Package org.apache.tika.ml.chardetect

Interface StatisticalSpecialist

All Known Implementing Classes:
Utf16SpecialistEncodingDetector
public interface StatisticalSpecialist
SPI contract for an MoE charset-detection specialist. Discovered via ServiceLoader at META-INF/services/org.apache.tika.ml.chardetect.StatisticalSpecialist. Implementations must be thread-safe.

  • Method Summary

    Modifier and Type
    Method
    Description
    String
    getName()
    Short name: "utf16", "sbcs", etc.
    SpecialistOutput
    score(byte[] probe)
    Per-class logits for the probe, or null to decline (probe too short, hard-gated, etc.).

  • Method Details

    • getName

      String getName()
      Short name: "utf16", "sbcs", etc.

    • score

      SpecialistOutput score(byte[] probe)
      Per-class logits for the probe, or null to decline (probe too short, hard-gated, etc.). Declining contributes nothing; a low-scoring result contributes weak signal.