import os import re from datetime import datetime from typing import List import streamlit as st from pypdf import PdfReader from reportlab.lib.pagesizes import A4 from reportlab.pdfgen import canvas from reportlab.lib.units import mm KEY_RE = re.compile(r"\b\d{44}\b") def extract_keys_from_pdf(pdf_path: str) -> List[str]: reader = PdfReader(pdf_path) found: List[str] = [] seen = set() for page in reader.pages: text = page.extract_text() or "" for k in KEY_RE.findall(text): if k not in seen: seen.add(k) found.append(k) return found def safe_date_default() -> str: return datetime.now().strftime("%d/%m/%Y") def normalize_hora(hora_input: str) -> str: if not hora_input: return "_____ : _____" s = hora_input.strip() if s in {"_____:_____", "_____ : _____", "____ : ____"}: return "_____ : _____" return s def render_print_pdf(out_path: str, data_coleta: str, hora_coleta: str, keys: List[str]): c = canvas.Canvas(out_path, pagesize=A4) _, h = A4 left = 20 * mm top = h - 20 * mm line_h = 6.2 * mm def draw(text: str, y: float, size=11, bold=False) -> float: c.setFont("Helvetica-Bold" if bold else "Helvetica", size) c.drawString(left, y, text) return y - line_h y = top # Cabeçalho (modelo) y = draw("CLIENTE:", y) y -= line_h * 1.3 y = draw("DATA DA COLETA:", y) y = draw(data_coleta, y) y -= line_h * 0.6 y = draw("HORA DA COLETA:", y) y = draw(hora_coleta, y) y -= line_h * 0.9 # Título da lista y = draw("CHAVES DE ACESSO:", y) y -= line_h * 0.2 c.setFont("Helvetica", 10.8) for k in keys: if y < 40 * mm: c.showPage() y = top c.setFont("Helvetica", 10.8) c.drawString(left, y, k) y -= line_h * 0.85 y -= line_h * 0.8 y = draw(f"TOTAL DA REMESSA: {len(keys)} VOLUMES", y) # Assinaturas no rodapé if y < 55 * mm: c.showPage() y_sig = 25 * mm c.setFont("Helvetica", 9.5) c.drawString(left + 70 * mm, y_sig + 12, "ASSINATURA DO REPRESENTANTE") c.drawString(left + 140 * mm, y_sig + 12, "ASSINATURA DO MOTORISTA") c.line(left + 55 * mm, y_sig + 10, left + 118 * mm, y_sig + 10) c.line(left + 130 * mm, y_sig + 10, left + 193 * mm, y_sig + 10) c.save() st.set_page_config(page_title="Extrator de Chaves NF-e", layout="centered") st.title("Extrator de Chaves NF-e (PDF → PDF para imprimir)") pdf = st.file_uploader("Envie o PDF", type=["pdf"]) data_coleta = st.text_input("Data da coleta", value=safe_date_default()) hora_coleta = st.text_input("Hora da coleta", value="_____ : _____") if pdf is not None: st.caption(f"Arquivo: {pdf.name}") if st.button("Gerar arquivos", disabled=(pdf is None)): tmp_in = os.path.join("/tmp", f"entrada_{datetime.now().strftime('%Y%m%d-%H%M%S')}.pdf") with open(tmp_in, "wb") as f: f.write(pdf.read()) keys = extract_keys_from_pdf(tmp_in) if not keys: st.error("Não encontrei chaves de acesso (44 dígitos) no PDF.") st.stop() data_coleta = (data_coleta or safe_date_default()).strip() hora_coleta_norm = normalize_hora(hora_coleta) ts = datetime.now().strftime("%Y%m%d-%H%M%S") out_pdf = os.path.join("/tmp", f"chaves_prontas_{ts}.pdf") out_txt = os.path.join("/tmp", f"chaves_{ts}.txt") render_print_pdf(out_pdf, data_coleta, hora_coleta_norm, keys) with open(out_txt, "w", encoding="utf-8") as f: f.write("\n".join(keys) + "\n") st.success(f"{len(keys)} chaves encontradas.") with open(out_pdf, "rb") as f: st.download_button( "Baixar PDF pronto para imprimir", data=f, file_name=os.path.basename(out_pdf), mime="application/pdf", ) with open(out_txt, "rb") as f: st.download_button( "Baixar TXT (opcional)", data=f, file_name=os.path.basename(out_txt), mime="text/plain", ) st.text_area("Prévia (primeiras 10 chaves)", value="\n".join(keys[:10]), height=200)