static void Main(string[] args)
{
var n = DateTime.Now;
var tempDi = new DirectoryInfo(string.Format("ExampleOutput-{0:00}-{1:00}-{2:00}-{3:00}{4:00}{5:00}", n.Year - 2000, n.Month, n.Day, n.Hour, n.Minute, n.Second));
tempDi.Create();
WmlDocument originalWml = new WmlDocument("../../Original.docx");
List <WmlRevisedDocumentInfo> revisedDocumentInfoList = new List <WmlRevisedDocumentInfo>()
{
new WmlRevisedDocumentInfo()
{
RevisedDocument = new WmlDocument("../../RevisedByBob.docx"),
Revisor = "Bob",
Color = Color.LightBlue,
},
new WmlRevisedDocumentInfo()
{
RevisedDocument = new WmlDocument("../../RevisedByMary.docx"),
Revisor = "Mary",
Color = Color.LightYellow,
},
};
WmlComparerSettings settings = new WmlComparerSettings();
WmlDocument consolidatedWml = WmlComparer.Consolidate(
originalWml,
revisedDocumentInfoList,
settings);
consolidatedWml.SaveAs(Path.Combine(tempDi.FullName, "Consolidated.docx"));
}
private static void AnnotateElementWithProps(
OpenXmlPart part,
XElement element,
List <ComparisonUnitAtom> comparisonUnitAtomList,
XName[] childElementPropertyNames,
WmlComparerSettings settings)
{
IEnumerable <XElement> runChildrenToProcess;
if (childElementPropertyNames == null)
{
runChildrenToProcess = element.Elements();
}
else
{
runChildrenToProcess = element
.Elements()
.Where(e => !childElementPropertyNames.Contains(e.Name));
}
foreach (XElement item in runChildrenToProcess)
{
CreateComparisonUnitAtomListRecurse(part, item, comparisonUnitAtomList, settings);
}
}
static void Main(string[] args)
{
var n = DateTime.Now;
var tempDi = new DirectoryInfo(string.Format("ExampleOutput-{0:00}-{1:00}-{2:00}-{3:00}{4:00}{5:00}", n.Year - 2000, n.Month, n.Day, n.Hour, n.Minute, n.Second));
tempDi.Create();
WmlComparerSettings settings = new WmlComparerSettings();
WmlDocument result = WmlComparer.Compare(
new WmlDocument("../../Source1.docx"),
new WmlDocument("../../Source2.docx"),
settings);
result.SaveAs(Path.Combine(tempDi.FullName, "Compared.docx"));
var revisions = WmlComparer.GetRevisions(result, settings);
foreach (var rev in revisions)
{
Console.WriteLine("Author: " + rev.Author);
Console.WriteLine("Revision type: " + rev.RevisionType);
Console.WriteLine("Revision text: " + rev.Text);
Console.WriteLine();
}
}
private static WmlDocument HashBlockLevelContent(
WmlDocument source,
WmlDocument sourceAfterProc,
WmlComparerSettings settings)
{
using (var msSource = new MemoryStream())
using (var msAfterProc = new MemoryStream())
{
msSource.Write(source.DocumentByteArray, 0, source.DocumentByteArray.Length);
msAfterProc.Write(sourceAfterProc.DocumentByteArray, 0, sourceAfterProc.DocumentByteArray.Length);
using (WordprocessingDocument wDocSource = WordprocessingDocument.Open(msSource, true))
using (WordprocessingDocument wDocAfterProc = WordprocessingDocument.Open(msAfterProc, true))
{
// create Unid dictionary for source
XDocument sourceMainXDoc = wDocSource.MainDocumentPart.GetXDocument();
XElement sourceMainRoot = sourceMainXDoc.Root ?? throw new ArgumentException();
Dictionary <string, XElement> sourceUnidDict = sourceMainRoot
.Descendants()
.Where(d => d.Name == W.p || d.Name == W.tbl || d.Name == W.tr)
.ToDictionary(d => (string)d.Attribute(PtOpenXml.Unid));
XDocument afterProcMainXDoc = wDocAfterProc.MainDocumentPart.GetXDocument();
XElement afterProcMainRoot = afterProcMainXDoc.Root ?? throw new ArgumentException();
IEnumerable <XElement> blockLevelElements = afterProcMainRoot
.Descendants()
