[ Team LiB ] |
14.1 Using the XsdInference NamespaceLike XmlDiffPatch, XsdInference can be added to your Global Assembly Cache for your convenience, and you can add it as a reference in your Visual Studio .NET project or Makefile. The XsdInference namespace contains one class, Infer. Infer has a single overloaded method, InferSchema( ), which returns an XmlSchemaCollection. One overload takes just an XmlReader, and the other one takes an XmlReader and an XmlSchemaCollection. Given an XmlReader for a particular XML document, you can create an XML Schema definition and write it to the console using the code in Example 14-1. Example 14-1. Using the XsdInference assemblyusing System; using System.Xml; using System.Xml.Schema; using Microsoft.XsdInference; public class XsdInfer { public static void Main(string [ ] args) { try { XmlSchemaCollection schemas = null; Infer infer = new Infer( ); XmlTextReader reader = new XmlTextReader(args[0]); schemas = infer.InferSchema(reader); foreach (XmlSchema schema in schemas) { schema.Write(Console.Out); } } catch (Exception e) { Console.Error.WriteLine(e); } } } The usage of the Infer class is fairly straightforward. In Example 14-1, you can see that an instance of Infer is created, and its Infer( ) method is called. Infer( ) returns an XmlSchemaCollection. Example 14-2 shows another pattern of usage for the Infer class, in which several XML documents are used to progressively refine an XML Schema. The changed lines are highlighted. Example 14-2. Using Infer to progressively infer a schemausing System; using System.Xml; using System.Xml.Schema; using Microsoft.XsdInference; public class XsdInfer { public static void Main(string [ ] args) { try { XmlSchemaCollection schemas = new XmlSchemaCollection( ); Infer infer = new Infer( ); foreach (string filename in args) { XmlTextReader reader = new XmlTextReader(filename); schemas = infer.InferSchema(reader, schemas); } foreach (XmlSchema schema in schemas) { schema.Write(Console.Out); } } catch (Exception e) { Console.Error.WriteLine(e); } } } The first change is to create an XmlSchemaCollection to pass into the InferSchema( ) method call. The first time in, this collection will be empty, and InferSchema( ) will infer a schema and place it into the collection. The second change is to repeatedly call InferSchema( ) for each filename on the command line, each time passing in the XmlSchemaCollection returned from the previous call. The schema will be refined by each call, and the final inferred schema will be able to validate any of the XML documents that were used to infer it. Note that this mechanism will work for any number of XML documents. One unique XML schema will be inferred for each different XML namespace in the XML documents. |
[ Team LiB ] |