add tidycjk and tidyEmoji

Start by comparing how existing CRAN packages are listed in the Natural Language Processing task view, then review the tidycjk and tidyEmoji CRAN pages linked in the issue. Done means both packages appear in the view with accurate, concise descriptions and links.

Two packages that may fit this view, both tidy-style toolkits for text that whitespace-based tooling handles poorly.

tidycjk (0.1.0), for Chinese, Japanese and Korean. CJK writing does not separate words with whitespace, so a standard tokeniser returns either one undifferentiated blob per sentence or isolated characters. It provides cjk_segment() and cjk_tokens() with a pluggable segmenter backend (cjk_segmenters(), register_cjk_segmenter()), script and language identification (cjk_script(), cjk_blocks(), cjk_detect_language(), has_cjk(), cjk_ratio()), counts and summaries (cjk_char_counts(), cjk_summary()), and the width handling that matters whenever CJK text is aligned or truncated (cjk_width(), cjk_pad(), cjk_truncate(), to_fullwidth(), to_halfwidth()).

tidyEmoji (0.4.0), for emoji in text columns. Not every code point is an emoji and a single emoji can span several code points, so emoji statistics are fiddly to get right. It covers extraction (emoji_extract_unnest(), emoji_extract_nest(), emoji_tokens()), counting and density (emoji_frequency(), emoji_density(), emoji_ratio(), emoji_position()), categorisation (emoji_categorize()), sentiment and emotion scoring against published lexicons (emoji_sentiment(), emoji_emotion(), emoji_lexicons(), register_emoji_lexicon()), shortcode conversion both ways for preprocessing and accessibility (as_emoji_shortcode(), as_emoji_name(), emoji_to_text(), text_to_emoji(), emoji_sanitize()), context and collocation (emoji_context(), emoji_collocations(), emoji_ngrams(), emoji_cooccurrence()), and emoji_dfm() for document-by-emoji feature tables. emoji_token_cost() reports what emoji cost under a tokeniser, which matters for LLM preprocessing.

The sentiment lexicon is the Emoji Sentiment Ranking (Kralj Novak et al. 2015, doi:10.1371/journal.pone.0144296).

Disclosure: I am the author of both packages.