Merge pull request OpenXmlDev#8 from Codeuctivity/SymbolFontToUnicode

Enabled image handler and text transform handler injection
huan086 · Jan 10, 2021 · 8e943a6 · 8e943a6
2 parents d5879a7 + 37b79a2
commit 8e943a6
Show file tree

Hide file tree

Showing 91 changed files with 901 additions and 1,440 deletions.
diff --git a/.editorconfig b/.editorconfig
@@ -199,3 +199,6 @@ dotnet_naming_style.begins_with_i.required_prefix = I
 dotnet_naming_style.begins_with_i.required_suffix = 
 dotnet_naming_style.begins_with_i.word_separator = 
 dotnet_naming_style.begins_with_i.capitalization = pascal_case
+
+# CS1591: Missing XML comment for publicly visible type or member
+dotnet_diagnostic.CS1591.severity = suggestion
diff --git a/.vscode/launch.json b/.vscode/launch.json
@@ -10,7 +10,7 @@
             "request": "launch",
             "preLaunchTask": "build",
             // If you have changed target frameworks, make sure to update the program path.
-            "program": "${workspaceFolder}/OpenXmlPowerToolsExamples/DocumentAssembler/bin/Debug/netcoreapp3.1/DocumentAssembler.dll",
+            "program": "${workspaceFolder}/OpenXmlPowerToolsExamples/DocumentAssembler/bin/Debug/net5.0/DocumentAssembler.dll",
             "args": [],
             "cwd": "${workspaceFolder}/OpenXmlPowerToolsExamples/DocumentAssembler",
             // For more information about the 'console' field, see https://aka.ms/VSCode-CS-LaunchJson-Console

diff --git a/OpenXmlPowerTools.Tests/ChartUpdaterTests.cs b/OpenXmlPowerTools.Tests/ChartUpdaterTests.cs
@@ -1,5 +1,6 @@
 using DocumentFormat.OpenXml.Packaging;
 using OpenXmlPowerTools;
+using OpenXmlPowerTools.Tests;
 using System;
 using System.IO;
 using Xunit;

diff --git a/OpenXmlPowerTools.Tests/DocumentAssemblerTests.cs b/OpenXmlPowerTools.Tests/DocumentAssemblerTests.cs
@@ -1,6 +1,7 @@
 using DocumentFormat.OpenXml.Packaging;
 using DocumentFormat.OpenXml.Validation;
 using OpenXmlPowerTools;
+using OpenXmlPowerTools.Tests;
 using System;
 using System.Collections.Generic;
 using System.IO;

diff --git a/OpenXmlPowerTools.Tests/DocumentBuilderTests.cs b/OpenXmlPowerTools.Tests/DocumentBuilderTests.cs
@@ -1,6 +1,7 @@
 using DocumentFormat.OpenXml.Packaging;
 using DocumentFormat.OpenXml.Validation;
 using OpenXmlPowerTools;
+using OpenXmlPowerTools.Tests;
 using System;
 using System.Collections.Generic;
 using System.IO;

diff --git a/OpenXmlPowerTools.Tests/FormattingAssemblerTests.cs b/OpenXmlPowerTools.Tests/FormattingAssemblerTests.cs
@@ -1,6 +1,7 @@
 using DocumentFormat.OpenXml.Packaging;
 using DocumentFormat.OpenXml.Validation;
 using OpenXmlPowerTools;
+using OpenXmlPowerTools.Tests;
 using System;
 using System.Collections.Generic;
 using System.IO;

diff --git a/OpenXmlPowerTools.Tests/HtmlConverterTests.cs b/OpenXmlPowerTools.Tests/HtmlConverterTests.cs
diff --git a/OpenXmlPowerTools.Tests/HtmlToWmlConverterTests.cs b/OpenXmlPowerTools.Tests/HtmlToWmlConverterTests.cs
@@ -1,6 +1,7 @@
 using DocumentFormat.OpenXml.Packaging;
 using DocumentFormat.OpenXml.Validation;
 using OpenXmlPowerTools;
+using OpenXmlPowerTools.Tests;
 using System;
 using System.IO;
 using System.Linq;
@@ -360,10 +361,7 @@ public void HW001(string name)
 
