static void Main(string[] args)
{
// Extract all images from 1st PDF page
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// This property is necessary only for registered version
// f.Serial = "XXXXXXXXXXX";
string pdfFile = @"..\..\simple text.pdf";
string imageDir = new DirectoryInfo(Directory.GetCurrentDirectory()).CreateSubdirectory("images").FullName;
List <PdfFocus.PdfImage> pdfImages = null;
f.OpenPdf(pdfFile);
if (f.PageCount > 0)
{
pdfImages = f.ExtractImages(1, 1);
// Show all extracted images.
if (pdfImages != null && pdfImages.Count > 0)
{
for (int i = 0; i < pdfImages.Count; i++)
{
string imageFile = Path.Combine(imageDir, String.Format("img{0}.png", i + 1));
pdfImages[i].Picture.Save(imageFile);
}
System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(imageDir)
{
UseShellExecute = true
});
}
}
}
static void Main(string[] args)
{
// Extract all images from PDF
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// This property is necessary only for registered version
// f.Serial = "XXXXXXXXXXX";
string pdfFile = @"..\..\..\..\..\simple text.pdf";
string imageDir = Path.GetDirectoryName(pdfFile);
List <PdfFocus.PdfImage> pdfImages = null;
f.OpenPdf(pdfFile);
if (f.PageCount > 0)
{
// Rasterize all vector graphics
f.ImageExtractionOptions.RasterizeComplexGraphics = true;
pdfImages = f.ExtractImages();
// Show all extracted images.
if (pdfImages != null && pdfImages.Count > 0)
{
for (int i = 0; i < pdfImages.Count; i++)
{
string imageFile = Path.Combine(imageDir, String.Format("img{0}.png", i + 1));
pdfImages[i].Picture.Save(imageFile);
System.Diagnostics.Process.Start(imageFile);
}
}
}
}
static void Main(string[] args)
{
//foreach (string file in Directory.EnumerateFiles(@"C:\Users\yanswong\source\repos\Convert PDF file to Excel file\C#\SW\", "*.pdf"))
foreach (string file in Directory.EnumerateFiles(@"C:\Users\yanswong\Desktop\CMM\May-2017\", "*.pdf", SearchOption.AllDirectories))
//foreach (string [] file in Directory.GetDirectories(@"\\pngnas2.mys.agilent.com\EUC_Gshare\vpdpenang\VPD LDA\VPD LDA\CMM MEASUREMENT RESULT\2018\","*",SearchOption.AllDirectories))
{
//string pathToPdf = @"C:\Users\yanswong\source\repos\Convert PDF file to Excel file\C#\SW\3700391-115-01-18-2.pdf";
//string pathToExcel = Path.ChangeExtension(pathToPdf, ".xls");
string pathToExcel = Path.ChangeExtension(file, ".xls");
// Convert PDF file to Excel file
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// 'true' = Convert all data to spreadsheet (tabular and even textual).
// 'false' = Skip textual data and convert only tabular (tables) data.
f.ExcelOptions.ConvertNonTabularDataToSpreadsheet = true;
// 'true' = Preserve original page layout.
// 'false' = Place tables before text.
f.ExcelOptions.PreservePageLayout = true;
//f.OpenPdf(pathToPdf);
f.OpenPdf(file);
if (f.PageCount > 0)
{
int result = f.ToExcel(pathToExcel);
//Open a produced Excel workbook
//if (result==0)
//{
// System.Diagnostics.Process.Start(pathToExcel);
//}
}
}
}
private void button1_Click(object sender, EventArgs e)
{
this.Cursor = Cursors.WaitCursor;
lblConversao.Text = "";
lblCaminho.Text = "";
try
{
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
caminho = @escolherArquivo();
f.OpenPdf(@caminho);
//string caminho = (Environment.GetFolderPath caminho);
if (f.PageCount > 0)
{
string destino = @Environment.GetFolderPath(Environment.SpecialFolder.MyDocuments).ToString() + @"\Conversao"
+ nomeArquivo.Replace(".", "") + ".docx";
f.WordOptions.Format = SautinSoft.PdfFocus.CWordOptions.eWordDocument.Docx;
f.ToWord(destino);
lblConversao.ForeColor = Color.Green;
lblConversao.Text = destino;
}
}
catch (Exception ex)
{
lblConversao.ForeColor = Color.Red;
lblConversao.Text = "Erro: " + ex.Message;
}
this.Cursor = Cursors.Default;
}
static void Main(string[] args)
{
string pdfFile = @"..\..\..\..\..\Text.pdf";
string htmlFile = Path.ChangeExtension(pdfFile, ".htm");
// Convert PDF file to HTML file
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// Let's change all text to Verdana 8pt.
