Deduplicate and curate MIME types database

- Run dedup-mime-types.py to resolve conflicts, preferring image/ types
  over other categories, non-x- subtypes over x- variants, and avoiding
  application/octet-stream
- Commented-out alternatives remain visible for manual adjustment
- Manual overrides: model/ types for 3D formats (obj, ply, stl, xyz),
  image/wmf over windows/metafile, text/javascript for .js,
  video/mp4 for .mp4/.m4v, application/xml for .xml
- Added entries: .ijm (text/x-imagej-macro), .r (text/x-r-source),
  .ply (model/ply), .obj (model/obj)
- Removed malformed entries (.aff with no MIME type, .aif -> .aff)

Claude performed the deduplication via the following Python script:

    #!/usr/bin/env python3
    """
    Deduplicate mime-types.txt by keeping the best MIME type for each extension.
    Dropped alternatives are commented out in-place so they remain visible.

    Preference order (lower score = better):
      0 - image/*   non-x- subtype (e.g. image/tiff)
      1 - other     non-x- subtype (e.g. audio/aiff, video/mp4)
      2 - image/x-* subtype        (e.g. image/x-tiff)
      3 - other     x-   subtype   (e.g. audio/x-aiff)
      4 - application/octet-stream (generic fallback, avoid)
    """

    import sys
    from collections import defaultdict

    INPUT = "src/main/resources/org/scijava/desktop/mime-types.txt"

    def preference(mime):
        if mime == "application/octet-stream":
            return (4, mime)
        category, _, subtype = mime.partition("/")
        x_prefixed = subtype.startswith("x-")
        if category == "image" and not x_prefixed:
            return (0, mime)
        if not x_prefixed:
            return (1, mime)
        if category == "image":
            return (2, mime)
        return (3, mime)

    leading = []              # comment/blank lines before first data line
    best = {}                 # ext -> winning mime
    dropped = defaultdict(list)  # ext -> [dropped mimes in order encountered]
    ext_order = []            # insertion order of extensions
    in_data = False

    with open(INPUT) as f:
        for line in f:
            stripped = line.rstrip("\n")
            if not in_data and (stripped.startswith("#") or not stripped):
                leading.append(stripped)
                continue
            in_data = True
            if stripped.startswith("#") or not stripped:
                # comments interspersed in data section — preserve as leading context
                leading.append(stripped)
                continue
            parts = stripped.split("\t", 1)
            if len(parts) < 2:
                print(f"  skipping (no MIME type): {stripped!r}", file=sys.stderr)
                continue
            ext, mime = parts
            if ext not in best:
                best[ext] = mime
                ext_order.append(ext)
            elif preference(mime) < preference(best[ext]):
                print(f"  {ext}: preferring {mime!r} over {best[ext]!r}", file=sys.stderr)
                dropped[ext].append(best[ext])
                best[ext] = mime
            else:
                print(f"  {ext}: keeping {best[ext]!r}, dropping {mime!r}", file=sys.stderr)
                dropped[ext].append(mime)

    with open(INPUT, "w") as f:
        for line in leading:
            f.write(line + "\n")
        for ext in ext_order:
            f.write(f"{ext}\t{best[ext]}\n")
            for alt in dropped[ext]:
                f.write(f"#{ext}\t{alt}\n")

    print(f"\n{len(ext_order)} unique extensions written.", file=sys.stderr)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: ctrueden