.Where(d => d.Name == W.p || d.Name == W.tbl || d.Name == W.tr);
foreach (XElement blockLevelContent in blockLevelElements)
{
var cloneBlockLevelContentForHashing = (XElement)CloneBlockLevelContentForHashing(
wDocAfterProc.MainDocumentPart,
blockLevelContent,
true,
settings);
string shaString = cloneBlockLevelContentForHashing
.ToString(SaveOptions.DisableFormatting)
.Replace(" xmlns=\"http://schemas.openxmlformats.org/wordprocessingml/2006/main\"", "");
string sha1Hash = WmlComparerUtil.SHA1HashStringForUTF8String(shaString);
var thisUnid = (string)blockLevelContent.Attribute(PtOpenXml.Unid);
if (thisUnid != null)
{
if (sourceUnidDict.ContainsKey(thisUnid))
{
XElement correlatedBlockLevelContent = sourceUnidDict[thisUnid];
correlatedBlockLevelContent.Add(new XAttribute(PtOpenXml.CorrelatedSHA1Hash, sha1Hash));
}
}
}
wDocSource.MainDocumentPart.PutXDocument();
}
var sourceWithCorrelatedSHA1Hash = new WmlDocument(source.FileName, msSource.ToArray());
return(sourceWithCorrelatedSHA1Hash);
}
}
private static void SaveDocumentIfDesired(WmlDocument source, string name, WmlComparerSettings settings)
{
if (SaveIntermediateFilesForDebugging && settings.DebugTempFileDi != null)
{
var fileInfo = new FileInfo(Path.Combine(settings.DebugTempFileDi.FullName, name));
source.SaveAs(fileInfo.FullName);
}
}
private static List <ComparisonUnitAtom> CreateComparisonUnitAtomListInternal(
OpenXmlPart part,
XElement contentParent,
WmlComparerSettings settings)
{
var comparisonUnitAtomList = new List <ComparisonUnitAtom>();
CreateComparisonUnitAtomListRecurse(part, contentParent, comparisonUnitAtomList, settings);
return(comparisonUnitAtomList);
}
private static string GetSha1HashStringForElement(XElement contentElement, WmlComparerSettings settings)
{
string text = contentElement.Value;
if (settings.CaseInsensitive)
{
text = text.ToUpper(settings.CultureInfo);
}
return(contentElement.Name.LocalName + text);
}
public ComparisonUnitAtom(
XElement contentElement,
XElement[] ancestorElements,
OpenXmlPart part,
WmlComparerSettings settings)
{
ContentElement = contentElement;
AncestorElements = ancestorElements;
Part = part;
RevTrackElement = GetRevisionTrackingElementFromAncestors(contentElement, AncestorElements);
if (RevTrackElement == null)
{
CorrelationStatus = CorrelationStatus.Equal;
}
else
{
if (RevTrackElement.Name == W.del)
{
CorrelationStatus = CorrelationStatus.Deleted;
}
else if (RevTrackElement.Name == W.ins)
{
CorrelationStatus = CorrelationStatus.Inserted;
}
}
var sha1Hash = (string)contentElement.Attribute(PtOpenXml.SHA1Hash);
if (sha1Hash != null)
{
SHA1Hash = sha1Hash;
}
else
{
string shaHashString = GetSha1HashStringForElement(ContentElement, settings);
SHA1Hash = WmlComparerUtil.SHA1HashStringForUTF8String(shaHashString);
}
}
internal static ComparisonUnitAtom[] CreateComparisonUnitAtomList(
OpenXmlPart part,
XElement contentParent,
WmlComparerSettings settings)
{
VerifyNoInvalidContent(contentParent);
AssignUnidToAllElements(contentParent); // add the Guid id to every element
MoveLastSectPrIntoLastParagraph(contentParent);
ComparisonUnitAtom[] cal = CreateComparisonUnitAtomListInternal(part, contentParent, settings).ToArray();
if (False)
{
var sb = new StringBuilder();
foreach (ComparisonUnitAtom item in cal)
{
sb.Append(item + Environment.NewLine);
}
string sbs = sb.ToString();
TestUtil.NotePad(sbs);
}
return(cal);
}
// The following method must be made internal if we ever turn this part of the partial class
// into its own class.