f.HtmlOptions.SingleFontFamily = "Verdana";
f.HtmlOptions.SingleFontSize = 8;
// After purchasing the license, please insert your serial number here to activate the component:
//f.Serial = "123456789";
f.OpenPdf(pdfFile);
if (f.PageCount > 0)
{
int from = 1;
int to = (3 > f.PageCount) ? f.PageCount : 3;
int result = f.ToHtml(htmlFile, from, to);
// Show resulted HTML document in a browser.
if (result == 0)
{
System.Diagnostics.Process.Start(htmlFile);
}
}
}
static void Main(string[] args)
{
// Extract all images with width and height more than 200px
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
string pdfFile = @"D:\Personal_GitHub\ExtractimagesFromPDF\PdfUtils\ReadPdfImage\Data\dipak.pdf";
string imageDir = Path.GetDirectoryName(pdfFile);
List <PdfFocus.PdfImage> pdfImages = null;
f.OpenPdf(pdfFile);
if (f.PageCount > 0)
{
// Specify to extract only images which have width and height
// more than 200px
f.ImageExtractionOptions.MinSize = new System.Drawing.Size(200, 200);
pdfImages = f.ExtractImages();
// Show all extracted images.
if (pdfImages != null && pdfImages.Count > 0)
{
for (int i = 0; i < pdfImages.Count; i++)
{
string imageFile = Path.Combine(imageDir, String.Format("img{0}.png", i + 1));
pdfImages[i].Picture.Save(imageFile);
System.Diagnostics.Process.Start(imageFile);
}
}
}
}
static void Main(string[] args)
{
// Extract all images with width and height more than 200px
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// This property is necessary only for registered version
// f.Serial = "XXXXXXXXXXX";
string pdfFile = @"..\..\..\..\..\simple text.pdf";
string imageDir = Path.GetDirectoryName(pdfFile);
List <PdfFocus.PdfImage> pdfImages = null;
f.OpenPdf(pdfFile);
if (f.PageCount > 0)
{
// Specify to extract only images which have width and height
// more than 200px
f.ImageExtractionOptions.MinSize = new System.Drawing.Size(200, 200);
pdfImages = f.ExtractImages();
// Show all extracted images.
if (pdfImages != null && pdfImages.Count > 0)
{
for (int i = 0; i < pdfImages.Count; i++)
{
string imageFile = Path.Combine(imageDir, String.Format("img{0}.png", i + 1));
pdfImages[i].Picture.Save(imageFile);
System.Diagnostics.Process.Start(imageFile);
}
}
}
}
static void Main(string[] args)
{
string pathToPdf = @"d:\Tempos\Table.pdf";
string pathToExcel = Path.ChangeExtension(pathToPdf, ".xls");
// Convert PDF file to Excel file
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// 'true' = Convert all data to spreadsheet (tabular and even textual).
// 'false' = Skip textual data and convert only tabular (tables) data.
f.ExcelOptions.ConvertNonTabularDataToSpreadsheet = true;
// 'true' = Preserve original page layout.