             var doc = HtmlToWmlConverter.ConvertHtmlToWml(defaultCss, usedAuthorCss, userCss, html, settings, null, s_ProduceAnnotatedHtml ? annotatedHtmlFi.FullName : null);
             Assert.NotNull(doc);
-            if (doc != null)
-            {
-                SaveValidateAndFormatMainDocPart(destDocxFi, doc);
-            }
+            SaveValidateAndFormatMainDocPart(destDocxFi, doc);
         }
 
         [Theory]
@@ -397,6 +395,11 @@ private static void SaveValidateAndFormatMainDocPart(FileInfo destDocxFi, WmlDoc
         {
             WmlDocument formattedDoc;
 
+            if (File.Exists(destDocxFi.FullName))
+            {
+                File.Delete(destDocxFi.FullName);
+            }
+
             doc.SaveAs(destDocxFi.FullName);
             using (var ms = new MemoryStream())
             {
@@ -407,12 +410,9 @@ private static void SaveValidateAndFormatMainDocPart(FileInfo destDocxFi, WmlDoc
                     document.MainDocumentPart.PutXDocumentWithFormatting();
                     var validator = new OpenXmlValidator();
                     var errors = validator.Validate(document);
-                    var errorsString = errors
-                        .Select(e => e.Description + Environment.NewLine)
-                        .StringConcatenate();
+                    var errorsString = errors.Select(e => e.Description + Environment.NewLine).StringConcatenate();
 
-                    // Assert that there were no errors in the generated document.
-                    Assert.Equal("", errorsString);
+                    Assert.True(errorsString.Length == 0, $"Error in {destDocxFi.FullName}\n{errorsString}");
                 }
                 formattedDoc = new WmlDocument(destDocxFi.FullName, ms.ToArray());
             }

