private static void ProcessWordDocument(
string outputDocumentFullName,
string sourceLanguage,
string targetLanguage)
{
using (WordprocessingDocument doc = WordprocessingDocument.Open(outputDocumentFullName, true))
{
OpenXmlPowerTools.SimplifyMarkupSettings settings = new OpenXmlPowerTools.SimplifyMarkupSettings
{
AcceptRevisions = true,
NormalizeXml = true, //setting this to false reduces translation quality, but if true some documents have XML format errors when opening
RemoveBookmarks = true,
RemoveComments = true,
RemoveContentControls = true,
RemoveEndAndFootNotes = true,
RemoveFieldCodes = true,
RemoveGoBackBookmark = true,
//RemoveHyperlinks = false,
RemoveLastRenderedPageBreak = true,
RemoveMarkupForDocumentComparison = true,
RemovePermissions = false,
RemoveProof = true,
RemoveRsidInfo = true,
RemoveSmartTags = true,
RemoveSoftHyphens = true,
RemoveWebHidden = true,
ReplaceTabsWithSpaces = false
};
OpenXmlPowerTools.MarkupSimplifier.SimplifyMarkup(doc, settings);
}
List <DocumentFormat.OpenXml.Wordprocessing.Text> texts = new List <DocumentFormat.OpenXml.Wordprocessing.Text>();
using (WordprocessingDocument doc = WordprocessingDocument.Open(outputDocumentFullName, true))
{
var body = doc.MainDocumentPart.Document.Body;
texts.AddRange(body.Descendants <DocumentFormat.OpenXml.Wordprocessing.Text>().Where(text => !String.IsNullOrEmpty(text.Text) && text.Text.Length > 0));
var headers = doc.MainDocumentPart.HeaderParts.Select(p => p.Header);
foreach (var header in headers)
{
texts.AddRange(header.Descendants <DocumentFormat.OpenXml.Wordprocessing.Text>().Where(text => !String.IsNullOrEmpty(text.Text) && text.Text.Length > 0));
}
var footers = doc.MainDocumentPart.FooterParts.Select(p => p.Footer);
foreach (var footer in footers)
{
texts.AddRange(footer.Descendants <DocumentFormat.OpenXml.Wordprocessing.Text>().Where(text => !String.IsNullOrEmpty(text.Text) && text.Text.Length > 0));
}
var exceptions = new ConcurrentQueue <Exception>();
// Extract Text for Translation
var batch = texts.Select(text => text.Text);
// Do Translation
var batches = SplitList(batch, TranslationServiceFacade.maxelements, TranslationServiceFacade.maxrequestsize);
Parallel.For(
0,
batches.Count(),
new ParallelOptions {
MaxDegreeOfParallelism = 1
},
l =>
{
try
{
var translationOutput = TranslationServiceFacade.TranslateArray(
batches[l].ToArray(),
sourceLanguage,
targetLanguage);
int batchStartIndexInDocument = 0;
for (int i = 0; i < l; i++)
{
batchStartIndexInDocument = batchStartIndexInDocument + batches[i].Count();
}
// Apply translated batch to document
for (int j = 0; j < translationOutput.Length; j++)
{
int indexInDocument = j + batchStartIndexInDocument + 1;
var newValue = translationOutput[j];
texts.Take(indexInDocument).Last().Text = newValue;
}
}
catch (Exception ex)
{
exceptions.Enqueue(ex);
}
});
// Throw the exceptions here after the loop completes.
