public static void Main()
{
// NOTE:
// When used in trial mode, the library imposes some restrictions.
// Please visit http://bitmiracle.com/pdf-library/trial-restrictions.aspx
// for more information.
using (PdfDocument pdf = new PdfDocument(@"..\Sample data\jfif3.pdf"))
{
// Extract plain text from document
string documentTextFile = "Document text.txt";
using (StreamWriter writer = new StreamWriter(documentTextFile))
writer.Write(pdf.GetText());
// Extract text with formatting from document
string documentTextFormattedFile = "Document text with formatting.txt";
using (StreamWriter writer = new StreamWriter(documentTextFormattedFile))
writer.Write(pdf.GetTextWithFormatting());
// Only extract visible plain text from first page
string firstPageTextFile = "First page text.txt";
using (StreamWriter writer = new StreamWriter(firstPageTextFile))
{
var options = new PdfTextExtractionOptions
{
WithFormatting = false,
SkipInvisibleText = true
};
writer.Write(pdf.Pages[0].GetText(options));
}
}
Console.WriteLine($"The output is located in {Environment.CurrentDirectory}");
}
public void ProcessFile()
{
string filePath = FileTable.FileLocation;
using (var pdf = new PdfDocument(filePath))
{
PdfTextExtractionOptions options = new PdfTextExtractionOptions
{
SkipInvisibleText = true,
WithFormatting = true
};
string formattedText = pdf.GetText(options);
string modifiedText = Regex.Replace(formattedText, @"(?<=\b)\p{Zs}(?=\b)", string.Empty);
string[] strArray = Regex.Split(modifiedText, @"\s+");
foreach (var map in Mappings)
{
OutputData.Add(new Output
{
FileId = FileTable.FileId,
DataType = map.MapType,
Value = GetNextIndexValue(strArray, map.MapName),
});
}
}
}
private string GetTextFromPdf(string filePath)
{
string text = "";
using (var pdf = new PdfDocument(filePath))
{
var options = new PdfTextExtractionOptions
{
SkipInvisibleText = true,
WithFormatting = false
};
text = pdf.GetText(options);
}
return(text);
}
static string getNomeFantasia(string path)
{
string identificador = "";
using (var pdf = new PdfDocument(path))
{
var page = pdf.Pages[1];
//CNPJ
var options = new PdfTextExtractionOptions
{
Rectangle = new PdfRectangle(230.30, 285.33, 11.06, 13.61),
WithFormatting = false
};
identificador = page.GetText(options);
Console.WriteLine(identificador);
}
return(identificador);
}
static string getIdentificadoMatriz(string path)
{
string identificador = "";
using (var pdf = new PdfDocument(path))
{
var page = pdf.Pages[1];
//CNPJ
var options = new PdfTextExtractionOptions
{
Rectangle = new PdfRectangle(227.50, 205.33, 5.56, 13.61),
WithFormatting = false
};
identificador = page.GetText(options);
Console.WriteLine(identificador);
}
return(identificador);
}
static string getCNPJ(string path)
{
string cnpj = "";
using (var pdf = new PdfDocument(path))
{
var page = pdf.Pages[1];
//CNPJ
var options = new PdfTextExtractionOptions
{
Rectangle = new PdfRectangle(227.50, 162.33, 11.06, 13.61),
WithFormatting = false
};
cnpj = page.GetText(options);
Console.WriteLine(cnpj);
}
return(cnpj);
}
static string getcnpjSocio(string path)
{
string cnpjSocio = "";
using (var pdf = new PdfDocument(path))
{
var page = pdf.Pages[4];
//CNPJ
var options = new PdfTextExtractionOptions
{
Rectangle = new PdfRectangle(212.20, 686.23, 11.15, 13.61),
WithFormatting = false
};
cnpjSocio = page.GetText(options);
Console.WriteLine(cnpjSocio);
}
return(cnpjSocio);
}
static string getidentificadoSocio(string path)
{
string identificadoSocio = "";
using (var pdf = new PdfDocument(path))
{
var page = pdf.Pages[4];
//CNPJ
var options = new PdfTextExtractionOptions
{
Rectangle = new PdfRectangle(212.20, 552.23, 5.56, 13.61),
WithFormatting = false
};
identificadoSocio = page.GetText(options);
Console.WriteLine(identificadoSocio);
}
return(identificadoSocio);
}
public Task <string> GetAsText(Stream fileStream, CancellationToken cancellationToken = default)
{
try
{
using (var pdf = new PdfDocument(fileStream))
{
var options = new PdfTextExtractionOptions
{
SkipInvisibleText = true,
WithFormatting = false
};
return(Task.FromResult(pdf.GetText()));
}
}
catch (Exception ex)
{
_logger.LogError(ex, ex.Message);
return(Task.FromResult(string.Empty));
}
}