tests/regression/example/config/postProcess.R

In dfalster/dataMashR: Mash data transparantly and reproducibly

## This is checking that is specific to the baad data set
postProcess <- function(data, lookupSpecies="none"){

  fillDerivedVariables <- function(x){

    x <- within(x, {
      c.d <- as.numeric(c.d)
      m.rt <- as.numeric(m.rt)
      a.cp <- as.numeric(a.cp)
      d.cr <- as.numeric(d.cr)

      # Missing leaf area when leaf mass and LMA are OK.
      ii <- is.na(a.lf) & !is.na(m.lf) & !is.na(ma.ilf)
      a.lf[ii] <- m.lf[ii] / ma.ilf[ii]

      # Height to crown base if tree height and crown depth are OK.
      ii <- is.na(h.c) & !is.na(c.d) & !is.na(h.t)
      h.c[ii] <- h.t[ii] - c.d[ii]

      # Crown depth if tree height and height to crown base are OK
      ii <- is.na(c.d) & !is.na(h.t) & !is.na(h.c)
      c.d[ii] <- h.t[ii] - h.c[ii]

      # Stem area  / DBH at breast height or base.
      ii <- is.na(a.stbh) & !is.na(d.bh)
      a.stbh[ii] <- (pi/4)*d.bh[ii]^2

      ii <- is.na(a.stba) & !is.na(d.ba)
      a.stba[ii] <- (pi/4)*d.ba[ii]^2

      # Again, the other way around.
      ii <- !is.na(a.stbh) & is.na(d.bh)
      d.bh[ii] <- sqrt(a.stbh[ii] / (pi/4))

      ii <- !is.na(a.stba) & is.na(d.ba)
      d.ba[ii] <- sqrt(a.stba[ii] / (pi/4))

      # Stem mass
      ii <- is.na(m.st) & !is.na(m.ss) & !is.na(m.sh) &
        !is.na(m.sb)
      m.st[ii] <- m.ss[ii] + m.sh[ii] + m.sb[ii]

      # total aboveground mass.
      ii <- is.na(m.so) & !is.na(m.lf) & !is.na(m.st)
      m.so[ii] <- m.lf[ii] + m.st[ii]

      # total root mass
      ii <- is.na(m.rt) & !is.na(m.rf) & !is.na(m.rc)
      m.rt[ii] <- m.rf[ii] + m.rc[ii]

      # Total mass.
      ii <- is.na(m.to) & !is.na(m.rt) & !is.na(m.so)
      m.to[ii] <- m.rt[ii] + m.so[ii]

      # crown width
      ii <- is.na(d.cr) & !is.na(a.cp)
      d.cr[ii] <- sqrt(a.cp[ii]/(pi/4))

      # crown area
      ii <- !is.na(d.cr) & is.na(a.cp)
      a.cp[ii] <- (pi/4)*d.cr[ii]^2

      # clean up
      ii <- NULL
    })

    x
  }

  # Fill missing derived variables (e.g. m.so= m.st+m.lf, etc.)
  fillDerivedVariables(data)
}