Class Word2VecReader

java.lang.Object
org.episteme.core.io.AbstractResourceReader<Map<String, Vector<Real>>>
org.episteme.social.linguistics.loaders.Word2VecReader
All Implemented Interfaces:
ResourceIO<Map<String, Vector<Real>>>, ResourceReader<Map<String, Vector<Real>>>
public class Word2VecReader extends AbstractResourceReader<Map<String, Vector<Real>>>
Loader for Word2Vec models (text format). Modernized to extend AbstractResourceReader for standardized embedding loading.
Since:
1.0
Author:
Silvere Martin-Michiellot, Gemini AI (Google DeepMind)

  • Constructor Details

    • Word2VecReader

      public Word2VecReader()
      Default constructor.

  • Method Details

    • loadFromSource

      protected Map<String, Vector<Real>> loadFromSource(String path) throws Exception
      Specified by:
      loadFromSource in class AbstractResourceReader<Map<String, Vector<Real>>>
      Throws:
      Exception

    • loadFromInputStream

      protected Map<String, Vector<Real>> loadFromInputStream(InputStream is, String id) throws Exception
      Overrides:
      loadFromInputStream in class AbstractResourceReader<Map<String, Vector<Real>>>
      Throws:
      Exception

    • loadTextModel

      public static Map<String, Vector<Real>> loadTextModel(File file) throws IOException
      Loads a Word2Vec model in text format.
      Throws:
      IOException

    • getLongDescription

      public String getLongDescription()
      Description copied from interface: ResourceIO
      Returns a long description of this resource handler. MUST be implemented with I18N support.
      Returns:
      the long description

    • cosineSimilarity

      public static double cosineSimilarity(Vector<Real> v1, Vector<Real> v2)
      Calculates cosine similarity between two vectors.

    • getResourceType

      public Class<Map<String, Vector<Real>>> getResourceType()
      Description copied from interface: ResourceIO
      Returns the type of resource.

    • getName

      public String getName()
      Description copied from interface: ResourceIO
      Returns the display name of this resource handler. MUST be implemented with I18N support.
      Returns:
      the display name

    • getDescription

      public String getDescription()
      Description copied from interface: ResourceIO
      Returns a short description of this resource handler. MUST be implemented with I18N support.
      Returns:
      the description

    • getCategory

      public String getCategory()
      Description copied from interface: ResourceIO
      Returns the category for grouping. MUST be implemented with I18N support.
      Returns:
      the category name

    • getResourcePath

      public String getResourcePath()
      Description copied from interface: ResourceIO
      Returns the base path where this resource is located.

    • getSupportedVersions

      public String[] getSupportedVersions()
      Description copied from interface: ResourceIO
      Returns the supported versions of the format this reader/writer handles.

      Each implementation MUST override this method to declare which versions of the underlying format are supported. The returned array should contain version strings in the format's standard notation (e.g., "3.0", "2.1", "Level 3 Version 2").

      Examples:

      • MathML: {"3.0", "2.0"}
      • SBML: {"Level 3 Version 2", "Level 3 Version 1", "Level 2 Version 5"}
      • PhyloXML: {"1.10", "1.00"}

      Returns:
      array of supported version strings, never null (empty array if version-agnostic)