// 'false' = Place tables before text.
f.ExcelOptions.PreservePageLayout = true;
f.OpenPdf(pathToPdf);
if (f.PageCount > 0)
{
int result = f.ToExcel(pathToExcel);
//Open a produced Excel workbook
if (result == 0)
{
System.Diagnostics.Process.Start(pathToExcel);
}
}
}
static void Main(string[] args)
{
string pathToPdf = @"d:\Tempos\table.pdf";
string pathToHtml = Path.ChangeExtension(pathToPdf, ".htm");
// Convert PDF file to HTML file
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// You may download the latest version of SDK here:
// www.sautinsoft.com/products/pdf-focus/download.php
// Let's force the component to store images inside HTML document
// using base-64 encoding
f.HtmlOptions.IncludeImageInHtml = true;
f.HtmlOptions.Title = "Simple text";
// This property is necessary only for registered version
f.OpenPdf(pathToPdf);
if (f.PageCount > 0)
{
int result = f.ToHtml(pathToHtml);
//Show HTML document in browser
if (result == 0)
{
System.Diagnostics.Process.Start(pathToHtml);
}
}
}
static void Main(string[] args)
{
string pdfFile = @"..\..\..\..\..\simple text.pdf";
string htmlFile = Path.ChangeExtension(pdfFile, ".htm");
// Convert PDF file to HTML file
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// Path (must exist) to a directory to store images after converting. Notice also to the property "ImageSubFolder".
f.HtmlOptions.ImageFolder = Path.GetDirectoryName(pdfFile);
// A folder (will be created by the component) without any drive letters, only the folder as "myfolder".
f.HtmlOptions.ImageSubFolder = String.Format("{0}_images", Path.GetFileNameWithoutExtension(pdfFile));
// We recommend to use PNG type for storing images.
f.HtmlOptions.ImageType = PdfFocus.CHtmlOptions.eHtmlImageType.Png;
// How to store images: Inside HTML document as base64 images or as linked separate image files.
f.HtmlOptions.IncludeImageInHtml = false;
// Set <title>...</title>
f.HtmlOptions.Title = "Simple text";
// After purchasing the license, please insert your serial number here to activate the component:
//f.Serial = "123456789";
f.OpenPdf(pdfFile);
if (f.PageCount > 0)
{
int result = f.ToHtml(htmlFile);
// Show resulted HTML document in a browser.
if (result == 0)
{
System.Diagnostics.Process.Start(htmlFile);
}
}
}
static void Main(string[] args)
{
// Here you will find various ways to store images
string pdfFile = @"..\..\..\..\..\simple text.pdf";
string htmlFile = Path.ChangeExtension(pdfFile, ".htm");
// Convert PDF file to HTML file
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// Way 1: Images will be stored as physical PNG files in: ImageFolder + ImageSubFolder.
// For example: "d:\" + "special folder" = "d:\special folder\"
/*
* f.HtmlOptions.ImageFolder = @"d:\";
* f.HtmlOptions.ImageSubFolder = "special folder";
* f.HtmlOptions.ImageType = PdfFocus.CHtmlOptions.eHtmlImageType.Png;
*/
// Way 2: Images will be stored as PNG files in the same directory with HTML file.
// All images on each page will be combined in a single image.
/*
* f.HtmlOptions.ImageFolder = Path.GetDirectoryName(pdfFile);
* f.HtmlOptions.ImageType = PdfFocus.CHtmlOptions.eHtmlImageType.Png;
* f.HtmlOptions.ImageSubFolder = "";
* f.HtmlOptions.CombineImages = true;
*/
// Way 3: Images will be stored as Jpeg files in a special folder "my images".
// Images will have name "picture100.jpg", "picture101.jpg" .. "pictureN.jpg".
// Let's set the quality for jpeg images to 95 percents.
f.HtmlOptions.ImageFolder = Path.GetDirectoryName(pdfFile);
f.HtmlOptions.ImageType = PdfFocus.CHtmlOptions.eHtmlImageType.Jpeg;
f.HtmlOptions.JpegQuality = 95;
f.HtmlOptions.ImageSubFolder = "my images";
f.HtmlOptions.ImageFileName = "picture";
f.HtmlOptions.ImageNumStart = 100;
f.HtmlOptions.CombineImages = false;
// Way 4: Images will be stored inside HTML document as base64 images.