private static ComparisonUnit[] GetComparisonUnitList(
ComparisonUnitAtom[] comparisonUnitAtomList,
WmlComparerSettings settings)
{
var seed = new Atgbw
{
Key = null,
ComparisonUnitAtomMember = null,
NextIndex = 0
};
IEnumerable <Atgbw> groupingKey = comparisonUnitAtomList
.Rollup(seed, (sr, prevAtgbw, i) =>
{
int?key;
int nextIndex = prevAtgbw.NextIndex;
if (sr.ContentElement.Name == W.t)
{
string chr = sr.ContentElement.Value;
char ch = chr[0];
if (ch == '.' || ch == ',')
{
var beforeIsDigit = false;
if (i > 0)
{
ComparisonUnitAtom prev = comparisonUnitAtomList[i - 1];
if (prev.ContentElement.Name == W.t && char.IsDigit(prev.ContentElement.Value[0]))
{
beforeIsDigit = true;
}
}
var afterIsDigit = false;
if (i < comparisonUnitAtomList.Length - 1)
{
ComparisonUnitAtom next = comparisonUnitAtomList[i + 1];
if (next.ContentElement.Name == W.t && char.IsDigit(next.ContentElement.Value[0]))
{
afterIsDigit = true;
}
}
if (beforeIsDigit || afterIsDigit)
{
key = nextIndex;
}
else
{
nextIndex++;
key = nextIndex;
nextIndex++;
}
}
else if (settings.WordSeparators.Contains(ch))
{
nextIndex++;
key = nextIndex;
nextIndex++;
}
else
{
key = nextIndex;
}
}
else if (WordBreakElements.Contains(sr.ContentElement.Name))
{
nextIndex++;
key = nextIndex;
nextIndex++;
}
else
{
key = nextIndex;
}
return(new Atgbw
{
Key = key,
ComparisonUnitAtomMember = sr,
NextIndex = nextIndex
});
})
.ToArray();
if (False)
{
var sb = new StringBuilder();
foreach (Atgbw item in groupingKey)
{
sb.Append(item.Key + Environment.NewLine);
sb.Append(" " + item.ComparisonUnitAtomMember.ToString(0) + Environment.NewLine);
}
string sbs = sb.ToString();
TestUtil.NotePad(sbs);
}
IEnumerable <IGrouping <int?, Atgbw> > groupedByWords = groupingKey
.GroupAdjacent(gc => gc.Key)
.ToArray();
if (False)
{
var sb = new StringBuilder();
foreach (IGrouping <int?, Atgbw> group in groupedByWords)
{
sb.Append("Group ===== " + @group.Key + Environment.NewLine);
foreach (Atgbw gc in @group)
{
sb.Append(" " + gc.ComparisonUnitAtomMember.ToString(0) + Environment.NewLine);
}
}
string sbs = sb.ToString();
TestUtil.NotePad(sbs);
}
WithHierarchicalGroupingKey[] withHierarchicalGroupingKey = groupedByWords
.Select(g =>
{
string[] hierarchicalGroupingArray = g
.First()
.ComparisonUnitAtomMember
.AncestorElements
.Where(a => ComparisonGroupingElements.Contains(a.Name))
.Select(a => a.Name.LocalName + ":" + (string)a.Attribute(PtOpenXml.Unid))
.ToArray();
return(new WithHierarchicalGroupingKey
{
ComparisonUnitWord = new ComparisonUnitWord(g.Select(gc => gc.ComparisonUnitAtomMember)),
HierarchicalGroupingArray = hierarchicalGroupingArray
});
}
)
.ToArray();
if (False)
{
var sb = new StringBuilder();
foreach (WithHierarchicalGroupingKey group in withHierarchicalGroupingKey)
{
sb.Append("Grouping Array: " +
@group.HierarchicalGroupingArray.Select(gam => gam + " - ").StringConcatenate() +
Environment.NewLine);
foreach (ComparisonUnit gc in @group.ComparisonUnitWord.Contents)
{
sb.Append(" " + gc.ToString(0) + Environment.NewLine);
}
}
string sbs = sb.ToString();
TestUtil.NotePad(sbs);
}
ComparisonUnit[] cul = GetHierarchicalComparisonUnits(withHierarchicalGroupingKey, 0).ToArray();
if (False)
{
string str = ComparisonUnit.ComparisonUnitListToString(cul);
TestUtil.NotePad(str);
}
return(cul);
}
// the following gets a flattened list of ComparisonUnitAtoms, with status indicated in each ComparisonUnitAtom: Deleted, Inserted, or Equal
// for any deleted or inserted rows, we go into the w:trPr properties, and add the appropriate w:ins or w:del element, and therefore
// when generating the document, the appropriate row will be marked as deleted or inserted.
