Saving EditableDocument to stream Leave feedback

This article shows and explains the advances techniques and approaches while working with EditableDocument in advanced level — saving to stream with resource callback, saving resources separately from HTML markup, and saving HTML markup with adjustable resource links.

Note

Information, explained in this article, is applicable for the GroupDocs.Editor for Java version 24.4 and higher.

In the latest versions of the GroupDocs.Editor for Java, especially in the version 24.4, the EditableDocument class, which is one of the most important among all types in the GroupDocs.Editor, was significantly expanded and improved. These new features are focused on saving the EditableDocument to the HTML format into the stream using different ways, and allowing to specify a resource template string. This article covers all of them.

Starting from the version 23.9, the EditableDocument class contains a new instane method getContent, which allows to save an HTML marjup of the HTML document to the specified stream with specified encoding. Its signature is below:

public String getContent()

This method internally converts the content of the EditableDocument instance to the HTML markup, where CSS stylesheets and images are not stored inside the markup. These external resources (CSS and images) are referenced through the links (URIs) by specifying their filenames. So, for example, if one HTML document has one stylesheet and two images with filenames “foo.jpeg” and “bar.png”, they will be referenced in the HTML markup by the next:

<link href="stylesheet.css" rel="stylesheet" />
.....
<IMG src="foo.jpeg" />
.....
<IMG src="bar.png" />

With this method, there is no option to specify or adjust the reference to the external resources somehow. Also, this method is useful for saving the HTML markup only, but not the external resources.

For dealing with these issues the next way was introduced.

Starting from the version 24.4 a new save() method overload was added:

void save(Writer htmlMarkup, HtmlSaveOptions saveOptions)

This method enables the most advanced, detailed and adjustable saving of the EditableDocument instance to the HTML format.

First parameter is an output parameter, into which the HTML markup will be written. User must specify any valid implementation of the java.io.Writer abstract class, usually java.io.Writer or java.io.FileWriter. It cannot be null.

Second parameter is the new public class HtmlSaveOptions, which is described below. This argument also canot be null.

Class HtmlSaveOptions contains 4 properties, only one of them is mandatory and cannot be null.

• HtmlTagCase property controls how the HTML tag names will be present in HTML markup: All lower case (default value), All upper case, or First letter upper case. Its value is a special enum - TagRenderingCase.
• AttributeValueDelimiter property controls which delimiter around the attribute values in HTML elements will be used: single quote (default value) or double quote. Its value is a special value type - QuoteType.
• EmbedStylesheetsIntoMarkup property is a boolean flag, which controls where to store the CSS stylesheet(s): as external resources (false), or embed them into the HTML markup, inside the STYLE element in the HTML->HEAD section (true). By default is false — do not embed inside the HTML markup.
• SavingCallback is the most important and the only mandatory property, that myust be specified by user and cannot be null. It is a reference to the user-defined resource callback, which implements an interface IHtmlSavingCallback.

When the save method is called, the GroupDocs.Editor scans the content of EditableDocument instance and gathers all resources: images, stylesheets, audio, etc. For each of these resources it calls the saveOneResource method, which is defined in the IHtmlSavingCallback interface and implemented by the user. When calling this method, the GroupDocs.Editor passes the resource itself as a type, which implements a IHtmlResource interface — it will be JpegImage for JPEG images, CssText for CSS stylesheets, Mp3Audio for MP3 audio files and so on.

When user obtains such a resource instance in the saveOneResource method, he can do with it what he wants: save to the disk, to the database, convert it somehow, or even do nothing. What is important, the user must return a link to this resource through the return value. So when the GroupDocs.Editor will start to form an HTML markup in order to write it to the java.io.Writer htmlMarkup argument, it will use the user-provided links to the resources in it.

