From 84c74e26ce1ca080631742dc6668ba916c934ab4 Mon Sep 17 00:00:00 2001 From: TheIrvin Date: Fri, 2 Oct 2026 16:53:05 -0500 Subject: [PATCH] Complete local Markdown conversion flow --- .github/workflows/ci.yml | 19 +++ .gitignore | 2 + MarkItDownWeb.slnx | 3 + README.md | 90 +++++++++++ azure/deploy.sh | 44 ----- .../MarkItDownWeb.Infrastructure.csproj | 4 +- .../Services/MarkItDownService.cs | 17 +- src/Web/Controllers/ConversionController.cs | 28 +++- src/Web/MarkItDownWeb.Web.csproj | 2 +- src/Web/Views/Conversion/Index.cshtml | 65 +++++--- src/Web/appsettings.json | 2 +- src/Web/wwwroot/css/site.css | 90 ++++++++++- src/Web/wwwroot/js/conversion.js | 151 +++++++++++++----- .../FileSignatureValidatorTests.cs | 68 ++++++++ .../MarkItDownWeb.Tests.csproj | 22 +++ 15 files changed, 489 insertions(+), 118 deletions(-) create mode 100644 .github/workflows/ci.yml create mode 100644 README.md delete mode 100644 azure/deploy.sh create mode 100644 tests/MarkItDownWeb.Tests/FileSignatureValidatorTests.cs create mode 100644 tests/MarkItDownWeb.Tests/MarkItDownWeb.Tests.csproj diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml new file mode 100644 index 0000000..4b53fa3 --- /dev/null +++ b/.github/workflows/ci.yml @@ -0,0 +1,19 @@ +name: CI + +on: + push: + branches: [master] + pull_request: + +permissions: + contents: read + +jobs: + test: + runs-on: ubuntu-latest + steps: + - uses: actions/checkout@v7 + - uses: actions/setup-dotnet@v6 + with: + dotnet-version: 10.0.x + - run: dotnet test MarkItDownWeb.slnx --configuration Release diff --git a/.gitignore b/.gitignore index f14a154..f3970af 100644 --- a/.gitignore +++ b/.gitignore @@ -5,4 +5,6 @@ Storage/ .vs/ *.user *.db +*.db-* publish/ +.venv/ diff --git a/MarkItDownWeb.slnx b/MarkItDownWeb.slnx index d0e6c14..9a37ff9 100644 --- a/MarkItDownWeb.slnx +++ b/MarkItDownWeb.slnx @@ -12,4 +12,7 @@ + + + diff --git a/README.md b/README.md new file mode 100644 index 0000000..5547eef --- /dev/null +++ b/README.md @@ -0,0 +1,90 @@ +# FileToMarkdown + +Aplicación web local para convertir documentos a Markdown. Selecciona un archivo, revisa el resultado en el navegador y descarga un `.md` con el nombre del documento original. + +## Funciones + +- Interfaz web responsive con vista previa, copia y descarga del Markdown. +- Historial local de conversiones con búsqueda y reapertura de resultados. +- Conversión de PDF, DOCX, PPTX, XLSX, HTML y CSV mediante [Microsoft MarkItDown](https://github.com/microsoft/markitdown). +- Base de datos SQLite y archivos de resultado guardados en el equipo. +- Validación de extensión, tamaño máximo de 100 MiB y firma de archivos PDF/Office. + +La aplicación se ejecuta en tu equipo. Los archivos se envían al proceso web local y a MarkItDown; esta configuración no proporciona autenticación, así que mantenla accesible solo desde tu propio equipo y no la expongas a una red pública. + +## Requisitos + +- [.NET SDK 10](https://dotnet.microsoft.com/download/dotnet/10.0) +- Python y `pip` + +## Ejecutar en local + +Desde la raíz del repositorio: + +```bash +python -m venv .venv +``` + +Activa el entorno virtual y luego instala MarkItDown: + +```bash +# Windows PowerShell +.venv\Scripts\Activate.ps1 +python -m pip install --upgrade pip +python -m pip install 'markitdown[all]' +``` + +```bash +# Linux / macOS +source .venv/bin/activate +python -m pip install --upgrade pip +python -m pip install 'markitdown[all]' +``` + +Inicia la interfaz: + +```bash +dotnet run --project src/Web/MarkItDownWeb.Web.csproj --launch-profile http +``` + +Abre [http://localhost:5241](http://localhost:5241). La ruta del ejecutable de Python se puede configurar con `MarkItDown__PythonPath`; por defecto se usa `python` del entorno activo. Por ejemplo, en PowerShell: + +```powershell +$env:MarkItDown__PythonPath = (Get-Command python).Source +dotnet run --project src/Web/MarkItDownWeb.Web.csproj --launch-profile http +``` + +## Comprobar cambios + +```bash +dotnet test MarkItDownWeb.slnx --configuration Release +``` + +GitHub Actions ejecuta la misma comprobación al abrir o actualizar un pull request y al hacer push a `master`. + +## Uso + +1. Elige un PDF, DOCX, PPTX, XLSX, HTML o CSV de hasta 100 MiB. +2. Pulsa **Convertir a Markdown** y revisa el contenido generado. +3. Cópialo o descárgalo como `.md`. +4. Abre una conversión anterior desde el historial. Los datos permanecen en el equipo. + +## Datos locales + +La base de datos SQLite se crea automáticamente. Los archivos temporales se eliminan al terminar la conversión; el historial conserva los resultados Markdown y los metadatos. Las carpetas se pueden cambiar desde `src/Web/appsettings.json` (`Storage:UploadsFolder`, `Storage:MarkdownFolder` y `ConnectionStrings:DefaultConnection`). No se conserva el documento original después de convertirlo. + +## Estructura + +```text +src/ + Domain/ Entidades y tipos de conversión + Application/ Flujo de conversión e interfaces + Infrastructure/ MarkItDown, almacenamiento y SQLite + Web/ Interfaz ASP.NET Core MVC +``` + +## Limitaciones conocidas + +- La calidad del Markdown depende del formato y del contenido del documento original. +- MarkItDown puede necesitar dependencias adicionales para ciertos formatos; consulta su documentación si la instalación completa no está disponible en tu plataforma. +- Esta versión está pensada para uso local individual; no incluye cuentas, aislamiento entre usuarios ni almacenamiento en la nube. diff --git a/azure/deploy.sh b/azure/deploy.sh deleted file mode 100644 index 65e47e7..0000000 --- a/azure/deploy.sh +++ /dev/null @@ -1,44 +0,0 @@ -#!/usr/bin/env bash -set -euo pipefail - -# === Azure Deployment Script for MarkItDownWeb === -# Adjust variables below before running. - -RESOURCE_GROUP="rg-markitdownweb" -LOCATION="eastus" -APP_SERVICE_PLAN="plan-markitdownweb" -WEB_APP_NAME="markitdownweb-app" -SQL_SERVER="sql-markitdownweb" -SQL_DB="MarkItDownWeb" -SQL_ADMIN="sqladmin" -SQL_PASSWORD="YourStrongPassword123!" - -az login - -# Resource group -az group create --name "$RESOURCE_GROUP" --location "$LOCATION" - -# App Service Plan + Web App -az appservice plan create --name "$APP_SERVICE_PLAN" --resource-group "$RESOURCE_GROUP" --sku B1 --is-linux -az webapp create --resource-group "$RESOURCE_GROUP" --plan "$APP_SERVICE_PLAN" \ - --name "$WEB_APP_NAME" --runtime "DOTNETCORE:9.0" - -# SQL Server + Database -az sql server create --name "$SQL_SERVER" --resource-group "$RESOURCE_GROUP" \ - --admin-user "$SQL_ADMIN" --admin-password "$SQL_PASSWORD" -az sql db create --resource-group "$RESOURCE_GROUP" --server "$SQL_SERVER" \ - --name "$SQL_DB" --service-objective S0 - -# Connection string -CS="Server=tcp:$SQL_SERVER.database.windows.net,1433;Database=$SQL_DB;User ID=$SQL_ADMIN;Password=$SQL_PASSWORD;Trusted_Connection=False;Encrypt=True;" -az webapp config connection-string set --resource-group "$RESOURCE_GROUP" \ - --name "$WEB_APP_NAME" --settings DefaultConnection="$CS" --connection-string-type SQLAzure - -# App settings -az webapp config appsettings set --resource-group "$RESOURCE_GROUP" --name "$WEB_APP_NAME" \ - --settings MarkItDown__PythonPath="/usr/bin/python3" Storage__UploadsFolder="/home/site/wwwroot/Uploads" - -# Publish and deploy -dotnet publish src/Web -c Release -o ./publish -cd publish && zip -r ../publish.zip . && cd .. -az webapp deploy --resource-group "$RESOURCE_GROUP" --name "$WEB_APP_NAME" --src-path publish.zip --type zip diff --git a/src/Infrastructure/MarkItDownWeb.Infrastructure.csproj b/src/Infrastructure/MarkItDownWeb.Infrastructure.csproj index 52d741c..b122e7f 100644 --- a/src/Infrastructure/MarkItDownWeb.Infrastructure.csproj +++ b/src/Infrastructure/MarkItDownWeb.Infrastructure.csproj @@ -6,11 +6,11 @@ - + runtime; build; native; contentfiles; analyzers; buildtransitive all - + diff --git a/src/Infrastructure/Services/MarkItDownService.cs b/src/Infrastructure/Services/MarkItDownService.cs index 2a58b2d..393ca1b 100644 --- a/src/Infrastructure/Services/MarkItDownService.cs +++ b/src/Infrastructure/Services/MarkItDownService.cs @@ -1,6 +1,7 @@ using System.Diagnostics; using System.Text; using MarkItDownWeb.Application.Interfaces; +using MarkItDownWeb.Infrastructure.Security; using Microsoft.Extensions.Configuration; namespace MarkItDownWeb.Infrastructure.Services; @@ -19,6 +20,10 @@ public async Task ConvertFileAsync(string absoluteFilePath, Cancellation if (!File.Exists(absoluteFilePath)) throw new FileNotFoundException("File not found", absoluteFilePath); + var extension = Path.GetExtension(absoluteFilePath); + if (!FileSignatureValidator.IsValid(absoluteFilePath, extension)) + throw new InvalidOperationException("The file contents do not match its extension."); + var psi = new ProcessStartInfo { FileName = _pythonPath, @@ -41,7 +46,17 @@ public async Task ConvertFileAsync(string absoluteFilePath, Cancellation process.BeginOutputReadLine(); process.BeginErrorReadLine(); - await process.WaitForExitAsync(ct); + try + { + await process.WaitForExitAsync(ct); + } + catch (OperationCanceledException) + { + if (!process.HasExited) + process.Kill(entireProcessTree: true); + + throw; + } if (process.ExitCode != 0) throw new InvalidOperationException($"MarkItDown failed: {stderr}"); diff --git a/src/Web/Controllers/ConversionController.cs b/src/Web/Controllers/ConversionController.cs index a9b7176..a25a66f 100644 --- a/src/Web/Controllers/ConversionController.cs +++ b/src/Web/Controllers/ConversionController.cs @@ -15,16 +15,23 @@ public ConversionController(ConversionService conversionService) public IActionResult Index() => View(new ConversionViewModel()); [HttpPost] - [RequestSizeLimit(100_000_000)] - public async Task Convert(IFormFile file, CancellationToken ct) + [RequestSizeLimit(105_906_176)] + public async Task Convert(IFormFile? file, CancellationToken ct) { if (file is null || file.Length == 0) return BadRequest(new { error = "No file was received." }); - await using var stream = file.OpenReadStream(); - var result = await _conversionService.ConvertAsync(stream, file.FileName, file.Length, ct); + try + { + await using var stream = file.OpenReadStream(); + var result = await _conversionService.ConvertAsync(stream, file.FileName, file.Length, ct); - return Json(result); + return Json(result); + } + catch (InvalidOperationException ex) + { + return BadRequest(new { error = ex.Message }); + } } [HttpGet] @@ -37,8 +44,15 @@ public async Task History(string? search, DateTime? date, Cancell [HttpGet] public async Task Item(Guid id, CancellationToken ct) { - var result = await _conversionService.GetHistoryItemAsync(id, ct); - return Json(result); + try + { + var result = await _conversionService.GetHistoryItemAsync(id, ct); + return Json(result); + } + catch (KeyNotFoundException) + { + return NotFound(new { error = "Conversion not found." }); + } } [HttpPost] diff --git a/src/Web/MarkItDownWeb.Web.csproj b/src/Web/MarkItDownWeb.Web.csproj index bd94e24..88a8df8 100644 --- a/src/Web/MarkItDownWeb.Web.csproj +++ b/src/Web/MarkItDownWeb.Web.csproj @@ -6,7 +6,7 @@ - + runtime; build; native; contentfiles; analyzers; buildtransitive all diff --git a/src/Web/Views/Conversion/Index.cshtml b/src/Web/Views/Conversion/Index.cshtml index 2e95f01..54e1731 100644 --- a/src/Web/Views/Conversion/Index.cshtml +++ b/src/Web/Views/Conversion/Index.cshtml @@ -3,39 +3,58 @@ ViewData["Title"] = "Convertir archivos a Markdown"; } -
-
-
-

