public string PdfFields() { string pdfText = String.Empty; PDFParser parser = new PDFParser(new BufferedInputStream(new FileInputStream(PdfFile))); parser.parse(); PDDocument originialPdfDoc = parser.getPDDocument(); bool isOriginalDocEncrypted = originialPdfDoc.isEncrypted(); if (isOriginalDocEncrypted) { originialPdfDoc.openProtection(new StandardDecryptionMaterial(PdfPassword)); } try { PDDocumentCatalog docCatalog = originialPdfDoc.getDocumentCatalog(); PDAcroForm acroForm = docCatalog.getAcroForm(); PDField field = acroForm.getField("Name"); if (field != null) { field.setValue("name"); } } catch (java.io.IOException ex) { throw ex; } return(pdfText); }
public string PdfToText() { string pdfText = String.Empty; PDFParser parser = new PDFParser(new BufferedInputStream(new FileInputStream(PdfFile))); parser.parse(); PDDocument originialPdfDoc = parser.getPDDocument(); bool isOriginalDocEncrypted = originialPdfDoc.isEncrypted(); if (isOriginalDocEncrypted) { originialPdfDoc.openProtection(new StandardDecryptionMaterial(PdfPassword)); } PDFTextStripper stripper = new PDFTextStripper(); try { pdfText = stripper.getText(originialPdfDoc); } catch (java.io.IOException ex) { throw ex; } return(pdfText); }
//Conversão do PDF public void ReadPDF() { IWebDriver driver = new ChromeDriver(); URL TestURL = new URL(driver.Url); BufferedInputStream TestFile = new BufferedInputStream(TestURL.openStream()); PDFParser TestPDF = new PDFParser(TestFile); TestPDF.parse(); String TestText = new PDFTextStripper().getText(TestPDF.getPDDocument()); System.IO.File.WriteAllText(@"C:\Users\Nicolas PC\Desktop\teste\PDFTESTE.txt", TestText); }
public void Infocrim() { var options = new ChromeOptions(); options.AddArguments("headless"); //using (IWebDriver driver = new ChromeDriver("C:/inetpub/wwwroot/wwwroot",options)) using (IWebDriver driver = new ChromeDriver()) { Actions builder = new Actions(driver); //Validação driver.Navigate().GoToUrl("http://ec2-18-231-116-58.sa-east-1.compute.amazonaws.com/ "); driver.FindElement(By.Id("username")).SendKeys("fiap"); driver.FindElement(By.Id("password")).SendKeys("mpsp"); driver.FindElement(By.Id("password")).SendKeys(Keys.Enter); driver.Navigate().GoToUrl("http://ec2-18-231-116-58.sa-east-1.compute.amazonaws.com/infocrim/login.html"); driver.FindElement(By.XPath("/html/body/table/tbody/tr[3]/td/table/tbody/tr/td/table/tbody/tr/td/table/tbody/tr[2]/td[4]/a/img")).Click(); driver.FindElement(By.XPath("/html/body/a/table[3]/tbody/tr/td[2]/table[1]/tbody/tr[3]/td/table/tbody/tr[2]/td/table/tbody/tr/td/div/a/img")).Click(); driver.FindElement(By.XPath("/html/body/table/tbody/tr[2]/td/table[3]/tbody/tr[2]/td[2]/a")).Click(); driver.FindElement(By.XPath("/html/body/table/tbody/tr/td/a[2]/img")).Click(); driver.FindElement(By.XPath("/html/body/print-preview-app//print-preview-sidebar//div[2]/print-preview-destination-settings//print-preview-settings-section[1]/div/print-preview-destination-select//select")).Click(); driver.FindElement(By.XPath("/html/body/print-preview-app//print-preview-sidebar//div[2]/print-preview-destination-settings//print-preview-settings-section[1]/div/print-preview-destination-select//select/option[2]")).Click(); URL url = new URL(driver.Url); BufferedInputStream fileToParse = new BufferedInputStream(url.openStream()); PDFParser parser = new PDFParser(fileToParse); parser.parse(); COSDocument cosDoc = parser.getDocument(); PDDocument pdDoc = new PDDocument(cosDoc); PDFTextStripper pdfStripper = new PDFTextStripper(); pdfStripper.setStartPage(1); pdfStripper.setEndPage(1); string parsedText = pdfStripper.getText(cosDoc); string saida = new PDFTextStripper().getText(parser.getPDDocument()); System.IO.File.WriteAllText(@"C:\Users\favar\Desktop\Texto\Infocrim.txt", saida); } }