/*
* f.HtmlOptions.IncludeImageInHtml = true;
*/
f.OpenPdf(pdfFile);
if (f.PageCount > 0)
{
int result = f.ToHtml(htmlFile);
// Show resulted HTML document in a browser.
if (result == 0)
{
System.Diagnostics.Process.Start(htmlFile);
}
}
}
static void Main(string[] args)
{
// Here we'll show you two modes of converting PDF to HTML:
// PDF Focus .Net offers you the Fixed and Flowing modes by your choice.
// HTML-Fixed (default) is better to use for rendering, because it completely
// repeats the PDF layout with the structure of pages.
// The markup of such documents is very complex and have a lot of tags styled by (x,y) coords.
// HTML-Flowing is better for further processing by a human: editing and combining.
// The markup of such documents is much simple inside and has the flowing structure.
// It's very simple for understanding by a human.
// But the resulting HTML document doesn't look exactly the same as input PDF pixel by pixel.
string pdfFile = @"..\..\License.pdf";
string htmlFileFixed = "Fixed.html";
string htmlFileFlowing = "Flowing.html";
// Convert PDF file to HTML (Fixed and Flowing) file
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// After purchasing the license, please insert your serial number here to activate the component:
//f.Serial = "XXXXXXXXXXX";
// How to store images: Inside HTML document as base64 images or as linked separate image files.
f.HtmlOptions.IncludeImageInHtml = true;
f.OpenPdf(pdfFile);
if (f.PageCount > 0)
{
// The HTML-Fixed mode.
f.HtmlOptions.Title = "Fixed";
f.HtmlOptions.RenderMode = PdfFocus.CHtmlOptions.eHtmlRenderMode.Fixed;
if (f.ToHtml(htmlFileFixed) == 0)
{
System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(htmlFileFixed)
{
UseShellExecute = true
});
}
// The HTML-Flowing mode.
f.HtmlOptions.Title = "Flowing";
f.HtmlOptions.RenderMode = PdfFocus.CHtmlOptions.eHtmlRenderMode.Flowing;
// Switch off character scaling and spacing to prevent
// adding of extra tags dividing the text by parts.
f.HtmlOptions.KeepCharScaleAndSpacing = false;
if (f.ToHtml(htmlFileFlowing) == 0)
{
System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(htmlFileFlowing)
{
UseShellExecute = true
});
}
}
}
public static void ConvertToExcel(object targ)
{
TArgument targum = (TArgument)targ;
string pdfFile = targum.PdfFile;
int page = targum.PageNumber;
string excelFile = Path.GetFileNameWithoutExtension(pdfFile) + ".xls";
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// 'true' = Convert all data to spreadsheet (tabular and even textual).
// 'false' = Skip textual data and convert only tabular (tables) data.
f.ExcelOptions.ConvertNonTabularDataToSpreadsheet = true;
// 'true' = Preserve original page layout.
// 'false' = Place tables before text.
f.ExcelOptions.PreservePageLayout = true;
// The information includes the names for the culture, the writing system,
// the calendar used, the sort order of strings, and formatting for dates and numbers.
System.Globalization.CultureInfo ci = new System.Globalization.CultureInfo("en-US");
ci.NumberFormat.NumberDecimalSeparator = ",";
ci.NumberFormat.NumberGroupSeparator = ".";
f.ExcelOptions.CultureInfo = ci;
f.OpenPdf(pdfFile);
bool done = false;
if (f.PageCount > 0)
{
if (page >= f.PageCount)
{
page = 1;
}
if (f.ToExcel(excelFile, page, page) == 0)
{
done = true;
}
f.ClosePdf();
}
if (done)
{
Console.WriteLine("{0}\t - Done!", Path.GetFileName(pdfFile));
System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(excelFile)
{
UseShellExecute = true
});
}
else
{
Console.WriteLine("{0}\t - Error!", Path.GetFileName(pdfFile));
}
}
private void btnPDF2Html_Click(object sender, EventArgs e) // PDF convert to Html Function
{
SautinSoft.PdfFocus Html = new SautinSoft.PdfFocus(); // 呼叫SautinSoft中的Html Convert Function
Html.OpenPdf(@"C:\Users\施宗佑\OneDrive\桌面\期末報告 - 物件導向程式設計\測試用文檔\Graph.pdf");
if (Html.PageCount > 0)
{
int result = Html.ToHtml(@"C:\Users\施宗佑\OneDrive\桌面\期末報告 - 物件導向程式設計\NewGraph.html");
}
}
/// <summary>
/// Converts PDF to DOCX, RTF, HTML, Text with OCR engine.