diff --git a/OpenXmlPowerTools.Tests/OpenXMLWordprocessingMLToHtmlConverter/WmlToHtmlConverterTests.cs b/OpenXmlPowerTools.Tests/OpenXMLWordprocessingMLToHtmlConverter/WmlToHtmlConverterTests.cs
@@ -0,0 +1,172 @@
+using DocumentFormat.OpenXml.Packaging;
+using OpenXmlPowerTools;
+using OpenXmlPowerTools.OpenXMLWordprocessingMLToHtmlConverter;
+using OpenXmlPowerTools.Tests;
+using System.IO;
+using System.Linq;
+using System.Text;
+using System.Xml.Linq;
+using Xunit;
+
+namespace OxPt
+{
+    public class WmlToHtmlConverterTests
+    {
+        // PowerShell oneliner that generates InlineData for all files in a directory
+        // dir | % { '[InlineData("' + $_.Name + '")]' } | clip
+
+        [Theory]
+        [InlineData("HC001-5DayTourPlanTemplate.docx")]
+        [InlineData("HC002-Hebrew-01.docx")]
+        [InlineData("HC003-Hebrew-02.docx")]
+        [InlineData("HC004-ResumeTemplate.docx")]
+        [InlineData("HC005-TaskPlanTemplate.docx")]
+        [InlineData("HC006-Test-01.docx")]
+        [InlineData("HC007-Test-02.docx")]
+        [InlineData("HC008-Test-03.docx")]
+        [InlineData("HC009-Test-04.docx")]
+        [InlineData("HC010-Test-05.docx")]
+        [InlineData("HC011-Test-06.docx")]
+        [InlineData("HC012-Test-07.docx")]
+        [InlineData("HC013-Test-08.docx")]
+        [InlineData("HC014-RTL-Table-01.docx")]
+        [InlineData("HC015-Vertical-Spacing-atLeast.docx")]
+        [InlineData("HC016-Horizontal-Spacing-firstLine.docx")]
+        [InlineData("HC017-Vertical-Alignment-Cell-01.docx")]
+        [InlineData("HC018-Vertical-Alignment-Para-01.docx")]
+        [InlineData("HC019-Hidden-Run.docx")]
+        [InlineData("HC020-Small-Caps.docx")]
+        [InlineData("HC021-Symbols.docx")]
+        [InlineData("HC022-Table-Of-Contents.docx")]
+        [InlineData("HC023-Hyperlink.docx")]
+        [InlineData("HC024-Tabs-01.docx")]
+        [InlineData("HC025-Tabs-02.docx")]
+        [InlineData("HC026-Tabs-03.docx")]
+        [InlineData("HC027-Tabs-04.docx")]
+        [InlineData("HC028-No-Break-Hyphen.docx")]
+        [InlineData("HC029-Table-Merged-Cells.docx")]
+        [InlineData("HC030-Content-Controls.docx")]
+        [InlineData("HC031-Complicated-Document.docx")]
+        [InlineData("HC032-Named-Color.docx")]
+        [InlineData("HC033-Run-With-Border.docx")]
+        [InlineData("HC034-Run-With-Position.docx")]
+        [InlineData("HC035-Strike-Through.docx")]
+        [InlineData("HC036-Super-Script.docx")]
+        [InlineData("HC037-Sub-Script.docx")]
+        [InlineData("HC038-Conflicting-Border-Weight.docx")]
+        [InlineData("HC039-Bold.docx")]
+        [InlineData("HC040-Hyperlink-Fieldcode-01.docx")]
+        [InlineData("HC041-Hyperlink-Fieldcode-02.docx")]
+        [InlineData("HC042-Image-Png.docx")]
+        [InlineData("HC043-Chart.docx")]
+        [InlineData("HC044-Embedded-Workbook.docx")]
+        [InlineData("HC045-Italic.docx")]
+        [InlineData("HC046-BoldAndItalic.docx")]
+        [InlineData("HC047-No-Section.docx")]
+        [InlineData("HC048-Excerpt.docx")]
+        [InlineData("HC049-Borders.docx")]
+        [InlineData("HC050-Shaded-Text-01.docx")]
+        [InlineData("HC051-Shaded-Text-02.docx")]
+        [InlineData("HC060-Image-with-Hyperlink.docx")]
+        [InlineData("HC061-Hyperlink-in-Field.docx")]
+        public void HC001(string name)
+        {
+            var sourceDir = new DirectoryInfo("../../../../TestFiles/");
+            var sourceDocx = new FileInfo(Path.Combine(sourceDir.FullName, name));
+
+            var oxPtConvertedDestHtml = new FileInfo(Path.Combine(TestUtil.TempDir.FullName, sourceDocx.Name.Replace(".docx", "-3-OxPt.html")));
+            ConvertToHtml(sourceDocx, oxPtConvertedDestHtml, false);
+        }
+
+        [Theory]
+        [InlineData("HC006-Test-01.docx")]
+        public void HC002_NoCssClasses(string name)
+        {
+            var sourceDir = new DirectoryInfo("../../../../TestFiles/");
+            var sourceDocx = new FileInfo(Path.Combine(sourceDir.FullName, name));
+
+            var oxPtConvertedDestHtml = new FileInfo(Path.Combine(TestUtil.TempDir.FullName, sourceDocx.Name.Replace(".docx", "-5-OxPt-No-CSS-Classes.html")));
+            ConvertToHtml(sourceDocx, oxPtConvertedDestHtml, true);
+        }
