public bool Validate(InternalSpreadsheet sheet, IWebSocketProgress progress)
{
// An export would have several others. But none of them is absolutely required except this one.
// (We could do without it, too, by assuming the first column contains them. But it's helpful to be
// able to recognize spreadsheets created without any knowledge at all of the expected content.)
// Note: depending on row content, this problem may be detected earlier in SpreadsheetIO while
// converting the file to an InternalSpreadsheet.
var rowTypeColumn = sheet.GetColumnForTag(InternalSpreadsheet.RowTypeColumnLabel);
if (rowTypeColumn < 0)
{
progress.MessageWithoutLocalizing(MissingHeaderMessage, ProgressKind.Error);
return(false);
}
var inputRows = sheet.ContentRows.ToList();
if (!inputRows.Any(r => r.GetCell(rowTypeColumn).Content.StartsWith("[")))
{
progress.MessageWithoutLocalizing("This spreadsheet has no data that Bloom knows how to import. Did you follow the standard format for Bloom spreadsheets?", ProgressKind.Warning);
// Technically this isn't a fatal error. We could just let the main loop do nothing. But reporting it as one avoids creating a spurious backup.
return(false);
}
return(true);
}
private void UpdateDataDivFromRow(ContentRow currentRow, string dataBookLabel)
{
if (dataBookLabel.Contains("branding"))
{
return; // branding data-div elements are complex and difficult and determined by current collection state
}
// Only a few of these are worth reporting
string whatsUpdated = null;
switch (dataBookLabel)
{
case "coverImage":
whatsUpdated = "the image on the cover";
break;
case "bookTitle":
whatsUpdated = "the book title";
break;
case "copyright":
whatsUpdated = "copyright information";
break;
}
if (whatsUpdated != null)
{
Progress($"Updating {whatsUpdated}.");
}
var xPath = "div[@data-book=\"" + dataBookLabel + "\"]";
var matchingNodes = _dataDivElement.SelectNodes(xPath);
XmlElement templateNode;
bool templateNodeIsNew = false;
if (matchingNodes.Count > 0)
{
templateNode = (XmlElement)matchingNodes[0];
}
else
{
templateNodeIsNew = true;
templateNode = _destinationDom.RawDom.CreateElement("div");
templateNode.SetAttribute("data-book", dataBookLabel);
}
var imageSrcCol = _sheet.GetColumnForTag(InternalSpreadsheet.ImageSourceColumnLabel);
var imageSrc = imageSrcCol >= 0 ? currentRow.GetCell(imageSrcCol).Content : null; // includes "images" folder
var imageFileName = Path.GetFileName(imageSrc);
bool specificLanguageContentFound = false;
bool asteriskContentFound = false;
//Whether or not a data-book div has a src attribute, we found that the innerText is used to set the
//src of the image in the actual pages of the document, though we haven't found a case where they differ.
//So during export we put the innerText into the image source column, and want to put it into
//both src and innertext on import, unless the element is in the noSrcAttribute list
if (imageFileName.Length > 0)
{
templateNode.SetAttribute("lang", "*");
templateNode.InnerText = imageFileName;
if (!SpreadsheetExporter.DataDivImagesWithNoSrcAttributes.Contains(dataBookLabel))
{
templateNode.SetAttribute("src", imageFileName);
}
if (templateNodeIsNew)
{
AddDataBookNode(templateNode);
}
if (_pathToSpreadsheetFolder != null)
{
// Make sure the image gets copied over too.
var fullSpreadsheetPath = Path.Combine(_pathToSpreadsheetFolder, imageSrc);
CopyImageFileToDestination(imageFileName, fullSpreadsheetPath);
}
}
else //This is not an image node
{
if (dataBookLabel.Equals("coverImage"))
{
Warn("No cover image found");
}
foreach (string lang in _sheet.Languages)
{
var langVal = currentRow.GetCell(_sheet.GetRequiredColumnForLang(lang)).Content;
var langXPath = "div[@data-book=\"" + dataBookLabel + "\" and @lang=\"" + lang + "\"]";
var langMatchingNodes = _dataDivElement.SelectNodes(langXPath).Cast <XmlElement>();
if (!string.IsNullOrEmpty(langVal))
{
//Found content in spreadsheet for this language and row
if (lang.Equals("*"))
{
asteriskContentFound = true;
}
else
{
specificLanguageContentFound = true;
}
if (langMatchingNodes.Count() > 0) //Found matching node in dom. Update node.
