Include pdf in result

src/Tests/PdfNormalizerTests.cs

@@ -0,0 +1,91 @@
+using System.Text;
+using Docnet.Core;
+
+[TestFixture]
+public class PdfNormalizerTests
+{
+    [Test]
+    public void NeutralizesVolatileValues()
+    {
+        var input =
+            "/ID [<A1B2C3D4E5F60718> <1122334455667788>] " +
+            "/CreationDate(D:20240115093000+05'30') " +
+            "/ModDate(D:20240115093000Z) " +
+            "<xmp:CreateDate>2024-01-15T09:30:00+05:30</xmp:CreateDate>" +
+            "<xmp:ModifyDate>2024-01-15T09:30:00Z</xmp:ModifyDate>" +
+            "<xmp:MetadataDate>2024-01-15T09:30:00Z</xmp:MetadataDate>" +
+            "<xmpMM:DocumentID>uuid:0f7b2c9a-1234-5678-9abc-def012345678</xmpMM:DocumentID>" +
+            "<xmpMM:InstanceID>xmp.iid:1a2b3c4d</xmpMM:InstanceID>";
+        var expected =
+            "/ID [<0000000000000000> <0000000000000000>] " +
+            "/CreationDate(D:00000000000000+00'00') " +
+            "/ModDate(D:00000000000000Z) " +
+            "<xmp:CreateDate>0000-00-00T00:00:00+00:00</xmp:CreateDate>" +
+            "<xmp:ModifyDate>0000-00-00T00:00:00Z</xmp:ModifyDate>" +
+            "<xmp:MetadataDate>0000-00-00T00:00:00Z</xmp:MetadataDate>" +
+            $"<xmpMM:DocumentID>{new string('0', 41)}</xmpMM:DocumentID>" +
+            $"<xmpMM:InstanceID>{new string('0', 16)}</xmpMM:InstanceID>";
+        Assert.That(Normalize(input), Is.EqualTo(expected));
+    }
+
+    [Test]
+    public void CollapsesDifferingValuesToTheSameOutput()
+    {
+        // The same producer emits a stable structure across runs, so two documents differing only
+        // in the volatile digits/hex normalize to identical bytes.
+        var a = "/ID [<A1B2C3D4>] /CreationDate(D:20240115093000+05'30')";
+        var b = "/ID [<99887766>] /CreationDate(D:19991231235959+11'45')";
+        Assert.That(a, Is.Not.EqualTo(b));
+        Assert.That(Normalize(a), Is.EqualTo(Normalize(b)));
+    }
+
+    [Test]
+    public void LeavesLookalikeKeysUntouched()
+    {
+        // /IDTree is a name-tree key (not the file identifier), /ModDateStamp is a different name,
+        // and a self-closing date element has no content: none should be altered.
+        var input = "/IDTree [1 2] /ModDateStamp(20240101) <xmp:CreateDate/>2024";
+        Assert.That(Normalize(input), Is.EqualTo(input));
+    }
+
+    [Test]
+    public void NormalizedDocumentStillLoads()
+    {
+        var data = File.ReadAllBytes("sample.pdf");
+        PdfNormalizer.Normalize(data);
+
+        using var reader = DocLib.Instance.GetDocReader(data, new(scalingFactor: 2));
+        Assert.That(reader.GetPageCount(), Is.EqualTo(2));
+    }
+
+    [Test]
+    public void IsIdempotent()
+    {
+        // A second pass has nothing left to change: normalizing already-normalized bytes is a no-op.
+        var once = File.ReadAllBytes("sample.pdf");
+        PdfNormalizer.Normalize(once);
+        var twice = (byte[])once.Clone();
+        PdfNormalizer.Normalize(twice);
+        Assert.That(twice, Is.EqualTo(once));
+    }
+
+    [Test]
+    public void NormalizedSinglePageSplitStillLoads()
+    {
+        // A page subset is re-serialized by pdfium (reintroducing volatile fields) then normalized;
+        // it must remain a valid one-page document.
+        var data = File.ReadAllBytes("sample.pdf");
+        var split = DocLib.Instance.Split(data, 1, 1);
+        PdfNormalizer.Normalize(split);
+
+        using var reader = DocLib.Instance.GetDocReader(split, new(scalingFactor: 2));
+        Assert.That(reader.GetPageCount(), Is.EqualTo(1));
+    }
+
+    static string Normalize(string value)
+    {
+        var bytes = Encoding.Latin1.GetBytes(value);
+        PdfNormalizer.Normalize(bytes);
+        return Encoding.Latin1.GetString(bytes);
+    }
+}

