move PuppeterSharp ParseAsync method

win7user10 · May 27, 2024 · 5c4c8f8 · 5c4c8f8
1 parent 9554a7d
commit 5c4c8f8
Show file tree

Hide file tree

Showing 2 changed files with 32 additions and 2 deletions.
diff --git a/src/Laraue.Crawling.Dynamic.PuppeterSharp/Extensions/DocumentSchemaParserExtensions.cs b/src/Laraue.Crawling.Dynamic.PuppeterSharp/Extensions/DocumentSchemaParserExtensions.cs
@@ -0,0 +1,31 @@
+using Laraue.Crawling.Abstractions;
+using PuppeteerSharp;
+
+namespace Laraue.Crawling.Dynamic.PuppeterSharp.Extensions;
+
+/// <summary>
+/// Transformers from opened browser page to the crawling model.
+/// </summary>
+public static class DocumentSchemaParserExtensions
+{
+    /// <summary>
+    /// Take the opened page, use the passed schema and returns the crawling result.
+    /// </summary>
+    /// <param name="parser"></param>
+    /// <param name="page"></param>
+    /// <param name="schema"></param>
+    /// <typeparam name="TResult"></typeparam>
+    /// <returns></returns>
+    /// <exception cref="InvalidOperationException"></exception>
+    public static async Task<TResult> ParseAsync<TResult>(
+        this IDocumentSchemaParser<IElementHandle, HtmlSelector> parser,
+        IPage page,
+        ICompiledDocumentSchema<IElementHandle, HtmlSelector, TResult> schema)
+    {
+        var element = await page.QuerySelectorAsync("body")
+            .ConfigureAwait(false);
+
+        return await parser.RunAsync(schema, element).ConfigureAwait(false)
+               ?? throw new InvalidOperationException("Tag <body> has not been found in the passed page");
+    }
+}
diff --git a/src/Laraue.Crawling.Dynamic.PuppeterSharp/Extensions/ServiceCollectionExtensions.cs b/src/Laraue.Crawling.Dynamic.PuppeterSharp/Extensions/ServiceCollectionExtensions.cs
@@ -24,7 +24,6 @@ public static class ServiceCollectionExtensions
     {
         return serviceCollection
             .AddSingleton<IBrowserFactory>(sp => new BrowserFactory(launchOptions, sp.GetRequiredService<ILoggerFactory>()))
-            .AddSingleton<IDocumentSchemaParser<IElementHandle, HtmlSelector>, PuppeterSharpParser>()
-            .AddSingleton<IPageParser, PageParser>();
+            .AddSingleton<IDocumentSchemaParser<IElementHandle, HtmlSelector>, PuppeterSharpParser>();
     }
 }