/// </summary>
public void ConvertPdfToAllWithOCR(string pdfPath)
{
// To perform OCR we'll use free OCR library by Nicomsoft.
// https://www.nicomsoft.com/products/ocr/download/
// The library is freeware and can be used in commercial application.
// Also you have to insert this key: AB2A4DD5FF2A.
NsOCR = new NSOCRLib.NSOCRClass();
NsOCR.Engine_SetLicenseKey("AB2A4DD5FF2A"); //required for licensed version only
NsOCR.Engine_InitializeAdvanced(out CfgObj, out OcrObj, out ImgObj);
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
f.OCROptions.Method += PerformOCRNicomsoft;
f.OCROptions.Mode = PdfFocus.COCROptions.eOCRMode.AllImages;
f.WordOptions.KeepCharScaleAndSpacing = false;
string pdfFile = pdfPath;
string outFile = String.Empty;
f.OpenPdf(pdfFile);
if (f.PageCount > 0)
{
// To Docx.
outFile = "Result.docx";
f.WordOptions.Format = PdfFocus.CWordOptions.eWordDocument.Docx;
if (f.ToWord(outFile) == 0)
{
System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(outFile)
{
UseShellExecute = true
});
}
// To HTML.
outFile = "Result.html";
f.HtmlOptions.KeepCharScaleAndSpacing = false;
if (f.ToHtml(outFile) == 0)
{
System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(outFile)
{
UseShellExecute = true
});
}
}
else
{
Console.WriteLine("Error: {0}!", f.Exception.Message);
Console.ReadLine();
}
}
private void btnConvertJPG_Click(object sender, EventArgs e) // PDF convert to Image Function
{
SautinSoft.PdfFocus JPG = new SautinSoft.PdfFocus(); // 呼叫SautinSoft中的Image Convert Function
JPG.OpenPdf(textBox1.Text);
if (JPG.PageCount > 0)
{
JPG.ImageOptions.Dpi = 200;
JPG.ImageOptions.ImageFormat = System.Drawing.Imaging.ImageFormat.Jpeg;
for (int page = 1; page <= JPG.PageCount; page++)
{
JPG.ToImage(textBox1.Text + "page" + page + ".jpg", page);
}
}
}
protected void Button1_Click(object sender, EventArgs e)
{
if (file.HasFile)
{
try
{
if (file.PostedFile.FileName.Length == 0 || file.FileBytes.Length == 0)
{
Result.Text = "Please select PDF file at first!";
return;
}
byte[] rtf = null;
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
f.OpenPdf(file.FileBytes);
if (f.PageCount > 0)
{
//Let's whole PDF document to Word (RTF)
f.WordOptions.Format = SautinSoft.PdfFocus.CWordOptions.eWordDocument.Rtf;
// You may also set an output format to Docx.