Convertir archivos a Markdown

+
+
+

CONVERSIÓN LOCAL

+

Convierte tus documentos a Markdown

+

Sube un archivo, revisa el resultado y descarga tu .md.

+
-
- - - +
+
+
+ + +
PDF, Word, PowerPoint, Excel, HTML o CSV · máximo 100 MB
+
+ + +
-
-
-
Markdown generado
+
+
- - - +

RESULTADO

+

Markdown generado

+
+
+ + +
- 

+                

             

-        

+        

 
-        

-            
Historial

-            
-            
    
-        
    
+        
     
    
 
    
 
 @section Scripts {
-    
+    
 }
diff --git a/src/Web/appsettings.json b/src/Web/appsettings.json
index 6d710c5..65b9e43 100644
--- a/src/Web/appsettings.json
+++ b/src/Web/appsettings.json
@@ -3,7 +3,7 @@
     "DefaultConnection": "Data Source=MarkItDownWeb.db"
   },
   "MarkItDown": {
-    "PythonPath": "C:\\Python314\\python.exe"
+    "PythonPath": "python"
   },
   "Storage": {
     "UploadsFolder": "Uploads",
diff --git a/src/Web/wwwroot/css/site.css b/src/Web/wwwroot/css/site.css
index 819f612..68ed224 100644
--- a/src/Web/wwwroot/css/site.css
+++ b/src/Web/wwwroot/css/site.css
@@ -19,6 +19,94 @@ html {
 
 body {
   margin-bottom: 60px;
+  background: #f5f7fa;
+  color: #17212f;
+}
+
+.navbar {
+  background: #fff;
+}
+
+.conversion-shell {
+  max-width: 1440px;
+  margin: 2.25rem auto;
+}
+
+.page-heading h1 {
+  max-width: 760px;
+  font-size: clamp(1.8rem, 3vw, 2.5rem);
+  font-weight: 700;
+  letter-spacing: -0.035em;
+}
+
+.eyebrow {
+  color: #64748b;
+  font-size: .7rem;
+  font-weight: 700;
+  letter-spacing: .12em;
+}
+
+.app-card {
+  height: 100%;
+  border: 1px solid #e4e9f0;
+  border-radius: 14px;
+  box-shadow: 0 8px 28px rgb(15 23 42 / 4%);
+}
+
+.status-text {
+  font-size: .9rem;
+}
+
+.markdown-output {
+  min-height: 360px;
+  max-height: 65vh;
+  overflow: auto;
+  margin: 0;
+  padding: 1rem;
+  border: 1px solid #e2e8f0;
+  border-radius: 9px;
+  background: #f8fafc;
+  color: #243247;
+  font: .875rem/1.6 ui-monospace, SFMono-Regular, Menlo, Consolas, monospace;
+  white-space: pre-wrap;
+  overflow-wrap: anywhere;
+}
+
+.history-list {
+  max-height: 65vh;
+  overflow-y: auto;
+}
+
+.history-entry {
+  display: grid;
+  width: 100%;
+  gap: .25rem;
+  padding: .4rem 0;
+  border: 0;
+  background: transparent;
+  color: inherit;
+  text-align: left;
+}
+
+.history-entry:not(:disabled):hover .history-file-name {
+  color: #0d6efd;
+}
+
+.history-entry:disabled {
+  color: #94a3b8;
+  cursor: not-allowed;
+}
+
+.history-file-name {
+  overflow: hidden;
+  font-weight: 600;
+  text-overflow: ellipsis;
+  white-space: nowrap;
+}
+
+.history-date {
+  color: #64748b;
+  font-size: .75rem;
 }
 
