autodesk/DocumentAutomation/src/DocxWorkbench.WordInterop/TemplateDocumentGenerator.cs

using System.IO.Compression;
using System.Security.Cryptography;
using System.Text;
using System.Text.Json;
using System.Xml.Linq;
using DocumentAutomation.Shared;
using DocxWorkbench.Core;

namespace DocxWorkbench.WordInterop;

public sealed class TemplateDocumentGenerator : ITemplateDocumentGenerator, IDisposable
{
    private readonly ITemplatePackageStore _packages;
    private readonly IWordDocumentNormalizer _normalizer;
    private readonly WordDocumentService _word = new();
    private readonly IDocumentInspector _inspector;
    public TemplateDocumentGenerator(ITemplatePackageStore packages, IWordDocumentNormalizer normalizer, IDocumentInspector inspector) { _packages = packages; _normalizer = normalizer; _inspector = inspector; }

    public async Task<TemplateGenerationResult> GenerateAsync(TemplateGenerationRequest request, IProgress<WordOperationProgress>? progress, CancellationToken cancellationToken)
    {
        var opened = _packages.Open(request.TemplatePackagePath); DocumentNormalizationResult? normalizedBody = null; var outputPdf = Path.ChangeExtension(request.OutputDocxPath, ".pdf");
        try
        {
            var bodyPath = request.ContentSourcePath;
            if (Path.GetExtension(bodyPath).Equals(".doc", StringComparison.OrdinalIgnoreCase)) { normalizedBody = await _normalizer.NormalizeAsync(new DocumentNormalizationRequest { SourcePath = bodyPath, WorkingDirectory = Path.Combine(opened.WorkingDirectory, "body") }, progress, cancellationToken); bodyPath = normalizedBody.NormalizedPath; }
            var workingTemplate = Path.Combine(opened.WorkingDirectory, "working-template.docx"); File.Copy(opened.TemplateDocxPath, workingTemplate, true); ReplaceScalarTokens(workingTemplate, request.ScalarValues); progress?.Report(new WordOperationProgress { Stage = "Generate", Message = "Building document from learned template" });
            var result = await _word.BuildDocumentAsync(new WordDocumentBuildRequest { TemplatePath = workingTemplate, BodyPath = bodyPath, OutputDocxPath = request.OutputDocxPath }, progress is null ? null : new Progress<string>(x => progress.Report(new WordOperationProgress { Stage = "Generate", Message = x })), cancellationToken); await _word.ExportPdfAsync(result.OutputDocxPath, outputPdf, cancellationToken);
            var validation = Validate(result.OutputDocxPath, opened.Template.TemplateId); var validationPath = Path.ChangeExtension(request.OutputDocxPath, ".validation.json"); File.WriteAllText(validationPath, JsonSerializer.Serialize(validation, new JsonSerializerOptions { WriteIndented = true })); return new TemplateGenerationResult { OutputDocxPath = result.OutputDocxPath, OutputPdfPath = outputPdf, GenerationLogPath = result.LogPath, ValidationLogPath = validationPath, Validation = validation };
        }
        finally { TryDeleteDirectory(opened.WorkingDirectory); }
    }

    private IReadOnlyList<RuleEvaluationResult> Validate(string output, string templateId)
    {
        var inspection = _inspector.Inspect(output); var result = new List<RuleEvaluationResult> { new() { RuleId = "DOCX_EXISTS", Status = File.Exists(output) ? RuleEvaluationStatus.Pass : RuleEvaluationStatus.Fail, Message = "Output DOCX exists.", DocumentLocation = output }, new() { RuleId = "NO_BODY_MARKER", Status = inspection.InsertionPoint.Type == "None" ? RuleEvaluationStatus.Pass : RuleEvaluationStatus.NeedsReview, Message = inspection.InsertionPoint.Type == "None" ? "No unresolved insertion marker remains." : "An insertion marker remains in the output.", DocumentLocation = "word/document.xml" }, new() { RuleId = "NO_TRACKED_CHANGES", Status = inspection.HasTrackedChanges ? RuleEvaluationStatus.Fail : RuleEvaluationStatus.Pass, Message = inspection.HasTrackedChanges ? "Tracked changes remain." : "No tracked changes detected." }, new() { RuleId = "TEMPLATE_ID", Status = RuleEvaluationStatus.Pass, Message = $"Generated from template {templateId}." } }; return result;
    }

    private static void ReplaceScalarTokens(string path, IReadOnlyDictionary<string, string> values)
    {
        if (values.Count == 0) return; using var archive = ZipFile.Open(path, ZipArchiveMode.Update); var w = XNamespace.Get("http://schemas.openxmlformats.org/wordprocessingml/2006/main");
        foreach (var entry in archive.Entries.Where(x => x.FullName.StartsWith("word/", StringComparison.OrdinalIgnoreCase) && x.FullName.EndsWith(".xml", StringComparison.OrdinalIgnoreCase)).ToArray())
        {
            XDocument document; using (var reader = new StreamReader(entry.Open())) document = XDocument.Parse(reader.ReadToEnd(), LoadOptions.PreserveWhitespace); var changed = false; foreach (var text in document.Descendants(w + "t")) { var value = text.Value; foreach (var pair in values) { var token = "{{" + pair.Key + "}}"; if (!value.Contains(token, StringComparison.OrdinalIgnoreCase)) continue; value = value.Replace(token, pair.Value, StringComparison.OrdinalIgnoreCase); changed = true; } text.Value = value; } if (!changed) continue; entry.Delete(); var replacement = archive.CreateEntry(entry.FullName, CompressionLevel.Optimal); using var stream = replacement.Open(); using var writer = new StreamWriter(stream, new UTF8Encoding(false)); document.Save(writer, SaveOptions.DisableFormatting);
        }
    }

    private static void TryDeleteDirectory(string path) { try { if (Directory.Exists(path)) Directory.Delete(path, true); } catch { } }
    public void Dispose() { _word.Dispose(); if (_normalizer is IDisposable disposable) disposable.Dispose(); }
}