public static List <WmlComparerRevision> GetRevisions(WmlDocument source, WmlComparerSettings settings)
{
using (var ms = new MemoryStream())
{
ms.Write(source.DocumentByteArray, 0, source.DocumentByteArray.Length);
using (WordprocessingDocument wDoc = WordprocessingDocument.Open(ms, true))
{
TestForInvalidContent(wDoc);
RemoveExistingPowerToolsMarkup(wDoc);
XElement contentParent = wDoc.MainDocumentPart.GetXDocument().Root?.Element(W.body);
ComparisonUnitAtom[] atomList =
CreateComparisonUnitAtomList(wDoc.MainDocumentPart, contentParent, settings).ToArray();
if (False)
{
var sb = new StringBuilder();
foreach (ComparisonUnitAtom item in atomList)
{
sb.Append(item + Environment.NewLine);
}
string sbs = sb.ToString();
TestUtil.NotePad(sbs);
}
List <IGrouping <string, ComparisonUnitAtom> > grouped = atomList
.GroupAdjacent(a =>
{
string key = a.CorrelationStatus.ToString();
if (a.CorrelationStatus != CorrelationStatus.Equal)
{
var rt = new XElement(a.RevTrackElement.Name,
new XAttribute(XNamespace.Xmlns + "w",
"http://schemas.openxmlformats.org/wordprocessingml/2006/main"),
a.RevTrackElement.Attributes().Where(a2 => a2.Name != W.id && a2.Name != PtOpenXml.Unid));
key += rt.ToString(SaveOptions.DisableFormatting);
}
return(key);
})
.ToList();
List <IGrouping <string, ComparisonUnitAtom> > revisions = grouped
.Where(k => k.Key != "Equal")
.ToList();
if (False)
{
var sb = new StringBuilder();
foreach (IGrouping <string, ComparisonUnitAtom> item in revisions)
{
sb.Append(item.Key + Environment.NewLine);
}
string sbs = sb.ToString();
TestUtil.NotePad(sbs);
}
List <WmlComparerRevision> mainDocPartRevisionList = revisions
.Select(rg =>
{
var rev = new WmlComparerRevision();
if (rg.Key.StartsWith("Inserted"))
{
rev.RevisionType = WmlComparerRevisionType.Inserted;
}
else if (rg.Key.StartsWith("Deleted"))
{
rev.RevisionType = WmlComparerRevisionType.Deleted;
}
XElement revTrackElement = rg.First().RevTrackElement;
rev.RevisionXElement = revTrackElement;
rev.Author = (string)revTrackElement.Attribute(W.author);
rev.ContentXElement = rg.First().ContentElement;
rev.Date = (string)revTrackElement.Attribute(W.date);
rev.PartUri = wDoc.MainDocumentPart.Uri;
rev.PartContentType = wDoc.MainDocumentPart.ContentType;
if (!RevElementsWithNoText.Contains(rev.ContentXElement.Name))
{
rev.Text = rg
.Select(rgc => rgc.ContentElement.Name == W.pPr ? NewLine : rgc.ContentElement.Value)
.StringConcatenate();
}
return(rev);
})
.ToList();
IEnumerable <WmlComparerRevision> footnotesRevisionList =
GetFootnoteEndnoteRevisionList(wDoc.MainDocumentPart.FootnotesPart, W.footnote, settings);
IEnumerable <WmlComparerRevision> endnotesRevisionList =
GetFootnoteEndnoteRevisionList(wDoc.MainDocumentPart.EndnotesPart, W.endnote, settings);
List <WmlComparerRevision> finalRevisionList = mainDocPartRevisionList
.Concat(footnotesRevisionList)
.Concat(endnotesRevisionList)
.ToList();
return(finalRevisionList);
}
}
}
private static IEnumerable <WmlComparerRevision> GetFootnoteEndnoteRevisionList(
OpenXmlPart footnotesEndnotesPart,
XName footnoteEndnoteElementName,
WmlComparerSettings settings)
{
if (footnotesEndnotesPart == null)
{
return(Enumerable.Empty <WmlComparerRevision>());
}
XDocument xDoc = footnotesEndnotesPart.GetXDocument();
IEnumerable <XElement> footnotesEndnotes =
xDoc.Root?.Elements(footnoteEndnoteElementName) ?? throw new OpenXmlPowerToolsException("Invalid document.");
var revisionsForPart = new List <WmlComparerRevision>();
foreach (XElement fn in footnotesEndnotes)
{
ComparisonUnitAtom[] atomList = CreateComparisonUnitAtomList(footnotesEndnotesPart, fn, settings).ToArray();
if (False)
{
var sb = new StringBuilder();
foreach (ComparisonUnitAtom item in atomList)
{
sb.Append(item + Environment.NewLine);
}
string sbs = sb.ToString();
TestUtil.NotePad(sbs);
}
List <IGrouping <string, ComparisonUnitAtom> > grouped = atomList
.GroupAdjacent(a =>
{