+
+        private static void CopyFormattingAssembledDocx(FileInfo source, FileInfo dest)
+        {
+            var ba = File.ReadAllBytes(source.FullName);
+            using var ms = new MemoryStream();
+            ms.Write(ba, 0, ba.Length);
+            using (var wordDoc = WordprocessingDocument.Open(ms, true))
+            {
+                RevisionAccepter.AcceptRevisions(wordDoc);
+                var simplifyMarkupSettings = new SimplifyMarkupSettings
+                {
+                    RemoveComments = true,
+                    RemoveContentControls = true,
+                    RemoveEndAndFootNotes = true,
+                    RemoveFieldCodes = false,
+                    RemoveLastRenderedPageBreak = true,
+
+                    RemovePermissions = true,
+                    RemoveProof = true,
+                    RemoveRsidInfo = true,
+                    RemoveSmartTags = true,
+                    RemoveSoftHyphens = true,
+                    RemoveGoBackBookmark = true,
+                    ReplaceTabsWithSpaces = false,
+                };
+                MarkupSimplifier.SimplifyMarkup(wordDoc, simplifyMarkupSettings);
+
+                var formattingAssemblerSettings = new FormattingAssemblerSettings
+                {
+                    RemoveStyleNamesFromParagraphAndRunProperties = false,
+                    ClearStyles = false,
+                    RestrictToSupportedLanguages = false,
+                    RestrictToSupportedNumberingFormats = false,
+                    CreateHtmlConverterAnnotationAttributes = true,
+                    OrderElementsPerStandard = false,
+                    ListItemRetrieverSettings =
+                        new ListItemRetrieverSettings()
+                        {
+                            ListItemTextImplementations = ListItemRetrieverSettings.DefaultListItemTextImplementations,
+                        },
+                };
+
+                FormattingAssembler.AssembleFormatting(wordDoc, formattingAssemblerSettings);
+            }
+            var newBa = ms.ToArray();
+            File.WriteAllBytes(dest.FullName, newBa);
+        }
+
+        private static void ConvertToHtml(FileInfo sourceDocx, FileInfo destFileName, bool fabricateCssClasses)
+        {
+            var byteArray = File.ReadAllBytes(sourceDocx.FullName);
+            using var memoryStream = new MemoryStream();
+            memoryStream.Write(byteArray, 0, byteArray.Length);
+            using var wDoc = WordprocessingDocument.Open(memoryStream, true);
+            var outputDirectory = destFileName.Directory;
+            destFileName = new FileInfo(Path.Combine(outputDirectory.FullName, destFileName.Name));
+            var imageDirectoryName = destFileName.FullName.Substring(0, destFileName.FullName.Length - 5) + "_files";
+            var pageTitle = (string)wDoc.CoreFilePropertiesPart.GetXDocument().Descendants(DC.title).FirstOrDefault();
+            if (pageTitle == null)
+            {
+                pageTitle = sourceDocx.FullName;
+            }
+
+            var settings = new WmlToHtmlConverterSettings()
+            {
+                PageTitle = pageTitle,
+                FabricateCssClasses = fabricateCssClasses,
+                CssClassPrefix = fabricateCssClasses ? "pt-" : null,
+                RestrictToSupportedLanguages = false,
+                RestrictToSupportedNumberingFormats = false
+            };
+
+            var html = WmlToHtmlConverter.ConvertToHtml(wDoc, settings);
+
+            // Note: the xhtml returned by ConvertToHtmlTransform contains objects of type XEntity.  PtOpenXmlUtil.cs define the XEntity class. See http://blogs.msdn.com/ericwhite/archive/2010/01/21/writing-entity-references-using-linq-to-xml.aspx for detailed explanation.
+            // If you further transform the XML tree returned by ConvertToHtmlTransform, you must do it correctly, or entities will not be serialized properly.
+
+            var htmlString = html.ToString(SaveOptions.DisableFormatting);
+            File.WriteAllText(destFileName.FullName, htmlString, Encoding.UTF8);
+        }
+    }
+}
diff --git a/OpenXmlPowerTools.Tests/OpenXmlPowerTools.Tests.csproj b/OpenXmlPowerTools.Tests/OpenXmlPowerTools.Tests.csproj
@@ -1,7 +1,7 @@
 <Project Sdk="Microsoft.NET.Sdk">
 
