r_dataset_details.Rmd

---
title: "Details of datasets with inapplicable data"
author: "Martin R. Smith"
---

```{R setup, echo=FALSE, message=FALSE}
library(TreeSearch)
citations <- TreeSearch::inapplicable.citations
phyData <- TreeSearch::inapplicable.phyData
gitRoot <- "https://raw.githubusercontent.com/ms609/ExploreInapplicable/master/"
studies <- c(matrices = "Manually revised matrices",
             inapplicable = "Inapplicable",
             extraState = "Extra State",
             ambigAbsent = "Ambiguous")
datasets <- lapply(names(citations), function (name) {
  ReadCharacters(file=paste0('matrices/', name, '.nex'))
})
```

Nexus files were constructed for each analysis using a [Perl script](https://github.com/ms609/ExploreInapplicable/blob/master/analyse_matrices.pl) and can be browsed on GitHub: `r paste0('[', studies, '](https://github.com/ms609/ExploreInapplicable/tree/master/', names(studies), ')', collapse=' | ')`

```{r datasets, echo=FALSE, results="asis"}
CharacterType <- function (datasetName) {
  vapply(readLines(paste0("charType/", datasetName, ".txt")),
         substr, character(1), 1, 1, USE.NAMES=FALSE)
}
CharacterNotes <- function(datasetName) {
  #notes <- vapply(readLines(paste0("charType/", datasetName, ".txt")), function (line) {
  #  substr(line, 3, nchar(line))
  #}, character(1), USE.NAMES=FALSE)
  notes <- readLines(paste0("charType/", datasetName, ".txt"))
  if (datasetName == 'Aria2015') {
    notes[23]
  } else if (datasetName == 'Liljeblad2008') {
    notes[42]
  } else {
    notes[grepl('\\]\\s*\\S+', notes, perl=TRUE)]
  }
}

AboutDataset <- function (i) {
  dataset <- phyData[[i]]
  datasetName <- names(phyData)[i]
  characterType <- CharacterType(datasetName)
  nTaxa <- length(dataset)
  nChar <- sum(attr(dataset, 'weight'))
  inapps <- apply(datasets[[i]], 2, function (x) any(x == '-'))
  cat(sprintf("\n\n### %s\n\n<!--Citation:-->%s\n\n - %s taxa, %s characters, of which %s contain inapplicable codings.\n\n - Download: %s\n\n - Inapplicable characters:\n\n   - Transformational: %s\n\n   - Neomorphic: %s\n\n%s\n\n",
          datasetName, citations[i], 
          nTaxa, nChar, sum(characterType %in% c('N', 'I', 'T') & inapps),
          paste0('[', studies, '](', gitRoot, names(studies), '/', 
                 datasetName, '.nex)', collapse=' | '),
          paste(which(characterType == 'T' & inapps), collapse=', '),
          paste(paste(which(characterType == 'N' & inapps), collapse=', '),
                if (any(characterType == 'I' & inapps)) '; with state 1 denoting absence:'
                else '',
                paste(which(characterType == 'I' & inapps), collapse=', '), collapse=''),
          if (length(CharacterNotes(datasetName)) > 0) {
            paste(c(' - Details:', CharacterNotes(datasetName)), collapse="\n\n   - ")
          } else ''
          ))
  0
}
oi <- vapply(seq_along(citations), AboutDataset, 0)
invisible()
```