string key = a.CorrelationStatus.ToString();
if (a.CorrelationStatus != CorrelationStatus.Equal)
{
var rt = new XElement(a.RevTrackElement.Name,
new XAttribute(XNamespace.Xmlns + "w",
"http://schemas.openxmlformats.org/wordprocessingml/2006/main"),
a.RevTrackElement.Attributes().Where(a2 => a2.Name != W.id && a2.Name != PtOpenXml.Unid));
key += rt.ToString(SaveOptions.DisableFormatting);
}
return(key);
})
.ToList();
List <IGrouping <string, ComparisonUnitAtom> > revisions = grouped
.Where(k => k.Key != "Equal")
.ToList();
IEnumerable <WmlComparerRevision> thisNoteRevisionList = revisions
.Select(rg =>
{
var rev = new WmlComparerRevision();
if (rg.Key.StartsWith("Inserted"))
{
rev.RevisionType = WmlComparerRevisionType.Inserted;
}
else if (rg.Key.StartsWith("Deleted"))
{
rev.RevisionType = WmlComparerRevisionType.Deleted;
}
XElement revTrackElement = rg.First().RevTrackElement;
rev.RevisionXElement = revTrackElement;
rev.Author = (string)revTrackElement.Attribute(W.author);
rev.ContentXElement = rg.First().ContentElement;
rev.Date = (string)revTrackElement.Attribute(W.date);
rev.PartUri = footnotesEndnotesPart.Uri;
rev.PartContentType = footnotesEndnotesPart.ContentType;
if (!RevElementsWithNoText.Contains(rev.ContentXElement.Name))
{
rev.Text = rg
.Select(rgc => rgc.ContentElement.Name == W.pPr ? NewLine : rgc.ContentElement.Value)
.StringConcatenate();
}
return(rev);
});
revisionsForPart.AddRange(thisNoteRevisionList);
}
return(revisionsForPart);
}
// prohibit
// - altChunk
// - subDoc
// - contentPart
// This strips all text nodes from the XML tree, thereby leaving only the structure.
private static object CloneBlockLevelContentForHashing(
OpenXmlPart mainDocumentPart,
XNode node,
bool includeRelatedParts,
WmlComparerSettings settings)
{
if (node is XElement element)
{
if (element.Name == W.bookmarkStart ||
element.Name == W.bookmarkEnd ||
element.Name == W.pPr ||
element.Name == W.rPr)
{
return(null);
}
if (element.Name == W.p)
{
var clonedPara = new XElement(element.Name,
element.Attributes().Where(a => a.Name != W.rsid &&
a.Name != W.rsidDel &&
a.Name != W.rsidP &&
a.Name != W.rsidR &&
a.Name != W.rsidRDefault &&
a.Name != W.rsidRPr &&
a.Name != W.rsidSect &&
a.Name != W.rsidTr &&
a.Name.Namespace != PtOpenXml.pt),
element.Nodes().Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings)));
IEnumerable <IGrouping <bool, XElement> > groupedRuns = clonedPara
.Elements()
.GroupAdjacent(e => e.Name == W.r &&
e.Elements().Count() == 1 &&
e.Element(W.t) != null);
var clonedParaWithGroupedRuns = new XElement(element.Name,
groupedRuns.Select(g =>
{
if (g.Key)
{
string text = g.Select(t => t.Value).StringConcatenate();
if (settings.CaseInsensitive)
{
text = text.ToUpper(settings.CultureInfo);
}
var newRun = (object)new XElement(W.r,
new XElement(W.t,
text));
return(newRun);
}
return(g);
}));
return(clonedParaWithGroupedRuns);
}
if (element.Name == W.r)
{
IEnumerable <XElement> clonedRuns = element
.Elements()
.Where(e => e.Name != W.rPr)
.Select(rc => new XElement(W.r,
CloneBlockLevelContentForHashing(mainDocumentPart, rc, includeRelatedParts, settings)));
return(clonedRuns);
}
if (element.Name == W.tbl)
{
var clonedTable = new XElement(W.tbl,
element.Elements(W.tr).Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings)));
return(clonedTable);
}
if (element.Name == W.tr)
{
var clonedRow = new XElement(W.tr,
element.Elements(W.tc).Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings)));
return(clonedRow);
}
if (element.Name == W.tc)
{
var clonedCell = new XElement(W.tc,
element.Elements().Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings)));
return(clonedCell);
}
if (element.Name == W.tcPr)
{
var clonedCellProps = new XElement(W.tcPr,
element.Elements(W.gridSpan).Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings)));
return(clonedCellProps);
}
if (element.Name == W.gridSpan)
{
var clonedGridSpan = new XElement(W.gridSpan,
