Encoder interface

1 files changed, 39 insertions, 0 deletions

diff --git a/linguistics/src/main/java/com/yahoo/language/process/Encoder.java b/linguistics/src/main/java/com/yahoo/language/process/Encoder.java
new file mode 100644
index 00000000000..91de16f669b
--- /dev/null
+++ b/linguistics/src/main/java/com/yahoo/language/process/Encoder.java
@@ -0,0 +1,39 @@
+// Copyright Yahoo. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
+package com.yahoo.language.process;
+
+import com.yahoo.language.Language;
+import com.yahoo.tensor.Tensor;
+import com.yahoo.tensor.TensorType;
+
+import java.util.List;
+
+/**
+ * An encoder converts a text string to a tensor or list of tokens
+ *
+ * @author bratseth
+ */
+public interface Encoder {
+
+    /**
+     * Encodes text into tokens in a list of ids.
+     *
+     * @param text the text to encode
+     * @param language the language of the text, or UNKNOWN to use language independent encoding
+     * @return the text encoded to a list of segment ids
+     * @throws IllegalArgumentException if the language is not supported by this encoder
+     */
+    List<Integer> encode(String text, Language language);
+
+    /**
+     * Encodes text into tokens in a tensor.
+     * The information contained in the encoding may depend on the tensor type.
+     *
+     * @param text the text to encode
+     * @param language the language of the text, or UNKNOWN to use language independent encoding
+     * @param tensorType the type of the ttensor to be returned
+     * @return the tex encoded into a tensor of the supplied type
+     * @throws IllegalArgumentException if the language or tensor type is not supported by this encoder
+     */
+    Tensor encode(String text, Language language, TensorType tensorType);
+
+}