refactor: Renamed CollectMetadata to CollectMetadataIfRequired.

HavenDV · HavenDV · commit c38570f7f345 · 2024-05-04T02:28:46.000+04:00
diff --git a/src/DocumentLoaders/Abstractions/src/DataSourceExtensions.cs b/src/DocumentLoaders/Abstractions/src/DataSourceExtensions.cs
@@ -28,7 +28,7 @@ public static IReadOnlyDictionary<string, object> ToMetadata(this DataSource dat
     /// <param name="settings"></param>
     /// <param name="dataSource"></param>
     /// <returns></returns>
-    public static IReadOnlyDictionary<string, object>? CollectMetadata(
+    public static IReadOnlyDictionary<string, object>? CollectMetadataIfRequired(
         this DocumentLoaderSettings? settings,
         DataSource dataSource)
     {
@@ -43,11 +43,11 @@ public static IReadOnlyDictionary<string, object> ToMetadata(this DataSource dat
     /// <param name="metadata"></param>
     /// <param name="additionalMetadata"></param>
     /// <returns></returns>
-    public static IReadOnlyDictionary<string, object>? With(
-        this IReadOnlyDictionary<string, object>? metadata,
+    public static IReadOnlyDictionary<string, object> With(
+        this IReadOnlyDictionary<string, object> metadata,
         IReadOnlyDictionary<string, object> additionalMetadata)
     {
-        return metadata?
+        return metadata
             .Concat(additionalMetadata)
             .ToDictionary(pair => pair.Key, pair => pair.Value);
     }
diff --git a/src/DocumentLoaders/Abstractions/src/FileLoader.cs b/src/DocumentLoaders/Abstractions/src/FileLoader.cs
@@ -35,8 +35,8 @@ public async Task<IReadOnlyCollection<Document>> LoadAsync(
         //     }
         // }
 
-        var metadata = settings.CollectMetadata(dataSource);
-
+        var metadata = settings.CollectMetadataIfRequired(dataSource);
+        
         return [new Document(content, metadata: metadata)];
     }
 }
diff --git a/src/DocumentLoaders/Pdf/src/AsposePdfSource.cs b/src/DocumentLoaders/Pdf/src/AsposePdfSource.cs
@@ -20,8 +20,8 @@ public async Task<IReadOnlyCollection<Document>> LoadAsync(
         var textAbsorber = new TextAbsorber();
         pdfDocument.Pages.Accept(textAbsorber);
 
-        var metadata = settings.CollectMetadata(dataSource);
-
+        var metadata = settings.CollectMetadataIfRequired(dataSource);
+        
         return [new Document(textAbsorber.Text, metadata: metadata)];
     }
 }
diff --git a/src/DocumentLoaders/Pdf/src/PdfPigPdfSource.cs b/src/DocumentLoaders/Pdf/src/PdfPigPdfSource.cs
@@ -18,11 +18,11 @@ public async Task<IReadOnlyCollection<Document>> LoadAsync(
         using var stream = await dataSource.GetStreamAsync(cancellationToken).ConfigureAwait(false);
         using var document = PdfDocument.Open(stream, new ParsingOptions());
 
-        var metadata = settings.CollectMetadata(dataSource);
+        var metadata = settings.CollectMetadataIfRequired(dataSource);
 
         return document
             .GetPages()
-            .Select(page => new Document(page.Text, metadata.With(new Dictionary<string, object>
+            .Select(page => new Document(page.Text, metadata?.With(new Dictionary<string, object>
             {
                 { "page", page.Number },
             })))
diff --git a/src/DocumentLoaders/WebBase/src/HtmlLoader.cs b/src/DocumentLoaders/WebBase/src/HtmlLoader.cs
@@ -33,7 +33,7 @@ public async Task<IReadOnlyCollection<Document>> LoadAsync(
             document.QuerySelector("html") ??
             throw new NotSupportedException("Not supported for pages without <html> tag");
 
-        var metadata = settings.CollectMetadata(dataSource);
+        var metadata = settings.CollectMetadataIfRequired(dataSource);
 
         return [new Document(html.TextContent, metadata: metadata)];
     }
diff --git a/src/DocumentLoaders/Word/src/WordLoader.cs b/src/DocumentLoaders/Word/src/WordLoader.cs
@@ -35,7 +35,7 @@ public async Task<IReadOnlyCollection<Document>> LoadAsync(
             }
         }
 
-        var metadata = settings.CollectMetadata(dataSource);
+        var metadata = settings.CollectMetadataIfRequired(dataSource);
 
         return documents
             .Select(text => new Document(text, metadata: metadata))

Original file line number	Diff line number	Diff line change
`@@ -35,8 +35,8 @@ public async Task<IReadOnlyCollection<Document>> LoadAsync(`
`35`	`35`	`// }`
`36`	`36`	`// }`
`37`	`37`
`38`		`- var metadata = settings.CollectMetadata(dataSource);`
`39`		`-`
	`38`	`+ var metadata = settings.CollectMetadataIfRequired(dataSource);`
	`39`	`+`
`40`	`40`	`return [new Document(content, metadata: metadata)];`
`41`	`41`	`}`
`42`	`42`	`}`
Original file line number	Diff line number	Diff line change
`@@ -20,8 +20,8 @@ public async Task<IReadOnlyCollection<Document>> LoadAsync(`
`20`	`20`	`var textAbsorber = new TextAbsorber();`
`21`	`21`	`pdfDocument.Pages.Accept(textAbsorber);`
`22`	`22`
`23`		`- var metadata = settings.CollectMetadata(dataSource);`
`24`		`-`
	`23`	`+ var metadata = settings.CollectMetadataIfRequired(dataSource);`
	`24`	`+`
`25`	`25`	`return [new Document(textAbsorber.Text, metadata: metadata)];`
`26`	`26`	`}`
`27`	`27`	`}`
Original file line number	Diff line number	Diff line change
`@@ -33,7 +33,7 @@ public async Task<IReadOnlyCollection<Document>> LoadAsync(`
`33`	`33`	`document.QuerySelector("html") ??`
`34`	`34`	`throw new NotSupportedException("Not supported for pages without <html> tag");`
`35`	`35`
`36`		`- var metadata = settings.CollectMetadata(dataSource);`
	`36`	`+ var metadata = settings.CollectMetadataIfRequired(dataSource);`
`37`	`37`
`38`	`38`	`return [new Document(html.TextContent, metadata: metadata)];`
`39`	`39`	`}`
Original file line number	Diff line number	Diff line change
`@@ -35,7 +35,7 @@ public async Task<IReadOnlyCollection<Document>> LoadAsync(`
`35`	`35`	`}`
`36`	`36`	`}`
`37`	`37`
`38`		`- var metadata = settings.CollectMetadata(dataSource);`
	`38`	`+ var metadata = settings.CollectMetadataIfRequired(dataSource);`
`39`	`39`
`40`	`40`	`return documents`
`41`	`41`	`.Select(text => new Document(text, metadata: metadata))`