new XAttribute("val", (string)element.Attribute(W.val)));
return(clonedGridSpan);
}
if (element.Name == W.txbxContent)
{
var clonedTextbox = new XElement(W.txbxContent,
element.Elements().Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings)));
return(clonedTextbox);
}
if (includeRelatedParts)
{
if (ComparisonUnitWord.ElementsWithRelationshipIds.Contains(element.Name))
{
var newElement = new XElement(element.Name,
element.Attributes()
.Where(a => a.Name.Namespace != PtOpenXml.pt)
.Where(a => !AttributesToTrimWhenCloning.Contains(a.Name))
.Select(a =>
{
if (!ComparisonUnitWord.RelationshipAttributeNames.Contains(a.Name))
{
return(a);
}
var rId = (string)a;
// could be an hyperlink relationship
try
{
OpenXmlPart oxp = mainDocumentPart.GetPartById(rId);
if (oxp == null)
{
throw new FileFormatException("Invalid WordprocessingML Document");
}
var anno = oxp.Annotation <PartSHA1HashAnnotation>();
if (anno != null)
{
return(new XAttribute(a.Name, anno.Hash));
}
if (!oxp.ContentType.EndsWith("xml"))
{
using (Stream str = oxp.GetStream())
{
byte[] ba;
using (var br = new BinaryReader(str))
{
ba = br.ReadBytes((int)str.Length);
}
string sha1 = WmlComparerUtil.SHA1HashStringForByteArray(ba);
oxp.AddAnnotation(new PartSHA1HashAnnotation(sha1));
return(new XAttribute(a.Name, sha1));
}
}
}
catch (ArgumentOutOfRangeException)
{
HyperlinkRelationship hr =
mainDocumentPart.HyperlinkRelationships.FirstOrDefault(z => z.Id == rId);
if (hr != null)
{
string str = hr.Uri.ToString();
return(new XAttribute(a.Name, str));
}
// could be an external relationship
ExternalRelationship er =
mainDocumentPart.ExternalRelationships.FirstOrDefault(z => z.Id == rId);
if (er != null)
{
string str = er.Uri.ToString();
return(new XAttribute(a.Name, str));
}
return(new XAttribute(a.Name, "NULL Relationship"));
}
return(null);
}),
element.Nodes().Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings)));
return(newElement);
}
}
if (element.Name == VML.shape)
{
return(new XElement(element.Name,
element.Attributes()
.Where(a => a.Name.Namespace != PtOpenXml.pt)
.Where(a => a.Name != "style" && a.Name != "id" && a.Name != "type"),
element.Nodes().Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings))));
}
if (element.Name == O.OLEObject)
{
var o = new XElement(element.Name,
element.Attributes()
.Where(a => a.Name.Namespace != PtOpenXml.pt)
.Where(a => a.Name != "ObjectID" && a.Name != R.id),
element.Nodes().Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings)));
return(o);
}
if (element.Name == W._object)
{
var o = new XElement(element.Name,
element.Attributes()
.Where(a => a.Name.Namespace != PtOpenXml.pt),
element.Nodes().Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings)));
return(o);
}
if (element.Name == WP.docPr)
{
return(new XElement(element.Name,
element.Attributes()
.Where(a => a.Name.Namespace != PtOpenXml.pt && a.Name != "id"),
element.Nodes().Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings))));
}
return(new XElement(element.Name,
element.Attributes()
.Where(a => a.Name.Namespace != PtOpenXml.pt)
.Where(a => !AttributesToTrimWhenCloning.Contains(a.Name)),
element.Nodes().Select(n =>
CloneBlockLevelContentForHashing(mainDocumentPart, n, includeRelatedParts, settings))));
}
if (settings.CaseInsensitive)
{
if (node is XText xt)
{
string newText = xt.Value.ToUpper(settings.CultureInfo);
return(new XText(newText));
}
}
return(node);
}
public static WmlDocument Compare(WmlDocument source1, WmlDocument source2, WmlComparerSettings settings)
{
return(CompareInternal(source1, source2, settings, true));
}
private static void SaveCleanedDocuments(WmlDocument source1, WmlDocument producedDocument, WmlComparerSettings settings)
{
if (SaveIntermediateFilesForDebugging && settings.DebugTempFileDi != null)
{
WmlDocument cleanedSource = CleanPowerToolsAndRsid(source1);
SaveDocumentIfDesired(cleanedSource, "Cleaned-Source.docx", settings);