   <PropertyGroup>
-    <TargetFrameworks Condition="'$(OS)' == 'Windows_NT' ">net472;net5.0</TargetFrameworks>
+    <TargetFrameworks Condition="'$(OS)' == 'Windows_NT' ">net472;net5.0;netcoreapp3.1</TargetFrameworks>
     <TargetFrameworks Condition="'$(OS)' != 'Windows_NT' ">net5.0</TargetFrameworks>
     <LangVersion>8.0</LangVersion>
     <EnableNETAnalyzers>true</EnableNETAnalyzers>

diff --git a/OpenXmlPowerTools.Tests/PresentationBuilderTests.cs b/OpenXmlPowerTools.Tests/PresentationBuilderTests.cs
@@ -1,5 +1,6 @@
 using DocumentFormat.OpenXml.Packaging;
 using OpenXmlPowerTools;
+using OpenXmlPowerTools.Tests;
 using System.Collections.Generic;
 using System.IO;
 using System.Linq;

diff --git a/OpenXmlPowerTools.Tests/RevisionAccepterTests.cs b/OpenXmlPowerTools.Tests/RevisionAccepterTests.cs
@@ -1,4 +1,5 @@
 using OpenXmlPowerTools;
+using OpenXmlPowerTools.Tests;
 using System.IO;
 using Xunit;
 

diff --git a/OpenXmlPowerTools.Tests/RevisionProcessorTests.cs b/OpenXmlPowerTools.Tests/RevisionProcessorTests.cs
@@ -1,4 +1,5 @@
 using OpenXmlPowerTools;
+using OpenXmlPowerTools.Tests;
 using System;
 using System.IO;
 using System.Linq;

diff --git a/OpenXmlPowerTools.Tests/SmlCellFormatterTests.cs b/OpenXmlPowerTools.Tests/SmlCellFormatterTests.cs
@@ -1,5 +1,6 @@
 using DocumentFormat.OpenXml.Packaging;
 using OpenXmlPowerTools;
+using OpenXmlPowerTools.Tests;
 using System.IO;
 using System.Linq;
 using Xunit;
@@ -119,7 +120,7 @@ public void CF002(string name, string sheetName, string range, string expected,
                 File.Copy(sourceXlsx.FullName, sourceCopiedToDestXlsx.FullName);
             }
 
-            using var sDoc = SpreadsheetDocument.Open(sourceXlsx.FullName, true);
+            using var sDoc = SpreadsheetDocument.Open(sourceCopiedToDestXlsx.FullName, true);
             var rangeXml = SmlDataRetriever.RetrieveRange(sDoc, sheetName, range);
             var displayValue = (string)rangeXml.Descendants("DisplayValue").FirstOrDefault();
             var displayColor = (string)rangeXml.Descendants("DisplayColor").FirstOrDefault();

diff --git a/OpenXmlPowerTools.Tests/SmlToHtmlConverterTests.cs b/OpenXmlPowerTools.Tests/SmlToHtmlConverterTests.cs
@@ -1,5 +1,6 @@
 using DocumentFormat.OpenXml.Packaging;
 using OpenXmlPowerTools;
+using OpenXmlPowerTools.Tests;
 using System.IO;
 using Xunit;
 