if (exceptions.Count > 0)
{
throw new AggregateException(exceptions);
}
//doc.MainDocumentPart.PutXDocument();
}
}
public static Boolean FillTable(string fileName, Dictionary <string, string> varDict, Dictionary <string, BildMitKommentar> dictBilder)
{
Boolean bRet = false;
szError = "";
try
{
File.Copy(Config.current.szWordTemplate, fileName, true);
using (WordprocessingDocument wordDocument
= WordprocessingDocument.Open(fileName, true))
{
OpenXmlPowerTools.SimplifyMarkupSettings settings = new OpenXmlPowerTools.SimplifyMarkupSettings
{
RemoveComments = true,
RemoveContentControls = true,
RemoveEndAndFootNotes = true,
RemoveFieldCodes = false,
RemoveLastRenderedPageBreak = true,
RemovePermissions = true,
RemoveProof = true,
RemoveRsidInfo = true,
RemoveSmartTags = true,
RemoveSoftHyphens = true,
// ReplaceTabsWithSpaces = true,
RemoveGoBackBookmark = true,
RemoveBookmarks = true,
RemoveMarkupForDocumentComparison = true,
RemoveWebHidden = true
};
OpenXmlPowerTools.MarkupSimplifier.SimplifyMarkup(wordDocument, settings);
MainDocumentPart mainPart = wordDocument.MainDocumentPart;
var paraList = mainPart.Document.Descendants <Paragraph>();
foreach (var para in paraList)
{
ReplaceVar(para, varDict);
}
var headerPart = mainPart.HeaderParts.FirstOrDefault();
paraList = headerPart.Header.Descendants <Paragraph>();
foreach (var para in paraList)
{
ReplaceVar(para, varDict);
}
var footerPart = mainPart.FooterParts.FirstOrDefault();
paraList = footerPart.Footer.Descendants <Paragraph>();
foreach (var para in paraList)
{
ReplaceVar(para, varDict);
}
var tableList = mainPart.Document.Descendants <Table>();
foreach (var table in tableList)
{
var rowList = table.Descendants <TableRow>();
foreach (var tr in rowList)
{
var tc = tr.Descendants <TableCell>().FirstOrDefault();
// get the first para
var p = tc.Descendants <Paragraph>().FirstOrDefault();
if (p.InnerText.Trim().Equals("${PictureTable}", StringComparison.InvariantCultureIgnoreCase))
{
AddInTable(mainPart, table, tr, dictBilder, varDict);
}
}
}
wordDocument.Save();
bRet = true;
}
}
catch (Exception ex)
{
szError = "FillTable: " + ex.ToString();
Logging.AddError(szError);
}
return(bRet);
}
/// <summary>
/// Translates an Word document
///
/// Based on method `ProcessWordDocument` (line 726 onwards) in
/// TranslationAssistant.Business/DocumentTranslationManager.cs in
/// MicrosoftTranslator/DocumentTranslator
/// </summary>
public async Task TranslateDocument(MemoryStream memoryStream, ITextTranslator textTranslator, string to, string from = null, bool ignoreHidden = false)
{
using (WordprocessingDocument doc = WordprocessingDocument.Open(memoryStream, true))
{
// Simply the Word document mark-up
OpenXmlPowerTools.SimplifyMarkupSettings settings = new OpenXmlPowerTools.SimplifyMarkupSettings
{
AcceptRevisions = true,
NormalizeXml = true, //setting this to false reduces translation quality, but if true some documents have XML format errors when opening
RemoveBookmarks = true,
RemoveComments = true,
RemoveContentControls = true,
RemoveEndAndFootNotes = true,
RemoveFieldCodes = true,
RemoveGoBackBookmark = true,
//RemoveHyperlinks = false,
RemoveLastRenderedPageBreak = true,
RemoveMarkupForDocumentComparison = true,
RemovePermissions = false,
RemoveProof = true,
RemoveRsidInfo = true,
RemoveSmartTags = true,
RemoveSoftHyphens = true,
RemoveWebHidden = true,
ReplaceTabsWithSpaces = false
};
OpenXmlPowerTools.MarkupSimplifier.SimplifyMarkup(doc, settings);
}
var texts = new List <Text>();
using (WordprocessingDocument doc = WordprocessingDocument.Open(memoryStream, true))
{
// Find all text nodes in the document (body, headers & footers)
