Replace MIME_LABELS map with Tika-derived display strings in FileSignatureDetector

[NotYuSheng]

Problem

FileSignatureDetector.java contains a hardcoded Map<String, String> MIME_LABELS with 47 entries mapping Tika MIME type strings to short badge labels (e.g. "application/pdf" → "PDF", "video/x-matroska" → "MKV"). Any new file type detected by Tika but not in this map falls through to a subtype-parsing fallback.

The fallback logic is already quite good:

String subtype = mime.substring(mime.indexOf('/') + 1);
subtype = subtype.replaceAll("^(x-|vnd\\.)", "").toUpperCase();
label = subtype.length() <= 12 ? subtype : null;

However the explicit map overrides are still needed for cases where the subtype is not a usable label (e.g. "vnd.openxmlformats-officedocument.wordprocessingml.document" → "DOCX").

Proposed Solution

Replace the map with a two-step approach using Tika's own registry:

1. Try MimeType.getExtension() from Tika's MimeTypes registry — the primary extension is typically the canonical short label (.pdf → "PDF", .docx → "DOCX", .mkv → "MKV").
2. Fall back to the subtype-stripping logic already in place for any type Tika does not have an extension for.

private static final MimeTypes TIKA_MIME_REPO = MimeTypes.getDefaultMimeTypes();

public static String labelFromMime(String mime) {
    // Try Tika's registered extension first
    try {
        MimeType mt = TIKA_MIME_REPO.forName(mime);
        String ext = mt.getExtension(); // e.g. ".pdf", ".docx"
        if (ext != null && !ext.isEmpty()) {
            return ext.substring(1).toUpperCase(); // strip leading dot
        }
    } catch (MimeTypeException ignored) {}

    // Fallback: derive from subtype string
    String subtype = mime.contains("/") ? mime.substring(mime.indexOf('/') + 1) : mime;
    subtype = subtype.replaceAll("^(x-|vnd\\.)", "").toUpperCase();
    return subtype.length() <= 12 ? subtype : null;
}

This approach works for any MIME type Tika recognises, not just the 47 currently listed.

Note on quality

Tika's getExtension() maps directly to the well-known file extension registered for each type, so application/vnd.openxmlformats-officedocument.wordprocessingml.document correctly returns .docx. For audio/video/image types the extension matches the common badge name exactly. This removes the need to maintain the explicit map.

Files to Change

• backend/src/main/java/com/tracepcap/analysis/service/FileSignatureDetector.java

Acceptance Criteria

• MIME_LABELS static map removed
• Label derived from MimeType.getExtension() with subtype fallback
• Existing known types (PDF, ZIP, DOCX, MP3, MKV, ELF) produce the same badge labels
• Previously unlabelled Tika types now receive a label where Tika has an extension