Help with the new ReadFileStreamAsync method

[ManuelHaas]

Following the update to 26.3.0, the ReadFileStreamAsync method is now automatically called instead of the ReadFileAsync method.

1. What needs to be enabled in the settings to ensure that the ReadFileAsync method is used again?
2. Can you provide an example that handles the use case “File is loaded from a database and displayed via the viewer” using the ReadFileStreamAsync method?

Thanks

[vladimir-litvinchik]

Hi, @ManuelHaas!

Thanks for reporting this, will take a look and update you.

[vladimir-litvinchik]

Hi, @ManuelHaas

After GetFileStreamAsync method was added in 26.3.0 the [BaseViewer.cs#L204](https://github.com/groupdocs-viewer/GroupDocs.Viewer-for-.NET-UI/blob/main/src/GroupDocs.Viewer.UI.SelfHost.Api.Shared/Viewers/BaseViewer.cs#L204] was updated to call ReadFileStreamAsync.

Previously this called ReadFileAsync and wrapped the resulting byte[] in a new MemoryStream. The change avoids buffering the whole file in memory important for large documents.

Unfortunately, there is no setting to turn previous logic back and you have to implement ReadFileStreamAsync to make it work.

I have prepared sample app that shows simulates DB in memory using Dictionary<string, byte[]>. In a real app you would replace the seeding with a query against your BLOB column.

The key part is Program.cs where DatabaseFileStorage is registered and DatabaseFileStorage.cs that implements IFileStorage where ReadFileStreamAsync returns a MemoryStream over the cached bytes.

Sample app: custom-file-storage.zip

To verify set breakpoint at ReadFileStreamAsync, run and open https://localhost:58023/viewer/?file=business-plan.docx

Let me know if it works for you.

[ManuelHaas]

@vladimir-litvinchik Thanks for the sample app. That’s exactly how I already implemented it. I thought there might be a more efficient way, so that only the part of the data that the viewer actually needs is loaded from the database.