Skip to contents

Convert R objects to JSON-compatible format for MCP

Source: R/type-conversion-to-json.R
to_mcp_json.Rd

Converts various R objects to a JSON-compatible format, preserving type information where possible. Handles special types like dates, factors, matrices, and special numeric values (Inf, -Inf, NaN).

Usage

to_mcp_json(
  x,
  auto_unbox = TRUE,
  size_limit = 1e+06,
  custom_serializers = list()
)

Arguments

x

An R object to convert

auto_unbox

Whether to automatically unbox single-element vectors

size_limit

Maximum object size in bytes before using large object handling (default: 1MB)

custom_serializers

List of custom serializers for specific classes

Value

A JSON-compatible representation of the R object

Details

The function handles the following R types:

  • Basic types: NULL, logical, numeric, character, integer

  • Special numeric values: Inf, -Inf, NaN

  • Date/time types: Date, POSIXct, POSIXlt

  • Complex numbers

  • Raw vectors (binary data)

  • Factors (with levels preserved)

  • Matrices and arrays (with dimensions)

  • Data frames

  • Lists (recursive conversion)

  • S3 and S4 objects

  • Formulas and language objects

  • Environments (replaced with markers)

Examples

# Basic types
to_mcp_json(list(a = 1, b = "hello"))
#> $a
#> [x] 1
#> 
#> $b
#> [x] "hello"
#> 
to_mcp_json(c(TRUE, FALSE, NA))
#> [1]  TRUE FALSE    NA

# Special numeric values
to_mcp_json(c(1, Inf, -Inf, NaN))
#> $values
#> [1] "1"    "Inf"  "-Inf" "NaN" 
#> 
#> $special_indices
#> [1] 2 3 4
#> 
#> $`_mcp_type`
#> [1] "numeric_vector_special"
#> 

# Dates and times
to_mcp_json(Sys.Date())
#> $values
#> [1] "2025-06-20"
#> 
#> $`_mcp_type`
#> [1] "Date"
#> 
to_mcp_json(Sys.time())
#> $values
#> [1] "2025-06-20T17:54:51"
#> 
#> $timezone
#> [1] "UTC"
#> 
#> $`_mcp_type`
#> [1] "POSIXct"
#> 

# Data frames
to_mcp_json(data.frame(x = 1:3, y = letters[1:3]))
#> $x
#> [1] 1 2 3
#> 
#> $y
#> [1] "a" "b" "c"
#> 
#> attr(,"_mcp_type")
#> [1] "data.frame"
#> attr(,"_mcp_nrow")
#> [1] 3

# Complex types
to_mcp_json(matrix(1:6, nrow = 2))
#> $data
#> [1] 1 2 3 4 5 6
#> 
#> $dim
#> [1] 2 3
#> 
#> $dimnames
#> NULL
#> 
#> $`_mcp_type`
#> [1] "matrix"
#> 
to_mcp_json(factor(c("a", "b", "a")))
#> $levels
#> [1] "a" "b"
#> 
#> $values
#> [1] 1 2 1
#> 
#> $`_mcp_type`
#> [1] "factor"
#> 
to_mcp_json(3 + 4i)
#> $real
#> [1] 3
#> 
#> $imaginary
#> [1] 4
#> 
#> $`_mcp_type`
#> [1] "complex"
#>