Search specified transcripts in getMatches()

[djvill]

getMatches() supports searching specified participants or transcript types. It'd be useful, especially for the purposes of testing queries, to add an id argument that takes a vector of transcript IDs to search.

As a further extension, getMatches() could throw an error if the user's specification of both participant.ids and id results in a null set of transcripts.

[robertfromont]

The underlying API actually already supports specifying an expression for identifying specific transcripts. Both transcripts and participants are identified using the same type of expressions as used with getMatchingTranscriptIds and getMatchingParticipantIds.

So in the next version (1.3-0) the getMatches function will now have the following parameters:

getMatches <- function(labbcat.url, pattern, 
  participant.expression=NULL, ## query expression for identifying participants to search the utterances of
  transcript.expression=NULL,  ## query expression for identifying transcripts to search in
  main.participant=TRUE, aligned=FALSE, matches.per.transcript=NULL, words.context=0, 
  max.matches=NULL, overlap.threshold=NULL, anchor.confidence.min=50, page.length=1000, no.progress=FALSE)

There will also be some new functions for easily generating expressions. So achieving the current functionality involves something like:

results <- getMatches(labbcat.url, pattern,
  participant.expression = expressionFromIds(c("AP511_MikeThorpe","BR2044_OllyOhlson")),
  transcript.expression = expressionFromTranscriptTypes(c("interview","monologue")))

...and the kind of thing you're after would be:

results <- getMatches(labbcat.url, pattern,
  transcript.expression = expressionFromIds(c("AP511_MikeThorpe.eaf","BR2044_OllyOhlson.eaf")))

...and you'll also be able to do things that don't involve knowing the specific IDs:

results <- getMatches(labbcat.url, pattern,
  participant.expression = paste(
    expressionFromAttributeValue("participant_gender", "M"),
    expressionFromAttributeValues("participant_languages_spoken", "en"),
    sep=" && "),
  transcript.expression = expressionFromTranscriptTypes("wordlist", not=TRUE))