{
XmlElement matchingNode = langMatchingNodes.First();
matchingNode.InnerXml = langVal;
if (langMatchingNodes.Count() > 1)
{
Warn("Found more than one " + dataBookLabel + " element for language "
+ lang + " in the book dom. Only the first will be updated.");
}
}
else //No node for this language and data-book. Create one from template and add.
{
XmlElement newNode = (XmlElement)templateNode.CloneNode(deep: true);
newNode.SetAttribute("lang", lang);
newNode.InnerXml = langVal;
AddDataBookNode(newNode);
}
}
else //Spreadsheet cell for this row and language is empty. Remove the corresponding node if present.
{
foreach (XmlNode n in langMatchingNodes.ToArray())
{
_dataDivElement.RemoveChild(n);
}
}
}
if (RemoveOtherLanguages)
{
HtmlDom.RemoveOtherLanguages(matchingNodes.Cast <XmlElement>().ToList(), _dataDivElement, _sheet.Languages);
}
if (asteriskContentFound && specificLanguageContentFound)
{
Warn(dataBookLabel + " information found in both * language column and other language column(s)");
}
}
}
public static void WriteSpreadsheet(InternalSpreadsheet spreadsheet, string outputPath, bool retainMarkup, IWebSocketProgress progress = null)
{
using (var package = new ExcelPackage())
{
var worksheet = package.Workbook.Worksheets.Add("BloomBook");
worksheet.DefaultColWidth = languageColumnWidth;
for (int i = 1; i <= spreadsheet.StandardLeadingColumns.Length; i++)
{
worksheet.Column(i).Width = standardLeadingColumnWidth;
}
var imageSourceColumn = spreadsheet.GetColumnForTag(InternalSpreadsheet.ImageSourceColumnLabel);
var imageThumbnailColumn = spreadsheet.GetColumnForTag(InternalSpreadsheet.ImageThumbnailColumnLabel);
// Apparently the width is in some approximation of 'characters'. This empirically determined
// conversion factor seems to do a pretty good job.
worksheet.Column(imageThumbnailColumn + 1).Width = defaultImageWidth / 6.88;
int r = 0;
foreach (var row in spreadsheet.AllRows())
{
r++;
for (var c = 0; c < row.Count; c++)
{
// Enhance: Excel complains about cells that contain pure numbers
// but are created as strings. We could possibly tell it that cells
// that contain simple numbers can be treated accordingly.
// It might be helpful for some uses of the group-on-page-index
// if Excel knew to treat them as numbers.
var sourceCell = row.GetCell(c);
var content = sourceCell.Content;
// Parse xml for markdown formatting on language columns,
// Display formatting in excel spreadsheet
ExcelRange currentCell = worksheet.Cells[r, c + 1];
if (!string.IsNullOrEmpty(sourceCell.Comment))
{
// Second arg is supposed to be the author.
currentCell.AddComment(sourceCell.Comment, "Bloom");
}
if (!retainMarkup &&
IsWysiwygFormattedColumn(row, c) &&
IsWysiwygFormattedRow(row))
{
MarkedUpText markedUpText = MarkedUpText.ParseXml(content);
if (markedUpText.HasFormatting)
{
currentCell.IsRichText = true;
foreach (MarkedUpTextRun run in markedUpText.Runs)
{
if (!run.Text.Equals(""))
{
ExcelRichText text = currentCell.RichText.Add(run.Text);
text.Bold = run.Bold;
text.Italic = run.Italic;
text.UnderLine = run.Underlined;
if (run.Superscript)
{
text.VerticalAlign = ExcelVerticalAlignmentFont.Superscript;
}
}
}
}
else
{
currentCell.Value = markedUpText.PlainText();
}
}
else
{
// Either the retainMarkup flag is set, or this is not book text. It could be header or leading column.
// Generally, we just want to blast our cell content into the spreadsheet cell.
// However, there are cases where we put an error message in an image thumbnail cell when processing the image path.