public string Parse(string fileName) { //Load in file. Using java.io because pdfbox is ported from java. var pdfFile = new FileInputStream(fileName); //Load file into the pdf parser var pdfParser = new PDFParser(pdfFile); //Parse the document, so that we can get it for the COSDocument pdfParser.parse(); /* COSDocument is the in-memory representation of the PDF. see https://pdfbox.apache.org/docs/1.8.4/javadocs/org/apache/pdfbox/cos/COSDocument.html */ var cosDocument = pdfParser.getDocument(); var pdDocument = new PDDocument(cosDocument); //Instantiate text stripper. var pdfTextStripper = new PDFTextStripper(); /* Needed for only stripping specific pages pdfTextStripper.setStartPage(0); pdfTextStripper.setEndPage(pdDocument.getNumberOfPages()); */ //Needed so that we can close the pdDocument before returning from this method var strippedText = pdfTextStripper.getText(pdDocument); //This closes all storage and delete the tmp files. pdDocument.close(); cosDocument.close(); return strippedText; }
public string Detran(PesquisaCPFCNPJ pesquisaCPFCNPJ) { var options = new ChromeOptions(); //options.AddArguments("headless"); options.AddArguments("no-sandbox"); using (IWebDriver driver = new ChromeDriver("C:/inetpub/wwwroot/wwwroot", options)) //using (IWebDriver driver = new ChromeDriver(options)) { Actions builder = new Actions(driver); driver.Navigate().GoToUrl("http://ec2-18-231-116-58.sa-east-1.compute.amazonaws.com/detran/login.html"); driver.FindElement(By.Id("form:j_id563205015_44efc15b")).Click(); driver.FindElement(By.Id("navigation_a_M_16")).Click(); driver.FindElement(By.XPath("//*[@id='navigation_a_F_16']")).Click(); driver.FindElement(By.Id("form:rg")).SendKeys(pesquisaCPFCNPJ.CPFCNPJ.ToString()); driver.FindElement(By.Id("form:nome")).SendKeys(pesquisaCPFCNPJ.Nome); driver.FindElement(By.LinkText("Pesquisar")).Click(); driver.SwitchTo().Window(driver.WindowHandles[1]); URL url = new URL(driver.Url); BufferedInputStream fileToParse = new BufferedInputStream(url.openStream()); PDFParser parser = new PDFParser(fileToParse); parser.parse(); COSDocument cosDoc = parser.getDocument(); PDDocument pdDoc = new PDDocument(cosDoc); PDFTextStripper pdfStripper = new PDFTextStripper(); pdfStripper.setStartPage(1); pdfStripper.setEndPage(1); string parsedText = pdfStripper.getText(pdDoc); string saida = new PDFTextStripper().getText(parser.getPDDocument()); driver.SwitchTo().Window(driver.WindowHandles[0]); driver.FindElement(By.Id("navigation_a_M_16")).Click(); driver.FindElement(By.PartialLinkText("Consultar Imagem da CNH")).Click(); driver.FindElement(By.LinkText("Pesquisar")).Click(); driver.SwitchTo().Window(driver.WindowHandles[2]); //string nomePai = driver.FindElement(By.XPath("/html/body/div[4]/div/table/tbody/tr/td/div/div/form/div[3]/div/table/tbody/tr/td/table/tbody/tr[2]/td/table/tbody/tr/td[2]")).Text; string nPai = driver.FindElement(By.XPath("/html/body/div[4]/div/table/tbody/tr/td/div/div/form/div[3]/div/table/tbody/tr/td/table/tbody/tr[2]/td/table/tbody/tr/td[2]/table/tbody/tr[3]/td/table/tbody/tr[2]/td/span")).Text; string nMae = driver.FindElement(By.XPath("/html/body/div[4]/div/table/tbody/tr/td/div/div/form/div[3]/div/table/tbody/tr/td/table/tbody/tr[2]/td/table/tbody/tr/td[2]/table/tbody/tr[4]/td/table/tbody/tr[2]/td/span")).Text; driver.SwitchTo().Window(driver.WindowHandles[0]); driver.FindElement(By.Id("navigation_a_M_18")).Click(); driver.FindElement(By.PartialLinkText("Consultar Veículo Base Estadual")).Click(); driver.FindElement(By.XPath("/html/body/div[4]/div/table/tbody/tr/td/div/div/form/div[1]/div[2]/table[2]/tbody/tr[2]/td[2]/input")).SendKeys(pesquisaCPFCNPJ.CPFCNPJ.ToString()); driver.FindElement(By.LinkText("Pesquisar")).Click(); driver.SwitchTo().Window(driver.WindowHandles[3]); URL urlCarro = new URL(driver.Url); BufferedInputStream fileToParseCarro = new BufferedInputStream(urlCarro.openStream()); PDFParser parserCarro = new PDFParser(fileToParseCarro); parserCarro.parse(); COSDocument cosDocCarro = parserCarro.getDocument(); PDDocument pdDocCarro = new PDDocument(cosDocCarro); PDFTextStripper pdfStripperCarro = new PDFTextStripper(); pdfStripper.setStartPage(1); pdfStripper.setEndPage(1); string