caption=scan_name, use_container_width=True) if scan_name not in st.session_state['snapshots']: st.session_state['snapshots'].append(scan_name) # File uploader uploads = st.file_uploader( "📂 Upload PNG/JPG images", type=["png","jpg","jpeg"], accept_multiple_files=True ) # --- Build combined list -------------------------------- all_records = [] # From snapshots for idx, path in enumerate(st.session_state['snapshots']): im = Image.open(path) w, h = im.size ar = round(w / h, 2) orient = "Square" if 0.9 <= ar <= 1.1 else ("Landscape" if ar > 1.1 else "Portrait") all_records.append({ "filename": os.path.basename(path), "source": path, "width": w, "height": h, "aspect_ratio": ar, "orientation": orient, "order": idx, }) # From uploads if uploads: for jdx, f in enumerate(uploads, start=len(all_records)): im = Image.open(f) w, h = im.size ar = round(w / h, 2) orient = "Square" if 0.9 <= ar <= 1.1 else ("Landscape" if ar > 1.1 else "Portrait") all_records.append({ "filename": f.name, "source": f, "width": w, "height": h, "aspect_ratio": ar, "orientation": orient, "order": jdx, }) # DataFrame df = pd.DataFrame(all_records) # Filter dims = st.sidebar.multiselect( "Include orientations:", options=["Landscape","Portrait","Square"], default=["Landscape","Portrait","Square"] ) df = df[df['orientation'].isin(dims)].reset_index(drop=True) st.markdown("#### Images & Scan Metadata (drag/order)") st.dataframe(df.style.format({"aspect_ratio":"{:.2f}"}), use_container_width=True) st.markdown("*Drag rows or edit `order` to set PDF page sequence.*") # Reordering try: edited = st.experimental_data_editor(df, num_rows="fixed", use_container_width=True) ordered_df = edited except Exception: edited = st.data_editor( df, column_config={"order": st.column_config.NumberColumn("Order", min_value=0, max_value=len(df)-1)}, hide_index=True, use_container_width=True, ) ordered_df = edited.sort_values('order').reset_index(drop=True) # Resolve ordered_sources list ordered_sources = [row['source'] for _, row in ordered_df.iterrows()] # --- Utility: Clean stems ------------------------------- def clean_stem(fn: str) -> str: return os.path.splitext(fn)[0].replace('-', ' ').replace('_', ' ') # --- PDF Creation: Image Sized + Captions -------------- def make_image_sized_pdf(sources): buf = io.BytesIO() c = canvas.Canvas(buf) for idx, src in enumerate(sources, start=1): im = Image.open(src) iw, ih = im.size cap_h = 20 pw, ph = iw, ih + cap_h c.setPageSize((pw, ph)) c.drawImage(ImageReader(im), 0, cap_h, iw, ih, preserveAspectRatio=True, mask='auto') caption = clean_stem(os.path.basename(src)) c.setFont('Helvetica', 12) c.drawCentredString(pw/2, cap_h/2, caption) c.setFont('Helvetica', 8) c.drawRightString(pw-10, 10, str(idx)) c.showPage() c.save(); buf.seek(0) return buf.getvalue() # --- Generate & Download ------------------------------- st.header("3️⃣ Generate & Download PDF") if st.button("🖋️ Generate Captioned PDF"): if not ordered_sources: st.warning("No images or scans to include.") else: central = pytz.timezone("US/Central") now = datetime.now(central) prefix = now.strftime("%Y-%m%d-%I%M%p") + "-" + now.strftime("%a").upper() stems = [clean_stem(os.path.basename(s)) for s in ordered_sources] # limit to 4 stems if too long stems = stems[:4] basename = " - ".join(stems) fname = f"{prefix}-{basename}.pdf" pdf_bytes = make_image_sized_pdf(ordered_sources) st.success(f"✅ PDF ready: **{fname}**") st.download_button("⬇️ Download PDF", data=pdf_bytes, file_name=fname, mime="application/pdf") st.markdown("#### Preview Page 1") try: import fitz doc = fitz.open(stream=pdf_bytes, filetype='pdf') pix = doc.load_page(0).get_pixmap(matrix=fitz.Matrix(1.5,1.5)) st.image(pix.tobytes(), use_container_width=True) except: st.info("Install `pymupdf` for preview.") # --- Footer ------------------------------------------------ st.sidebar.markdown("---") st.sidebar.markdown("Built by Aaron C. Wacker • Senior AI Engineer")