Caching taxonomy object model?

[ValeriyTitov]

For me, this line of code take about 30 second to complete (it parse hundreds of files)
SchemaSet.cs

public void Compile()
        {
            thisSchemaSet.Compile();
        }

As i understand, it parse xml files in Taxonomy to create Object model of Taxonomy.
After object model is built, Gepsio load all possible facts from instance document.

So, for each instance of xbrl report Gepsio doing same thing - Creating object model of Taxonomy from the scratch. But if multiply xbrl-Instance files reference same Taxononmy, it not nessesary to create this.

var doc = new XbrlDocument();
doc.Taxonomy.Build("http://example.com/Taxonomy.xsd");
doc.Taxonomy.SaveToSteam(Stream);
doc.Load("Instance.xml");

var doc2 = new XbrlDocument();
doc2.Taxonomy.LoadFromSteam(Stream);
doc2.Load("Instance2.xml");

It will be my problem, if taxonomy mismatch.
Or is there something i misunderstood ?

[JeffFerguson]

You're certainly right. There is definite room for improvement there. Gepsio could take the performance hit on the first compilation, but store the compiled results for future XBRL instance loads.

We could also take this one step further, if necessary. The schemas never change, once published. Once compiled, perhaps the results of the compilation could be serialized to disk so that they could be loaded more efficiently during subsequent runs. For a given schema, the schema set compilation output will never change, even between different processes running Gepsio. Can the results be stored somewhere on disk so that no compilations are ever again needed for a particular schema with a particular namespace? It's worth thinking about.

[KacperJab]

Hi,
I have a question regarding this topic. How can I view the taxonomy model (i.e. ESEF IFRS taxonomy) using gepsio?