@@ -53,8 +54,8 @@ public void SH005_ConvertSheet(string name, string sheetName)
             }
 
             var dataTemplateFileNameSuffix = "-2-Generated-XmlData-Entire-Sheet.xml";
-            var dataXmlFi = new FileInfo(Path.Combine(TestUtil.TempDir.FullName, sourceXlsx.Name.Replace(".xlsx", dataTemplateFileNameSuffix)));
-            using var sDoc = SpreadsheetDocument.Open(sourceXlsx.FullName, true);
+            var dataXmlFi = new FileInfo(Path.Combine(TestUtil.TempDir.FullName, sourceCopiedToDestXlsx.Name.Replace(".xlsx", dataTemplateFileNameSuffix)));
+            using var sDoc = SpreadsheetDocument.Open(sourceCopiedToDestXlsx.FullName, true);
             var settings = new SmlToHtmlConverterSettings();
             var rangeXml = SmlDataRetriever.RetrieveSheet(sDoc, sheetName);
             rangeXml.Save(dataXmlFi.FullName);
@@ -111,8 +112,8 @@ public void SH004_ConvertRange(string name, string sheetName, string range)
             }
 
             var dataTemplateFileNameSuffix = string.Format("-2-Generated-XmlData-{0}.xml", range.Replace(":", ""));
-            var dataXmlFi = new FileInfo(Path.Combine(TestUtil.TempDir.FullName, sourceXlsx.Name.Replace(".xlsx", dataTemplateFileNameSuffix)));
-            using var sDoc = SpreadsheetDocument.Open(sourceXlsx.FullName, true);
+            var dataXmlFi = new FileInfo(Path.Combine(TestUtil.TempDir.FullName, sourceCopiedToDestXlsx.Name.Replace(".xlsx", dataTemplateFileNameSuffix)));
+            using var sDoc = SpreadsheetDocument.Open(sourceCopiedToDestXlsx.FullName, true);
             var settings = new SmlToHtmlConverterSettings();
             var rangeXml = SmlDataRetriever.RetrieveRange(sDoc, sheetName, range);
             rangeXml.Save(dataXmlFi.FullName);
@@ -139,8 +140,8 @@ public void SH003_ConvertTable(string name, string tableName)
                 File.Copy(sourceXlsx.FullName, sourceCopiedToDestXlsx.FullName);
             }
 
-            var dataXmlFi = new FileInfo(Path.Combine(TestUtil.TempDir.FullName, sourceXlsx.Name.Replace(".xlsx", "-2-Generated-XmlData.xml")));
-            using var sDoc = SpreadsheetDocument.Open(sourceXlsx.FullName, true);
+            var dataXmlFi = new FileInfo(Path.Combine(TestUtil.TempDir.FullName, sourceCopiedToDestXlsx.Name.Replace(".xlsx", "-2-Generated-XmlData.xml")));
+            using var sDoc = SpreadsheetDocument.Open(sourceCopiedToDestXlsx.FullName, true);
             var settings = new SmlToHtmlConverterSettings();
             var rangeXml = SmlDataRetriever.RetrieveTable(sDoc, tableName);
             rangeXml.Save(dataXmlFi.FullName);
@@ -152,7 +153,7 @@ public void SH002_SheetNames(string name, int numberOfSheets)
         {
             var sourceDir = new DirectoryInfo("../../../../TestFiles/");
             var sourceXlsx = new FileInfo(Path.Combine(sourceDir.FullName, name));
-            using var sDoc = SpreadsheetDocument.Open(sourceXlsx.FullName, true);
+            using var sDoc = SpreadsheetDocument.Open(sourceXlsx.FullName, false);
             var sheetNames = SmlDataRetriever.SheetNames(sDoc);
             Assert.Equal(numberOfSheets, sheetNames.Length);
         }
@@ -164,7 +165,7 @@ public void SH001_TableNames(string name, int numberOfTables)
         {
             var sourceDir = new DirectoryInfo("../../../../TestFiles/");
             var sourceXlsx = new FileInfo(Path.Combine(sourceDir.FullName, name));
-            using var sDoc = SpreadsheetDocument.Open(sourceXlsx.FullName, true);
+            using var sDoc = SpreadsheetDocument.Open(sourceXlsx.FullName, false);
             var table = SmlDataRetriever.TableNames(sDoc);
             Assert.Equal(numberOfTables, table.Length);
         }