WmlDocument cleanedProduced = CleanPowerToolsAndRsid(producedDocument);
SaveDocumentIfDesired(cleanedProduced, "Cleaned-Produced.docx", settings);
}
}
private static void SaveDocumentsAfterProducingDocument(MemoryStream ms1, MemoryStream ms2, WmlComparerSettings settings)
{
if (SaveIntermediateFilesForDebugging && settings.DebugTempFileDi != null)
{
SaveDocumentIfDesired(new WmlDocument("after1.docx", ms1), "Source1-Step5-AfterProducingDocument.docx", settings);
SaveDocumentIfDesired(new WmlDocument("after2.docx", ms2), "Source2-Step5-AfterProducingDocument.docx", settings);
}
}
private static void CreateComparisonUnitAtomListRecurse(
OpenXmlPart part,
XElement element,
List <ComparisonUnitAtom> comparisonUnitAtomList,
WmlComparerSettings settings)
{
if (element.Name == W.body || element.Name == W.footnote || element.Name == W.endnote)
{
foreach (XElement item in element.Elements())
{
CreateComparisonUnitAtomListRecurse(part, item, comparisonUnitAtomList, settings);
}
return;
}
if (element.Name == W.p)
{
IEnumerable <XElement> paraChildrenToProcess = element
.Elements()
.Where(e => e.Name != W.pPr);
foreach (XElement item in paraChildrenToProcess)
{
CreateComparisonUnitAtomListRecurse(part, item, comparisonUnitAtomList, settings);
}
XElement paraProps = element.Element(W.pPr);
if (paraProps == null)
{
var pPrComparisonUnitAtom = new ComparisonUnitAtom(
new XElement(W.pPr),
element.AncestorsAndSelf()
.TakeWhile(a => a.Name != W.body && a.Name != W.footnotes && a.Name != W.endnotes).Reverse()
.ToArray(),
part,
settings);
comparisonUnitAtomList.Add(pPrComparisonUnitAtom);
}
else
{
var pPrComparisonUnitAtom = new ComparisonUnitAtom(
paraProps,
element.AncestorsAndSelf()
.TakeWhile(a => a.Name != W.body && a.Name != W.footnotes && a.Name != W.endnotes).Reverse()
.ToArray(),
part,
settings);
comparisonUnitAtomList.Add(pPrComparisonUnitAtom);
}
return;
}
if (element.Name == W.r)
{
IEnumerable <XElement> runChildrenToProcess = element
.Elements()
.Where(e => e.Name != W.rPr);
foreach (XElement item in runChildrenToProcess)
{
CreateComparisonUnitAtomListRecurse(part, item, comparisonUnitAtomList, settings);
}
return;
}
if (element.Name == W.t || element.Name == W.delText)
{
string val = element.Value;
foreach (char ch in val)
{
var sr = new ComparisonUnitAtom(
new XElement(element.Name, ch),
element.AncestorsAndSelf()
.TakeWhile(a => a.Name != W.body && a.Name != W.footnotes && a.Name != W.endnotes).Reverse()
.ToArray(),
part,
settings);
comparisonUnitAtomList.Add(sr);
}
return;
}
if (AllowableRunChildren.Contains(element.Name) || element.Name == W._object)
{
var sr3 = new ComparisonUnitAtom(
element,
element.AncestorsAndSelf().TakeWhile(a => a.Name != W.body && a.Name != W.footnotes && a.Name != W.endnotes)
.Reverse().ToArray(),
part,
settings);
comparisonUnitAtomList.Add(sr3);
return;
}
RecursionInfo re = RecursionElements.FirstOrDefault(z => z.ElementName == element.Name);
if (re != null)
{
AnnotateElementWithProps(part, element, comparisonUnitAtomList, re.ChildElementPropertyNames, settings);
return;
}
if (ElementsToThrowAway.Contains(element.Name))
{
return;
}
AnnotateElementWithProps(part, element, comparisonUnitAtomList, null, settings);
}
private static WmlDocument CompareInternal(
WmlDocument source1,
WmlDocument source2,
WmlComparerSettings settings,
bool preProcessMarkupInOriginal)
{
if (preProcessMarkupInOriginal)
{
source1 = PreProcessMarkup(source1, settings.StartingIdForFootnotesEndnotes + 1000);
}
source2 = PreProcessMarkup(source2, settings.StartingIdForFootnotesEndnotes + 2000);
SaveDocumentIfDesired(source1, "Source1-Step1-PreProcess.docx", settings);
SaveDocumentIfDesired(source2, "Source2-Step1-PreProcess.docx", settings);
// at this point, both source1 and source2 have unid on every element. These are the values that will
// enable reassembly of the XML tree. But we need other values.