var body = doc.MainDocumentPart.Document.Body;
texts.AddRange(body.Descendants <Text>()
.Where(text => !String.IsNullOrEmpty(text.Text) && text.Text.Length > 0));
var headers = doc.MainDocumentPart.HeaderParts.Select(p => p.Header);
foreach (var header in headers)
{
texts.AddRange(header.Descendants <Text>().Where(text =>
!String.IsNullOrEmpty(text.Text) && text.Text.Length > 0));
}
var footers = doc.MainDocumentPart.FooterParts.Select(p => p.Footer);
foreach (var footer in footers)
{
texts.AddRange(footer.Descendants <Text>().Where(text =>
!String.IsNullOrEmpty(text.Text) && text.Text.Length > 0));
}
if (ignoreHidden)
{
texts.RemoveAll(t => t.Parent.Descendants <Vanish>().Any());
}
// Extract text strings for translation
var values = texts.Select(text => text.Text);
// Do Translation
var translations = await textTranslator.TranslateTexts(values, to, from);
// Apply translations to document by iterating through both lists and
// replacing the original with its translation
using (var textsEnumerator = texts.GetEnumerator())
{
using (var translationsEnumerator = translations.GetEnumerator())
{
while (textsEnumerator.MoveNext() && translationsEnumerator.MoveNext())
{
textsEnumerator.Current.Text = translationsEnumerator.Current;
}
}
}
}
}
public DocxConversionData CreateCleanedCopy(string fullPath, bool ignoreHidden)
{
if (string.IsNullOrEmpty(fullPath))
{
throw new ArgumentNullException("path variable cannot be null or empty");
}
// ensure we have a valid path to a docx file
if (!validator.IsValid(fullPath))
{
return(null);
}
var origFileData = new DocxConversionData(fullPath);
var cleanedFileName = origFileData.FileName + "_cleaned" + origFileData.FileExtension;
var cleanedFileData = new DocxConversionData(Path.Join(origFileData.FilePath, cleanedFileName));
try
{
// ensure there is not an orphaned copy already on disk
File.Delete(cleanedFileData.FullPath);
// Make a copy of the original using the "cleaned" file name
File.Copy(origFileData.FullPath, cleanedFileData.FullPath);
//Clean Document of extra tags
using (WordprocessingDocument doc = WordprocessingDocument.Open(cleanedFileData.FullPath, true))
{
OpenXmlPowerTools.SimplifyMarkupSettings settings = new OpenXmlPowerTools.SimplifyMarkupSettings
{
AcceptRevisions = true,
NormalizeXml = false, //setting this to false reduces translation quality, but if true some documents have XML format errors when opening
RemoveBookmarks = true,
RemoveComments = true,
RemoveContentControls = true,
RemoveEndAndFootNotes = true,
RemoveFieldCodes = false, //FieldCode remove pagination
RemoveGoBackBookmark = true,
//RemoveHyperlinks = false,
RemoveLastRenderedPageBreak = true,
RemoveMarkupForDocumentComparison = true,
RemovePermissions = false,
RemoveProof = true,
RemoveRsidInfo = true,
RemoveSmartTags = true,
RemoveSoftHyphens = true,
RemoveWebHidden = true,
ReplaceTabsWithSpaces = false
};
OpenXmlPowerTools.MarkupSimplifier.SimplifyMarkup(doc, settings);
//Extract Strings for translation
var textBlocks = textExtractor.ExtractText(doc, ignoreHidden);
//var textBlocks = textExtractor.ExtractParagraph(doc, ignoreHidden);
cleanedFileData.Strings2Translate = textBlocks.Select(para => para.InnerText);
}
//Total the number of strings for translation
cleanedFileData.TotalStrings2Translate = cleanedFileData.Strings2Translate.Count();
if (File.Exists(cleanedFileData.FullPath))
{
cleanedFileData.Messages.Add("Success");
}
else
{
cleanedFileData.Messages.Add("Docx data failed to build properly.");
}
}
catch (Exception e)
{
cleanedFileData.Messages.Add(e.Message);
}
return(cleanedFileData);
}