 .form-floating > .form-control-plaintext::placeholder, .form-floating > .form-control::placeholder {
@@ -28,4 +116,4 @@ body {
 
 .form-floating > .form-control-plaintext:focus::placeholder, .form-floating > .form-control:focus::placeholder {
   text-align: start;
-}
\ No newline at end of file
+}
diff --git a/src/Web/wwwroot/js/conversion.js b/src/Web/wwwroot/js/conversion.js
index 0b853ff..f51bb3f 100644
--- a/src/Web/wwwroot/js/conversion.js
+++ b/src/Web/wwwroot/js/conversion.js
@@ -1,82 +1,157 @@
 const fileInput = document.getElementById('fileInput');
 const btnConvert = document.getElementById('btnConvert');
+const btnCopy = document.getElementById('btnCopy');
+const btnDownload = document.getElementById('btnDownload');
+const btnClear = document.getElementById('btnClear');
 const statusText = document.getElementById('statusText');
 const markdownOutput = document.getElementById('markdownOutput');
+const resultHeading = document.getElementById('resultHeading');
 const historyList = document.getElementById('historyList');
+const historyEmpty = document.getElementById('historyEmpty');
 const searchHistory = document.getElementById('searchHistory');
 
+let currentFileName = '';
+
+function setMarkdown(markdown, fileName = '') {
+    markdownOutput.textContent = markdown || '';
+    currentFileName = fileName;
+    resultHeading.textContent = fileName ? `${fileName} → Markdown` : 'Markdown generado';
+    const hasMarkdown = markdownOutput.textContent.length > 0;
+    btnCopy.disabled = !hasMarkdown;
+    btnDownload.disabled = !hasMarkdown;
+}
+
+async function readResponse(response) {
+    const body = await response.json().catch(() => ({}));
+    if (!response.ok) {
+        throw new Error(body.error || 'No se pudo completar la solicitud.');
+    }
+    return body;
+}
+
 btnConvert.addEventListener('click', async () => {
     const file = fileInput.files[0];
     if (!file) {
         statusText.textContent = 'Selecciona un archivo primero.';
+        fileInput.focus();
         return;
     }
 
     const formData = new FormData();
     formData.append('file', file);
-
     statusText.textContent = 'Convirtiendo...';
     btnConvert.disabled = true;
 
     try {
         const response = await fetch('/Conversion/Convert', { method: 'POST', body: formData });
-        const result = await response.json();
-
-        if (result.success) {
-            markdownOutput.textContent = result.markdown;
-            statusText.textContent = `Completado en ${result.conversionTimeSeconds.toFixed(2)}s`;
-            loadHistory();
-        } else {
-            statusText.textContent = result.errorMessage || 'Error al convertir.';
+        const result = await readResponse(response);
+        if (!result.success) {
+            throw new Error(result.errorMessage || 'No se pudo convertir el archivo.');
         }
-    } catch (err) {
-        statusText.textContent = 'Error de red.';
+
+        setMarkdown(result.markdown, result.fileName || file.name);
+        statusText.textContent = `Listo en ${Number(result.conversionTimeSeconds).toFixed(2)} s.`;
+        await loadHistory(searchHistory.value);
+    } catch (error) {
+        statusText.textContent = error.message || 'Error de conexión. Comprueba que la aplicación siga abierta.';
     } finally {
         btnConvert.disabled = false;
     }
 });
 