// In source1:
// - accept tracked revisions
// - determine hash code for every block-level element
// - save as attribute on every element
// - accept tracked revisions and reject tracked revisions leave the unids alone, where possible.
// - after accepting and calculating the hash, then can use the unids to find the right block-level
// element in the unmodified source1, and install the hash
// In source2:
// - reject tracked revisions
// - determine hash code for every block-level element
// - save as an attribute on every element
// - after rejecting and calculating the hash, then can use the unids to find the right block-level element
// in the unmodified source2, and install the hash
// - sometimes after accepting or rejecting tracked revisions, several paragraphs will get coalesced into a
// single paragraph due to paragraph marks being inserted / deleted.
// - in this case, some paragraphs will not get a hash injected onto them.
// - if a paragraph doesn't have a hash, then it will never correspond to another paragraph, and such
// issues will need to be resolved in the normal execution of the LCS algorithm.
// - note that when we do propagate the unid through for the first paragraph.
// Establish correlation between the two.
// Find the longest common sequence of block-level elements where hash codes are the same.
// this sometimes will be every block level element in the document. Or sometimes will be just a fair
// number of them.
// at the start of doing the LCS algorithm, we will match up content, and put them in corresponding unknown
// correlated comparison units. Those paragraphs will only ever be matched to their corresponding paragraph.
// then the algorithm can proceed as usual.
// need to call ChangeFootnoteEndnoteReferencesToUniqueRange before creating the wmlResult document, so that
// the same GUID ids are used for footnote and endnote references in both the 'after' document, and in the
// result document.
WmlDocument source1AfterAccepting = RevisionProcessor.AcceptRevisions(source1);
WmlDocument source2AfterRejecting = RevisionProcessor.RejectRevisions(source2);
SaveDocumentIfDesired(source1AfterAccepting, "Source1-Step2-AfterAccepting.docx", settings);
SaveDocumentIfDesired(source2AfterRejecting, "Source2-Step2-AfterRejecting.docx", settings);
// this creates the correlated hash codes that enable us to match up ranges of paragraphs based on
// accepting in source1, rejecting in source2
source1 = HashBlockLevelContent(source1, source1AfterAccepting, settings);
source2 = HashBlockLevelContent(source2, source2AfterRejecting, settings);
SaveDocumentIfDesired(source1, "Source1-Step3-AfterHashing.docx", settings);
SaveDocumentIfDesired(source2, "Source2-Step3-AfterHashing.docx", settings);
// Accept revisions in before, and after
source1 = RevisionProcessor.AcceptRevisions(source1);
source2 = RevisionProcessor.AcceptRevisions(source2);
SaveDocumentIfDesired(source1, "Source1-Step4-AfterAccepting.docx", settings);
SaveDocumentIfDesired(source2, "Source2-Step4-AfterAccepting.docx", settings);
// after accepting revisions, some unids may have been removed by revision accepter, along with the
// correlatedSHA1Hash codes, this is as it should be.
// but need to go back in and add guids to paragraphs that have had them removed.
using (var ms = new MemoryStream())
{
ms.Write(source2.DocumentByteArray, 0, source2.DocumentByteArray.Length);
using (WordprocessingDocument wDoc = WordprocessingDocument.Open(ms, true))
{
AddUnidsToMarkupInContentParts(wDoc);
}
}
var wmlResult = new WmlDocument(source1);
using (var ms1 = new MemoryStream())
using (var ms2 = new MemoryStream())
{
ms1.Write(source1.DocumentByteArray, 0, source1.DocumentByteArray.Length);
ms2.Write(source2.DocumentByteArray, 0, source2.DocumentByteArray.Length);
WmlDocument producedDocument;
using (WordprocessingDocument wDoc1 = WordprocessingDocument.Open(ms1, true))
using (WordprocessingDocument wDoc2 = WordprocessingDocument.Open(ms2, true))
{
producedDocument = ProduceDocumentWithTrackedRevisions(settings, wmlResult, wDoc1, wDoc2);
}
SaveDocumentsAfterProducingDocument(ms1, ms2, settings);
SaveCleanedDocuments(source1, producedDocument, settings);
return(producedDocument);
}
}