parsedTextCarro = pdfStripperCarro.getText(pdDocCarro); string saidaCarro = new PDFTextStripper().getText(parserCarro.getPDDocument()); string resultado = saida + nPai + nMae + saidaCarro; string[] strsplit = resultado.Replace("\r\n", ":").Split(':'); string cpf = strsplit[33].Trim(); string rg = strsplit[13].Trim(); string expeditor = strsplit[34].Trim(); string registro = strsplit[36].Trim(); string local = strsplit[38].Trim(); string espelhoPid = strsplit[40].Trim(); string emissaoCnh = strsplit[42].Trim(); string categoria = strsplit[46].Trim(); string primeiraHab = strsplit[48].Trim(); string statusCnh = strsplit[50].Trim(); string renach = strsplit[52].Trim(); string espelhoCnh = strsplit[54].Trim(); string validadeCnh = strsplit[56].Trim(); string pontuacao = strsplit[58].Trim(); string nomePai = strsplit[119].Trim(); string nomeMae = strsplit[120].Trim(); string placa = strsplit[144].Replace(" 7107 - SAO PAULO", "").Trim(); string municipioPlaca = strsplit[144].Replace("gge4223 ", "").Trim(); string renavam = strsplit[146].Replace(" 9AAAAVAU0J4001600 ", "").Trim(); string chassi = strsplit[146].Replace("01172566666 ", "").Trim(); string numMotor = strsplit[148].Replace(" 22/11/18 00", "").Trim(); string dataAltMotor = strsplit[148].Replace("CWL031481 ", "").Trim(); string tipo = strsplit[151].Replace(" 1 - IMPORTADO 16 - ALCO/GASOL", "").Trim(); string procedencia = strsplit[151].Replace("6 - AUTOMOVEL ", "").Replace(" 16 - ALCO/GASOL ", "").Trim(); string combustivel = strsplit[151].Replace("6 - AUTOMOVEL 1 - IMPORTADO ", "").Trim(); string cor = strsplit[153].Replace(" 162801 – VARIANT GL ", "").Trim(); string marcaModelo = strsplit[153].Replace("4 - BRANCA 162801 – ", "").Trim(); string categoriaAut = strsplit[155].Replace(" 1971 1972 ", "").Trim(); string anoFab = strsplit[155].Replace("1 - PARTICULAR ", "").Replace(" 1972 ", "").Trim(); string anoMod = strsplit[155].Replace("1 - PARTICULAR 1971 ", "").Trim(); string logradouro = strsplit[166].Replace(" 00121 ", "").Trim(); string numero = strsplit[166].Replace("AV LINS DE VASCONCELOS ", "").Trim(); string complemento = strsplit[182].Replace(" 010006-010 ", "").Trim(); string cep = strsplit[182].Replace("4 ANDAR ", "").Trim(); string bairro = strsplit[184].Replace(" 7107 - SAO PAULO SP ", "").Trim(); string licenciamento = strsplit[225].Replace(" 07/03/2019 ", "").Trim(); string dataLicenciamento = strsplit[225].Replace("2019 ", "").Trim(); string dataEmissaoCRV = strsplit[227].Trim(); DetranModel objDen = new DetranModel(); objDen.CNPJCPF = long.Parse(cpf.Replace(".", "").Replace("-", "")); objDen.RG = rg; objDen.Expeditor = expeditor; objDen.Registro = registro; objDen.Local = local; objDen.PID = espelhoPid; objDen.EmissaoCnh = emissaoCnh; objDen.Categoria = categoria; objDen.PrimeiraHabilitação = primeiraHab; objDen.StatusCnh = statusCnh; objDen.Renach = renach; objDen.EspelhoCnh = espelhoCnh; objDen.ValidadeCnh = validadeCnh; objDen.Pontuacao = pontuacao; objDen.NomePai = nPai; objDen.NomeMae = nMae; objDen.Placa = placa; objDen.MunicipioCarro = municipioPlaca; objDen.Renavam = renavam; objDen.Chassi = chassi; objDen.NumMotor = numMotor; objDen.DataAltMotor = dataAltMotor; objDen.Tipo = tipo; objDen.Procedencia = procedencia; objDen.Combustivel = combustivel; objDen.Cor = cor; objDen.MarcaModelo = marcaModelo; objDen.CategoriaAut = categoriaAut; objDen.Fabricacao = anoFab; objDen.Modelo = anoMod; objDen.Logradouro = logradouro; objDen.Numero = numero; objDen.Complemento = complemento; objDen.CEP = cep; objDen.Bairro = bairro; objDen.Licenciamento = licenciamento; objDen.DataLicenciamento = dataLicenciamento; objDen.DataEmissaoCRV = dataEmissaoCRV; detranRepository.Insert(objDen); string objjsonData = JsonConvert.SerializeObject(objDen, new JsonSerializerSettings { Formatting = Formatting.Indented }); //System.IO.File.WriteAllText(@"C:\Users\favar\Desktop\Texto\Detran.txt", objjsonData); return(objjsonData); } }