-document.getElementById('btnCopy').addEventListener('click', () => {
-    navigator.clipboard.writeText(markdownOutput.textContent);
+btnCopy.addEventListener('click', async () => {
+    try {
+        await navigator.clipboard.writeText(markdownOutput.textContent);
+        statusText.textContent = 'Markdown copiado al portapapeles.';
+    } catch {
+        statusText.textContent = 'No se pudo acceder al portapapeles en este navegador.';
+    }
 });
 
-document.getElementById('btnDownload').addEventListener('click', () => {
-    const blob = new Blob([markdownOutput.textContent], { type: 'text/markdown' });
+btnDownload.addEventListener('click', () => {
+    const blob = new Blob([markdownOutput.textContent], { type: 'text/markdown;charset=utf-8' });
     const url = URL.createObjectURL(blob);
-    const a = document.createElement('a');
-    a.href = url;
-    a.download = 'resultado.md';
-    a.click();
-    URL.revokeObjectURL(url);
+    const link = document.createElement('a');
+    const baseName = (currentFileName || 'documento').replace(/\.[^.]+$/, '');
+    const safeName = baseName.replace(/[\\/:*?"<>|\u0000-\u001f]/g, '_').trim() || 'documento';
+    link.href = url;
+    link.download = `${safeName}.md`;
+    document.body.appendChild(link);
+    link.click();
+    link.remove();
+    setTimeout(() => URL.revokeObjectURL(url), 1000);
+    statusText.textContent = `Descarga iniciada: ${safeName}.md`;
 });
 
-document.getElementById('btnClear').addEventListener('click', () => {
-    markdownOutput.textContent = '';
+btnClear.addEventListener('click', () => {
+    setMarkdown('');
     fileInput.value = '';
     statusText.textContent = '';
+    fileInput.focus();
 });
 
 async function loadHistory(search = '') {
-    const response = await fetch(`/Conversion/History?search=${encodeURIComponent(search)}`);
-    const items = await response.json();
-
-    historyList.innerHTML = '';
-    items.forEach(item => {
-        const li = document.createElement('li');
-        li.className = 'list-group-item list-group-item-action';
-        li.style.cursor = 'pointer';
-        li.textContent = item.fileName;
-        li.addEventListener('click', () => openHistoryItem(item.id));
-        historyList.appendChild(li);
-    });
+    try {
+        const response = await fetch(`/Conversion/History?search=${encodeURIComponent(search)}`);
+        const items = await readResponse(response);
+        historyList.replaceChildren();
+
+        items.forEach(item => {
+            const entry = document.createElement('li');
+            entry.className = 'list-group-item px-0';
+
+            const button = document.createElement('button');
+            button.type = 'button';
+            button.className = 'history-entry';
+            button.disabled = item.status !== 'Completed';
+            button.setAttribute('aria-label', `Abrir ${item.fileName}`);
+
+            const name = document.createElement('span');
+            name.className = 'history-file-name';
+            name.textContent = item.fileName;
+
+            const detail = document.createElement('span');
+            detail.className = 'history-date';
+            detail.textContent = item.status === 'Completed'
+                ? new Date(item.createdAt).toLocaleString()
+                : 'No se pudo convertir';
+
+            button.append(name, detail);
+            button.addEventListener('click', () => openHistoryItem(item.id));
+            entry.appendChild(button);
+            historyList.appendChild(entry);
+        });
+
+        historyEmpty.hidden = items.length > 0;
+        historyEmpty.textContent = search ? 'No hay resultados para esa búsqueda.' : 'Aún no hay conversiones guardadas.';
+    } catch {
+        historyEmpty.hidden = false;
+        historyEmpty.textContent = 'No se pudo cargar el historial.';
+    }
 }
 
 async function openHistoryItem(id) {
-    const response = await fetch(`/Conversion/Item?id=${id}`);
-    const result = await response.json();
-    markdownOutput.textContent = result.markdown;
+    statusText.textContent = 'Abriendo conversión...';
+    try {
+        const response = await fetch(`/Conversion/Item?id=${encodeURIComponent(id)}`);
+        const result = await readResponse(response);
+        setMarkdown(result.markdown, result.fileName);
+        statusText.textContent = result.markdown ? 'Conversión cargada desde el historial.' : 'Esta conversión no tiene Markdown disponible.';
+    } catch (error) {
+        statusText.textContent = error.message;
+    }
 }
 