//f.WordOptions.Format = SautinSoft.PdfFocus.CWordOptions.eWordDocument.Docx;
rtf = f.ToWord();
}
//show Word/rtf
if (rtf != null)
{
ShowResult(rtf, "Result.rtf", "application/msword");
}
else
{
Result.Text = "Converting failed!";
}
}
catch (Exception ex)
{
Response.Write("Error: " + ex.Message);
}
}
else
{
Result.Text = "Please choose a file to upload!";
}
}
static void Main(string[] args)
{
//converte pdf em imagem
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
f.OpenPdf(@"D:\Projetos\OCR_Demo\Resultado\PDF.pdf");
if (f.PageCount > 0)
{
f.ImageOptions.Dpi = 300;
f.ToMultipageTiff(@"D:\Projetos\OCR_Demo\Resultado\imagem\PDF.tiff");
}
ConvertTiffToJpeg(@"D:\Projetos\OCR_Demo\Resultado\imagem\PDF.tiff");
var files = Directory.GetFiles(@"D:\Projetos\OCR_Demo\Resultado\imagem\").Where(x => x.EndsWith("jpg"));
StringBuilder stringBuilder = new StringBuilder();
foreach (var file in files)
{
try
{
using (var engine = new TesseractEngine(@"tessdata", "por", EngineMode.Default))
{
using (var img = Pix.LoadFromFile(file))
{
using (var page = engine.Process(img))
{
var texto = page.GetText();
stringBuilder.Append(texto);
}
}
}
}
catch (Exception ex)
{
Console.WriteLine("Erro {0}", ex.Message);
}
finally
{
}
}
string filePath = @"D:\Projetos\OCR_Demo\Resultado\textos\teste.txt";
File.WriteAllText(filePath, stringBuilder.ToString());
}
public static void ConvertToExcel(object targ)
{
TArgument targum = (TArgument)targ;
string pdfFile = targum.PdfFile;
int page = targum.PageNumber;
string excelFile = Path.ChangeExtension(pdfFile, ".xls");
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// 'true' = Convert all data to spreadsheet (tabular and even textual).
// 'false' = Skip textual data and convert only tabular (tables) data.
f.ExcelOptions.ConvertNonTabularDataToSpreadsheet = true;
// 'true' = Preserve original page layout.
// 'false' = Place tables before text.
f.ExcelOptions.PreservePageLayout = true;
f.OpenPdf(pdfFile);
bool done = false;
if (f.PageCount > 0)
{
if (page >= f.PageCount)
{
page = 1;
}
if (f.ToExcel(excelFile, page, page) == 0)
{
done = true;
}
f.ClosePdf();
}
if (done)
{
Console.WriteLine("{0}\t - Done!", Path.GetFileName(pdfFile));
}
else
{
Console.WriteLine("{0}\t - Error!", Path.GetFileName(pdfFile));
}
}
public static void ConvertToPng(object targ)
{
TArgument targum = (TArgument)targ;
string pdfFile = targum.PdfFile;
int page = targum.PageNumber;
string pngFile = Path.GetFileNameWithoutExtension(pdfFile) + ".png";
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
f.ImageOptions.ImageFormat = System.Drawing.Imaging.ImageFormat.Png;
f.ImageOptions.Dpi = 300;
f.OpenPdf(pdfFile);
bool done = false;
if (f.PageCount > 0)
{
if (page >= f.PageCount)
{
page = 1;
}
if (f.ToImage(pngFile, page) == 0)
{
done = true;
}
f.ClosePdf();
}
if (done)
{
Console.WriteLine("{0}\t - Done!", Path.GetFileName(pdfFile));
System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(pngFile)
{
UseShellExecute = true
});
}
else
{
Console.WriteLine("{0}\t - Error!", Path.GetFileName(pdfFile));
}
}
static void Main(string[] args)
{
string pdfFile = @"..\..\simple text.pdf";
string htmlFile = "Result.html";
// Convert PDF file to HTML file
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// After purchasing the license, please insert your serial number here to activate the component:
//f.Serial = "XXXXXXXXXXX";
// Path (must exist) to a directory to store images after converting. Notice also to the property "ImageSubFolder".
f.HtmlOptions.ImageFolder = Path.GetDirectoryName(htmlFile);
// A folder (will be created by the component) without any drive letters, only the folder as "myfolder".
f.HtmlOptions.ImageSubFolder = String.Format("{0}_images", Path.GetFileNameWithoutExtension(pdfFile));
// Auto - the same image format as in the source PDF;
// 'Jpeg' to make the document size less;
// 'PNG' to keep the highest quality, but the highest size too.
f.EmbeddedImagesFormat = PdfFocus.eImageFormat.Auto;
// How to store images: Inside HTML document as base64 images or as linked separate image files.
