ProvStoreDF: ProvStoreDF constructors and class

In gmbecker/roprov: Low-Level Support for Provenance Capture Between in-Memory R Objects

Description

Functions to create a valid ProvStore object from raw data about evaluation history (including hashes of input and output variable values).

Usage

makeProvStore(invarhashes = NULL, outvarhashes = NULL, code = NULL,
  codehash = fastdigest(code), invars = names(invarhashes),
  outvars = names(outvarhashes), invarclasses = character(),
  outvarclasses = character(), agent = getUser(), hashprefix = "SpookyV2")

ProvStoreDF(outvars = character(), outvarhashes = character(),
  outvarclasses = character(), invars = character(),
  invarhashes = character(), invarclasses = character(),
  agent = character(), code = character(), codehash = vapply(code,
  fastdigest, character(1)), hashprefix = "SpookyV2", df = NULL)

Arguments

invarhashes	character. Hashes of values for input variables (or NULL).
outvarhashes	character. Hashes of values for output variables (or NULL).
code	character. The code associated with the input and output variables in question.
codehash	character. The hash of the (parsed and deparsed) code.
invars	character. The set of input variables (symbols) to code
outvars	character. The output variable(s) (symbols) to code.
invarclasses	character. The (top level) classes of the input variable values.
outvarclasses	character. The (top level) classes of the output variable values.
agent	character. a string identifying the user. Defaults to the output of a whoami system call.
hashprefix	character. A prefix to append to the hashes if not present already to make them more self-describing.
df	data.frame. Optional. An already constructed data.frame do use as the provdata of the constructed ProvStoreDF object.

Details

makeProvStore expects information about a single code unit. i.e., all input values will be counted as inputs for all outputs. For makeProvStore the number of inputs and outputs need not match, and the appropriate replication on inputs will happen automatically.

ProvStoreDF is a direct constructor and thus expects the replciation of inputs for multiple outputs to have already occurred.

Examples

## spoof the information needed to create a provstore
library(fastdigest)
code = c("x = 5", "y = x + 1")
outvars = c("x", "y")
outvarhashes = c(fastdigest(5), fastdigest(6))
outvarclasses = rep("numeric", 2)
invars = c("", "x")
invarhashes = c("", fastdigest(5))
invarclasses = c("", "numeric")

ps = ProvStoreDF(outvars = outvars,
                 outvarhashes = outvarhashes,
                 outvarclasses = outvarclasses,
                 invars = invars,
                 invarhashes = invarhashes,
                 invarclasses = invarclasses,
                 code = code,
                 agent = "coolguyorgirl")

 df = data.frame(outputvar = outvars,
                 outputvarhash = outvarhashes,
                 outputvarclass = outvarclasses,
                 inputvar= invars,
                 inputvarhash = invarhashes,
                 inputvarclass = invarclasses,
                 agent = "coolgirloruy",
                 code = code,
                 codehash = sapply(code, fastdigest),
                 stringsAsFactors = FALSE)

ps2 = ProvStoreDF(df = df)

gmbecker/roprov documentation built on May 27, 2019, 11:49 p.m.