-searchHistory.addEventListener('input', (e) => loadHistory(e.target.value));
+let historySearchTimer;
+searchHistory.addEventListener('input', event => {
+    clearTimeout(historySearchTimer);
+    historySearchTimer = setTimeout(() => loadHistory(event.target.value), 200);
+});
+
+fileInput.addEventListener('change', () => {
+    if (fileInput.files[0]) {
+        statusText.textContent = `${fileInput.files[0].name} seleccionado.`;
+    }
+});
 
 loadHistory();
diff --git a/tests/MarkItDownWeb.Tests/FileSignatureValidatorTests.cs b/tests/MarkItDownWeb.Tests/FileSignatureValidatorTests.cs
new file mode 100644
index 0000000..7bd68e2
--- /dev/null
+++ b/tests/MarkItDownWeb.Tests/FileSignatureValidatorTests.cs
@@ -0,0 +1,68 @@
+using MarkItDownWeb.Infrastructure.Security;
+using Xunit;
+
+namespace MarkItDownWeb.Tests;
+
+public class FileSignatureValidatorTests
+{
+    [Fact]
+    public void AcceptsPdfWithPdfHeader()
+        => AssertValid(".pdf", [0x25, 0x50, 0x44, 0x46, 0x2D]);
+
+    [Theory]
+    [InlineData(".docx")]
+    [InlineData(".pptx")]
+    [InlineData(".xlsx")]
+    public void AcceptsOfficeFilesWithZipHeader(string extension)
+        => AssertValid(extension, [0x50, 0x4B, 0x03, 0x04, 0x14, 0x00]);
+
+    [Theory]
+    [InlineData(".pdf")]
+    [InlineData(".docx")]
+    [InlineData(".pptx")]
+    [InlineData(".xlsx")]
+    public void RejectsContentThatDoesNotMatchExtension(string extension)
+        => AssertInvalid(extension, [0x00, 0x01, 0x02, 0x03]);
+
+    [Fact]
+    public void RejectsFileShorterThanSignature()
+        => AssertInvalid(".pdf", [0x25, 0x50]);
+
+    [Theory]
+    [InlineData(".html")]
+    [InlineData(".csv")]
+    public void TextFormatsDoNotRequireBinarySignature(string extension)
+        => AssertValid(extension, [0x00]);
+
+    private static void AssertValid(string extension, byte[] contents)
+    {
+        using var file = TemporaryFile.Create(extension, contents);
+        Assert.True(FileSignatureValidator.IsValid(file.Path, extension));
+    }
+
+    private static void AssertInvalid(string extension, byte[] contents)
+    {
+        using var file = TemporaryFile.Create(extension, contents);
+        Assert.False(FileSignatureValidator.IsValid(file.Path, extension));
+    }
+
+    private sealed class TemporaryFile : IDisposable
+    {
+        private TemporaryFile(string path) => Path = path;
+
+        public string Path { get; }
+
+        public static TemporaryFile Create(string extension, byte[] contents)
+        {
+            var path = System.IO.Path.Combine(System.IO.Path.GetTempPath(), $"{Guid.NewGuid()}{extension}");
+            File.WriteAllBytes(path, contents);
+            return new TemporaryFile(path);
+        }
+
+        public void Dispose()
+        {
+            if (File.Exists(Path))
+                File.Delete(Path);
+        }
+    }
+}
diff --git a/tests/MarkItDownWeb.Tests/MarkItDownWeb.Tests.csproj b/tests/MarkItDownWeb.Tests/MarkItDownWeb.Tests.csproj
new file mode 100644
index 0000000..c2ccc91
--- /dev/null
+++ b/tests/MarkItDownWeb.Tests/MarkItDownWeb.Tests.csproj
@@ -0,0 +1,22 @@
+
+
+  
+    
+    
+    
+      runtime; build; native; contentfiles; analyzers; buildtransitive
+      all
+    
+  
+
+  
+    
+  
+
+  
+    net10.0
+    enable
+    enable
+  
+
+