// We don't want to overwrite these. An easy way to prevent it is to not overwrite any cell that already has content.
// Since export is creating a new spreadsheet, cells we want to write will always be empty initially.
if (currentCell.Value == null)
{
currentCell.Value = content;
}
}
//Embed any images in the excel file
if (c == imageSourceColumn)
{
var imageSrc = sourceCell.Content;
// if this row has an image source value that is not a header
if (imageSrc != "" && !row.IsHeader)
{
var sheetFolder = Path.GetDirectoryName(outputPath);
var imagePath = Path.Combine(sheetFolder, imageSrc);
//Images show up in the cell 1 row greater and 1 column greater than assigned
//So this will put them in row r, column imageThumbnailColumn+1 like we want
var rowHeight = embedImage(imagePath, r - 1, imageThumbnailColumn);
worksheet.Row(r).Height = rowHeight * 72 / 96 + 3; //so the image is visible; height seems to be points
}
}
}
if (row is HeaderRow)
{
using (ExcelRange rng = GetRangeForRow(worksheet, r))
rng.Style.Font.Bold = true;
}
if (row.Hidden)
{
worksheet.Row(r).Hidden = true;
SetBackgroundColorOfRow(worksheet, r, InternalSpreadsheet.HiddenColor);
}
else if (row.BackgroundColor != default(Color))
{
SetBackgroundColorOfRow(worksheet, r, row.BackgroundColor);
}
}
worksheet.Cells[1, 1, r, spreadsheet.ColumnCount].Style.WrapText = true;
int embedImage(string imageSrcPath, int rowNum, int colNum)
{
int finalHeight = 30; // a reasonable default if we don't manage to embed an image.
try
{
using (Image image = Image.FromFile(imageSrcPath))
{
string imageName = Path.GetFileNameWithoutExtension(imageSrcPath);
var origImageHeight = image.Size.Height;
var origImageWidth = image.Size.Width;
int finalWidth = defaultImageWidth;
finalHeight = (int)(finalWidth * origImageHeight / origImageWidth);
var size = new Size(finalWidth, finalHeight);
using (Image thumbnail = ImageUtils.ResizeImageIfNecessary(size, image, false))
{
var excelImage = worksheet.Drawings.AddPicture(imageName, thumbnail);
excelImage.SetPosition(rowNum, 2, colNum, 2);
}
}
}
catch (Exception)
{
string errorText;
if (!RobustFile.Exists(imageSrcPath))
{
errorText = "Missing";
}
else if (Path.GetExtension(imageSrcPath).ToLowerInvariant().Equals(".svg"))
{
errorText = "Can't display SVG";
}
else
{
errorText = "Bad image file";
}
progress?.MessageWithoutLocalizing(errorText + ": " + imageSrcPath);
worksheet.Cells[r, imageThumbnailColumn + 1].Value = errorText;
}
return(Math.Max(finalHeight, 30));
}
foreach (var iColumn in spreadsheet.HiddenColumns)
{
// This is pretty yucky... our internal spreadsheet is all 0-based, but the EPPlus library is all 1-based...
var iColumn1Based = iColumn + 1;
worksheet.Column(iColumn1Based).Hidden = true;
SetBackgroundColorOfColumn(worksheet, iColumn1Based, InternalSpreadsheet.HiddenColor);
}
try
{
RobustFile.Delete(outputPath);
var xlFile = new FileInfo(outputPath);
package.SaveAs(xlFile);
}
catch (IOException ex) when((ex.HResult & 0x0000FFFF) == 32) //ERROR_SHARING_VIOLATION
{
Console.WriteLine("Writing Spreadsheet failed. Do you have it open in another program?");
Console.WriteLine(ex.Message);
Console.WriteLine(ex.StackTrace);
progress?.Message("Spreadsheet.SpreadsheetLocked", "",
"Bloom could not write to the spreadsheet because another program has it locked. Do you have it open in another program?",
ProgressKind.Error);
}
catch (Exception ex)
{
progress?.Message("Spreadsheet.ExportFailed", "",
"Export failed: " + ex.Message,
ProgressKind.Error);
}
}
}
public void SetCell(string columnName, string content)
{
int index = Spreadsheet.GetColumnForTag(columnName);
SetCell(index, content);
}