A partial implementation of a R expression -> latex math converter

latex-math.r

# User facing function
#
# to_math(x_1 + 1^{2 + 4} + 5 + sqrt(y) / 5 %/% 10)
# to_math(paste(x^2, y - 1, z_i))
# to_math(hat(tilde(ring(x))))
to_math <- function(x) {
  to_math_q(substitute(x))
}

to_math_q <- function(x) {
  if (is.integer(x) || is.numeric(x)) return(x)
  if (is.name(x)) {
    x2 <- as.character(x)
    x3 <- if (x2 %in% names(symbols)) symbols[[x2]] else x2
    return(x3)
  }

  eval(x, lenv)
}

lenv <- new.env(parent = emptyenv())

dots <- function(...) {
  eval(substitute(alist(...)))
}

# Convert a function into a fexpr: the function recieves only unevaluated args
fexpr <- function(f) {
  stopifnot(is.function(f))
  function(...) {
    args <- dots(...)
    do.call(f, args, quote = TRUE)
  }
}

# Helper functions
unary_op <- function(left, right) {
  fexpr(function(e1, e2) {
    paste0(left, to_math_q(e1), right)
  })
}
binary_op <- function(sep) {
  fexpr(function(e1, e2) {
    paste0(to_math_q(e1), " ", sep, " ", to_math_q(e2))
  })
}

# Binary operators
lenv$"+" <- binary_op("+")
lenv$"-" <- binary_op("-")
lenv$"*" <- binary_op("*")
lenv$"/" <- binary_op("/")
lenv$"%+-%" <- binary_op("\\pm")
lenv$"%/%" <- binary_op("\\")
lenv$"%*%" <- binary_op("\\times")
lenv$"%.%" <- binary_op("\\cdot")
lenv$"[" <- binary_op("_")
lenv$"^" <- binary_op("^")

# Grouping
lenv$"{" <- unary_op("{", "}")
lenv$"(" <- unary_op("(", ")")
lenv$paste <- fexpr(function(...) {
  paste0(unlist(lapply(list(...), to_math_q)), collapse = " ")
})

# Other math functions
lenv$sqrt <- unary_op("\\sqrt{", "}")
lenv$log <- unary_op("\\log{", "}")
lenv$inf <- unary_op("\\inf{", "}")
lenv$sup <- unary_op("\\sup{", "}")
lenv$frac <- fexpr(function(a, b) {
  paste0("\\frac{", to_math_q(a), "}{", to_math_q(b), "}")
})

# Labelling
lenv$hat <- unary_op("\\hat{", "}")
lenv$tilde <- unary_op("\\tilde{", "}")
lenv$dot <- unary_op("\\dot{", "}")
lenv$ring <- unary_op("\\ring{", "}")

Very cool! Could something like this get implemented in ROxygen in place of having to write out LaTeX and plain-txt version with \deqn{}{}? I suppose one could also consider the reverse parser (LaTeX to R), but that might be asking for trouble..

There is a bug in the function, as it does not evaluate negative numbers: to_math(-1), to_math_q(expression(-1)), to_math_q(expression(x^-x)). These lines result in error.
Is it possible to fix this bug?

We can fix the bug in the function, to evaluate negative numbers, by:

binary_op <- function(sep) {
force(sep)
function(e1, e2) {
if(missing(e2)){ paste0(" - ", e1) }else{ paste0(e1,sep,e2)}
}
}