f.HtmlOptions.IncludeImageInHtml = false;
// Set <title>...</title>
f.HtmlOptions.Title = String.Format("This HTML was converted from {0}.", Path.GetFileName(pdfFile));
f.OpenPdf(pdfFile);
if (f.PageCount > 0)
{
int res = f.ToHtml(htmlFile);
// Open the result for demonstration purposes.
if (res == 0)
{
System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(htmlFile)
{
UseShellExecute = true
});
}
}
}
public static void ConvertToWord(object targ)
{
TArgument targum = (TArgument)targ;
string pdfFile = targum.PdfFile;
int page = targum.PageNumber;
string docxFile = Path.GetFileNameWithoutExtension(pdfFile) + ".docx";
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
f.WordOptions.Format = PdfFocus.CWordOptions.eWordDocument.Docx;
f.WordOptions.RenderMode = PdfFocus.CWordOptions.eRenderMode.Flowing;
f.OpenPdf(pdfFile);
bool done = false;
if (f.PageCount > 0)
{
if (page >= f.PageCount)
{
page = 1;
}
if (f.ToWord(docxFile, page, page) == 0)
{
done = true;
}
f.ClosePdf();
}
if (done)
{
Console.WriteLine("{0}\t - Done!", Path.GetFileName(pdfFile));
System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(docxFile)
{
UseShellExecute = true
});
}
else
{
Console.WriteLine("{0}\t - Error!", Path.GetFileName(pdfFile));
}
}
public static void ConvertToXml(object targ)
{
TArgument targum = (TArgument)targ;
string pdfFile = targum.PdfFile;
int page = targum.PageNumber;
string xmlFile = Path.GetFileNameWithoutExtension(pdfFile) + ".xml";
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// Let's convert all data (textual and tabular) to XML.
f.XmlOptions.ConvertNonTabularDataToSpreadsheet = true;
f.OpenPdf(pdfFile);
bool done = false;
if (f.PageCount > 0)
{
if (page >= f.PageCount)
{
page = 1;
}
if (f.ToXml(xmlFile, page, page) == 0)
{
done = true;
}
f.ClosePdf();
}
if (done)
{
Console.WriteLine("{0}\t - Done!", Path.GetFileName(pdfFile));
System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(xmlFile)
{
UseShellExecute = true
});
}
else
{
Console.WriteLine("{0}\t - Error!", Path.GetFileName(pdfFile));
}
}
public override void convertToWord(string path, string tempFolder)
{
nameDocFile = Path.GetFileNameWithoutExtension(path) + extension;
convertFileName = nameDocFile;
string pathDocFile = Path.Combine(tempFolder, nameDocFile);
pathWordFile = pathDocFile;
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
if (File.Exists(pathDocFile))
{
File.Delete(pathDocFile);
}
f.OpenPdf(path);
if (f.PageCount > 0)
{
f.WordOptions.Format = SautinSoft.PdfFocus.CWordOptions.eWordDocument.Docx;
f.ToWord(pathDocFile);
}
}
public static void ConvertToHtml(object targ)
{
TArgument targum = (TArgument)targ;
string pdfFile = targum.PdfFile;
int page = targum.PageNumber;
string htmlFile = targum.HtmlFile;
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
f.EmbeddedImagesFormat = PdfFocus.eImageFormat.Auto;
f.HtmlOptions.IncludeImageInHtml = false;
f.HtmlOptions.ImageSubFolder = String.Format("{0}_images", Path.GetFileNameWithoutExtension(pdfFile));
f.HtmlOptions.Title = String.Format("This document was produced from {0}.", Path.GetFileName(pdfFile));
f.HtmlOptions.ImageFileName = "picture";
f.OpenPdf(pdfFile);
bool done = false;
if (f.PageCount > 0)
{
if (page >= f.PageCount)
{
page = 1;
}
if (f.ToHtml(htmlFile, page, page) == 0)