src/Tests/PdfSnapshotTests.cs

@@ -0,0 +1,21 @@
+using System.Runtime.CompilerServices;
+using Docnet.Core;
+
+[TestFixture]
+public class PdfSnapshotTests
+{
+    // The SinglePage snapshots are subset to only the rendered page, so the accepted pdf must load
+    // as a one-page document. A readable guard alongside the opaque binary verified files.
+    [TestCase("Samples.VerifyFirstPage#pdf.verified.pdf", 1)]
+    [TestCase("Samples.VerifySecondPage#pdf.verified.pdf", 1)]
+    [TestCase("Samples.VerifyPdf#pdf.verified.pdf", 2)]
+    public void SnapshotHasExpectedPageCount(string file, int expectedPages)
+    {
+        var data = File.ReadAllBytes(Path.Combine(SourceDirectory(), file));
+        using var reader = DocLib.Instance.GetDocReader(data, new(scalingFactor: 2));
+        Assert.That(reader.GetPageCount(), Is.EqualTo(expectedPages));
+    }
+
+    static string SourceDirectory([CallerFilePath] string path = "") =>
+        Path.GetDirectoryName(path)!;
+}

src/Tests/Samples.VerifyFirstPage.verified.txt

@@ -0,0 +1,19 @@
+{
+  Version: 1.4,
+  PageCount: 2,
+  Pages: [
+    {
+      Text:
+ A Simple PDF File 
+ This is a small demonstration .pdf file - 
+ just for use in the Virtual Mechanics tutorials. More text. And more 
+ text. And more text. And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. And more text. Boring, zzzzz. And more text. And more text. And 
+ more text. And more text. And more text. And more text. And more text. 
+ And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. And more text. And more text. Even more. Continued on page 2 ...
+    }
+  ]
+}

src/Tests/Samples.VerifyPageDimensions.verified.txt

@@ -0,0 +1,29 @@
+{
+  Version: 1.4,
+  PageCount: 2,
+  Pages: [
+    {
+      Text:
+ A Simple PDF File 
+ This is a small demonstration .pdf file - 
+ just for use in the Virtual Mechanics tutorials. More text. And more 
+ text. And more text. And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. And more text. Boring, zzzzz. And more text. And more text. And 
+ more text. And more text. And more text. And more text. And more text. 
+ And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. And more text. And more text. Even more. Continued on page 2 ...
+    },
+    {
+      Index: 1,
+      Text:
+ Simple PDF File 2 
+ ...continued from page 1. Yet more text. And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. Oh, how boring typing this stuff. But not as boring as watching 
+ paint dry. And more text. And more text. And more text. And more text. 
+ Boring. More, a little more text. The end, and just as well. 
+    }
+  ]
+}

src/Tests/Samples.VerifyPdf.verified.txt

@@ -0,0 +1,29 @@
+{
+  Version: 1.4,
+  PageCount: 2,
+  Pages: [
+    {
+      Text:
+ A Simple PDF File 
+ This is a small demonstration .pdf file - 
+ just for use in the Virtual Mechanics tutorials. More text. And more 
+ text. And more text. And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. And more text. Boring, zzzzz. And more text. And more text. And 
+ more text. And more text. And more text. And more text. And more text. 
+ And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. And more text. And more text. Even more. Continued on page 2 ...
+    },
+    {
+      Index: 1,
+      Text:
+ Simple PDF File 2 
+ ...continued from page 1. Yet more text. And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. Oh, how boring typing this stuff. But not as boring as watching 
+ paint dry. And more text. And more text. And more text. And more text. 
+ Boring. More, a little more text. The end, and just as well. 
+    }
+  ]
+}

