FEP-eb48: Hashtags

avidseeker · July 16, 2024, 11:37pm

Standardized method for identifying and displaying hashtags in posts across the Fediverse.

silverpill · July 18, 2024, 5:27pm

It is followed by one or more alphanumeric characters (letters A-Z, a-z, or digits 0-9).

What about non-ascii alphanums? Many Fediverse projects support unicode hashtags too.

avidseeker · July 18, 2024, 6:12pm

Of course they should be included.

Just to credit @neustradamus, the inspiration for this comes from this Mastodon post.

argrath · August 9, 2024, 2:22pm

Mastodon:

github.com

mastodon/mastodon/blob/6ca731e9b69962a0023df2c5314562293bb4b6b6/app/models/tag.rb#L33


      
          include Paginable
          # rubocop:disable Rails/HasAndBelongsToMany
          has_and_belongs_to_many :statuses
          has_and_belongs_to_many :accounts
          # rubocop:enable Rails/HasAndBelongsToMany
          
          has_many :passive_relationships, class_name: 'TagFollow', inverse_of: :tag, dependent: :destroy
          has_many :featured_tags, dependent: :destroy, inverse_of: :tag
          has_many :followers, through: :passive_relationships, source: :account
          
          HASHTAG_SEPARATORS = "_\u00B7\u30FB\u200c"
          HASHTAG_FIRST_SEQUENCE_CHUNK_ONE = "[[:word:]_][[:word:]#{HASHTAG_SEPARATORS}]*[[:alpha:]#{HASHTAG_SEPARATORS}]"
          HASHTAG_FIRST_SEQUENCE_CHUNK_TWO = "[[:word:]#{HASHTAG_SEPARATORS}]*[[:word:]_]"
          HASHTAG_FIRST_SEQUENCE = "(#{HASHTAG_FIRST_SEQUENCE_CHUNK_ONE}#{HASHTAG_FIRST_SEQUENCE_CHUNK_TWO})"
          HASHTAG_LAST_SEQUENCE = '([[:word:]_]*[[:alpha:]][[:word:]_]*)'
          HASHTAG_NAME_PAT = "#{HASHTAG_FIRST_SEQUENCE}|#{HASHTAG_LAST_SEQUENCE}"
          
          HASHTAG_RE = %r{(?<![=/)\p{Alnum}])#(#{HASHTAG_NAME_PAT})}
          HASHTAG_NAME_RE = /\A(#{HASHTAG_NAME_PAT})\z/i
          HASHTAG_INVALID_CHARS_RE = /[^[:alnum:]\u0E47-\u0E4E#{HASHTAG_SEPARATORS}]/

Misskey:

github.com

misskey-dev/mfm.js/blob/9cd9aaebe4db49b80aae72b03ea3fa2018f5face/src/internal/parser.ts#L617


      
          			}
          			// generate a text if mention is invalid
          			if (invalidMention) {
          				return P.success(resultIndex, input.slice(index, resultIndex));
          			}
          			const acct = modifiedHost != null ? `@${modifiedName}@${modifiedHost}` : `@${modifiedName}`;
          			return P.success(index + acct.length, M.MENTION(modifiedName, modifiedHost, acct));
          		});
          	},
          
          	hashtag: () => {
          		const mark = P.str('#');
          		const hashTagChar = P.seq(
          			P.notMatch(P.alt([P.regexp(/[ \u3000\t.,!?'"#:/[\]【】()「」（）<>]/), space, newLine])),
          			P.char,
          		).select(1);
          		const innerItem: P.Parser<unknown> = P.lazy(() => P.alt([
          			P.seq(
          				P.str('('), nest(innerItem, hashTagChar).many(0), P.str(')'),
          			),
          			P.seq(

FYI.