{
done = true;
}
f.ClosePdf();
}
if (done)
{
Console.WriteLine("{0}\t - Done!", Path.GetFileName(pdfFile));
}
else
{
Console.WriteLine("{0}\t - Error!", Path.GetFileName(pdfFile));
}
}
public static void ConvertToText(object targ)
{
TArgument targum = (TArgument)targ;
string pdfFile = targum.PdfFile;
int page = targum.PageNumber;
string textFile = Path.GetFileNameWithoutExtension(pdfFile) + ".txt";
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
f.WordOptions.ShowInvisibleText = true;
f.OpenPdf(pdfFile);
bool done = false;
if (f.PageCount > 0)
{
if (page >= f.PageCount)
{
page = 1;
}
if (f.ToText(textFile, page, page) == 0)
{
done = true;
}
f.ClosePdf();
}
if (done)
{
Console.WriteLine("{0}\t - Done!", Path.GetFileName(pdfFile));
System.Diagnostics.Process.Start(new System.Diagnostics.ProcessStartInfo(textFile)
{
UseShellExecute = true
});
}
else
{
Console.WriteLine("{0}\t - Error!", Path.GetFileName(pdfFile));
}
}
public static void ConvertToHtml(object targ)
{
TArgument targum = (TArgument)targ;
string pdfFile = targum.PdfFile;
int page = targum.PageNumber;
string htmlFile = Path.ChangeExtension(pdfFile, ".html");
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
f.HtmlOptions.ImageType = PdfFocus.CHtmlOptions.eHtmlImageType.Png;
f.HtmlOptions.IncludeImageInHtml = false;
f.HtmlOptions.ImageSubFolder = String.Format("{0}_images", Path.GetFileNameWithoutExtension(pdfFile));
f.HtmlOptions.ImageFileName = "pict";
f.OpenPdf(pdfFile);
bool done = false;
if (f.PageCount > 0)
{
if (page >= f.PageCount)
{
page = 1;
}
if (f.ToHtml(htmlFile, page, page) == 0)
{
done = true;
}
f.ClosePdf();
}
if (done)
{
Console.WriteLine("{0}\t - Done!", Path.GetFileName(pdfFile));
}
else
{
Console.WriteLine("{0}\t - Error!", Path.GetFileName(pdfFile));
}
}
static void Main(string[] args)
{
string pathToPdf = @"d:\simple text.pdf";
string pathToWord = @"d:\result.doc";
//Convert PDF file to Word file
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
f.OpenPdf(pathToPdf);
if (f.PageCount > 0)
{
int result = f.ToWord(pathToWord);
//Show Word document
if (result == 0)
{
System.Diagnostics.Process.Start(pathToWord);
}
}
}
public static void ConvertToText(object targ)
{
TArgument targum = (TArgument)targ;
string pdfFile = targum.PdfFile;
int page = targum.PageNumber;
string textFile = Path.ChangeExtension(pdfFile, ".txt");
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
f.WordOptions.ShowInvisibleText = true;
f.OpenPdf(pdfFile);
bool done = false;
if (f.PageCount > 0)
{
if (page >= f.PageCount)
{
page = 1;
}
if (f.ToText(textFile, page, page) == 0)
{
done = true;
}
f.ClosePdf();
}
if (done)
{
Console.WriteLine("{0}\t - Done!", Path.GetFileName(pdfFile));
}
else
{
Console.WriteLine("{0}\t - Error!", Path.GetFileName(pdfFile));
}
}
static void Main(string[] args)
{
//Convert PDF file to Multipage TIFF file
SautinSoft.PdfFocus f = new SautinSoft.PdfFocus();
// You may download the latest version of SDK here:
// www.sautinsoft.com/products/pdf-focus/download.php
string pdfPath = @"d:\Tempos\table.pdf";
string tiffPath = @"d:\Tempos\table.tiff";
f.OpenPdf(pdfPath);
if (f.PageCount > 0)
{
f.ImageOptions.Dpi = 120;
if (f.ToMultipageTiff(tiffPath) == 0)
{
System.Diagnostics.Process.Start(tiffPath);
}
}
}