The complete code sample below demonstrates the creating of EditableDocument instance by opening an arbitrary DOCX file for edit, then creationg the custom implementation of the IHtmlSavingCallback, which saves the provided resources to the database, and saving the content of the EditableDocument in the HTML format.

public void SavingToStreamPublicExample()
{
	//Define an input document
	String inputDocx = "Path\\SampleDoc.docx";

	//Create and instance of custom SavingCallback
	SavingCallback callback = new SavingCallback();

	//Create HTML saving options
	HtmlSaveOptions saveOptions = new HtmlSaveOptions();
	saveOptions.setEmbedStylesheetsIntoMarkup(false);//Do not embed CSS into markup
	saveOptions.setAttributeValueDelimiter(QuoteType.DoubleQuote.getCode());
	saveOptions.setHtmlTagCase(TagRenderingCase.UpperCase);
	saveOptions.setSavingCallback(callback);//Specify our custom callback

	//Create an Editor class instance - load options are not necessary for DOCX and are default here
	Editor editor = new Editor(inputDocx);
	{
		//Create an EditableDocument instance - edit options are not specified and thus are default
		using (EditableDocument document = editor.edit())
		{
			//Define a StringWriter, into which the HTML markup will be written
			StringWriter writer = new StringWriter;
			document.save(writer, saveOptions);
		}
	}
}

	private final static class SavingCallback implements IHtmlSavingCallback
	{
		public final String saveOneResource(IHtmlResource resource)
		{
			//save this stream to the DB or to the disk somehow
			InputStream resourceStream = resource.getByteContent();

			//form and return a reference to the resource depending on resource type
			if (Operators.is(resource, CssText.class))
			{
				return StringExtensions.format("https://mywebsite.web/stylesheet/{0}", resource.getFilenameWithExtension());
			}
			else
			{
				return StringExtensions.format("https://mywebsite.web/GetImage/name={0}", resource.getFilenameWithExtension());
			}
		}
	}

After executing this code, the markup variable of StringWriter type will contain an HTML markup, where the references to the external resources will be like:

<link href="https://mywebsite.web/stylesheet/stylesheet.css" rel="stylesheet"/>
.....
<IMG src="https://mywebsite.web/GetImage/name=foo.jpeg"/>
.....
<IMG src="https://mywebsite.web/GetImage/name=bar.png"/>

Sometimes it is necessary to save only markup, and do not save the resources. In such cases the approach, described above, may be treates as too redundant. The EditableDocument class from its beginning contains methods getContent() and getBodyContent(). These two methods have two overloads, but in general the first one returns the whole HTML markup of the document, while the second — only the inner content of the HTML->BODY element, without the BODY opening and closing tags.

Before the version 24.4 was released, these two methods allowed to specify the external resource prefixes, so when generating the HTML markup, these prefixes are added to resource names. For example, if the document has two images, let’s say, “foo.jpeg” and “bar.png”, then, by calling the EditableDocument.getBodyContent("https://mywebsite.web/image?name="), the output markup will contain two IMG elements with the next content:

<IMG src="https://mywebsite.web/image?name=foo.jpeg"/>
.....
<IMG src="https://mywebsite.web/image?name=bar.png"/>

However, this state of things had two main flaws:

• It was not possible to insert the resource name inside the request URI, only at the end.
• It was not possible to insert the resource name more than once in the single request URI.

GroupDocs.Editor version 24.4 fixes both these issues. The methods signatures are intact, but now they treat the specified strings not as the prefixes, but as the template format strings. Protocol is the next:

1. If specified template string is null or empty, then pure resource name will be placed in the HTML markup.
2. If specified template string contains one unique valid placeholder, which occurs one or more times, the resource name will replace the placeholder in the HTML markup.
3. Otherwise, if specified template string contains more than one unique placeholder, or contains no placeholder at all, it will be treated as a prefix string.

Due to this protocol the old-written user code will work the same as for the previous versions, because the template strings do not have the placeholders and thus are treated as prefixes.

But when using the template string, now, when calling the EditableDocument.getBodyContent("https://mywebsite.web/image?name={0}&stamp=456789&token={0}&expireIn=123987") method, the image named “foo.jpeg” will be presented in the output markup as the next:

<IMG src="https://mywebsite.web/image?name=foo.jpeg&stamp=456789&token=foo.jpeg&expireIn=123987"/>

So this new mechanism allows to specify the request URIs for the external images and stylesheets very precisely.