Tidy tables of a text hypergraph

Usage

# S3 method for class 'text_hypergraph'
hg_get(x, what = c("weights", "documents", "vocabulary", "sentences"), ...)

Arguments

x: A text_hypergraph() object.
what: Which table: "weights" (default) – for the bag construction one row per document-word pair (doc, word, count, weight); for the window construction one row per window-content/word membership (edge, word, weight = window count); for the sentence construction one row per sentence/word membership (edge, word, weight = in-sentence count); for the knn construction one row per hyperedge membership (doc, edge, weight = cosine similarity). "sentences" (sentence construction only) gives one row per sentence hyperedge (edge, doc, sentence, n_tokens). "documents" gives one row per document (with n_tokens/n_types for token-based constructions, plus any metadata columns carried from the input). "vocabulary" gives one row per word (word, count, doc_freq, and idf under tf-idf weighting; empty for the knn construction, which has no token layer).
...: Unused.

Value

A base data.frame as described under what.

Examples

hg <- text_hypergraph(c(a = "salt and soup", b = "soup and stars"))
hg_get(hg)
#>   doc  word count weight
#> 1   a   and     1      1
#> 2   a  salt     1      1
#> 3   a  soup     1      1
#> 4   b   and     1      1
#> 5   b  soup     1      1
#> 6   b stars     1      1
hg_get(hg, what = "documents")
#>   doc n_tokens n_types
#> 1   a        3       3
#> 2   b        3       3