src/Tests/Samples.VerifyPdfStream.verified.txt

@@ -0,0 +1,29 @@
+{
+  Version: 1.4,
+  PageCount: 2,
+  Pages: [
+    {
+      Text:
+ A Simple PDF File 
+ This is a small demonstration .pdf file - 
+ just for use in the Virtual Mechanics tutorials. More text. And more 
+ text. And more text. And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. And more text. Boring, zzzzz. And more text. And more text. And 
+ more text. And more text. And more text. And more text. And more text. 
+ And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. And more text. And more text. Even more. Continued on page 2 ...
+    },
+    {
+      Index: 1,
+      Text:
+ Simple PDF File 2 
+ ...continued from page 1. Yet more text. And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. Oh, how boring typing this stuff. But not as boring as watching 
+ paint dry. And more text. And more text. And more text. And more text. 
+ Boring. More, a little more text. The end, and just as well. 
+    }
+  ]
+}

src/Tests/Samples.VerifyPreserveTransparency.verified.txt

@@ -0,0 +1,29 @@
+{
+  Version: 1.4,
+  PageCount: 2,
+  Pages: [
+    {
+      Text:
+ A Simple PDF File 
+ This is a small demonstration .pdf file - 
+ just for use in the Virtual Mechanics tutorials. More text. And more 
+ text. And more text. And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. And more text. Boring, zzzzz. And more text. And more text. And 
+ more text. And more text. And more text. And more text. And more text. 
+ And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. And more text. And more text. Even more. Continued on page 2 ...
+    },
+    {
+      Index: 1,
+      Text:
+ Simple PDF File 2 
+ ...continued from page 1. Yet more text. And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. Oh, how boring typing this stuff. But not as boring as watching 
+ paint dry. And more text. And more text. And more text. And more text. 
+ Boring. More, a little more text. The end, and just as well. 
+    }
+  ]
+}

src/Tests/Samples.VerifySecondPage.verified.txt

@@ -0,0 +1,16 @@
+{
+  Version: 1.4,
+  PageCount: 2,
+  Pages: [
+    {
+      Index: 1,
+      Text:
+ Simple PDF File 2 
+ ...continued from page 1. Yet more text. And more text. And more text. 
+ And more text. And more text. And more text. And more text. And more 
+ text. Oh, how boring typing this stuff. But not as boring as watching 
+ paint dry. And more text. And more text. And more text. And more text. 
+ Boring. More, a little more text. The end, and just as well. 
+    }
+  ]
+}

src/Verify.DocNet/PageInfo.cs

@@ -0,0 +1,8 @@
+class PageInfo
+{
+    /// <summary>Zero based index of the page within the source document.</summary>
+    public int Index { get; init; }
+
+    /// <summary>Text extracted from the page, or null when the page has no text.</summary>
+    public string? Text { get; init; }
+}

src/Verify.DocNet/PdfInfo.cs

@@ -0,0 +1,13 @@
+class PdfInfo
+{
+    public string Version { get; init; } = "";
+
+    /// <summary>
+    /// Total number of pages in the source document. This is the full count regardless of any
+    /// <c>PagesToInclude</c>/<c>SinglePage</c> filter; the pages actually rendered are listed in
+    /// <see cref="Pages"/>, each carrying its <see cref="PageInfo.Index"/> within the document.
+    /// </summary>
+    public int PageCount { get; init; }
+
+    public IReadOnlyList<PageInfo> Pages { get; init; } = [];
+}