Converts an R model object into a parsed model — parse

Parses a fitted R model's structure and extracts the components needed to create a dplyr formula for prediction. The parsed model can be serialized (e.g., saved to YAML) and later used to generate predictions without the original model object.

Usage

parse_model(model)

Arguments

model: An R model object.

Value

A parsed model object with class parsed_model and a model-specific subclass (e.g., pm_xgb, pm_tree, pm_regression). The object contains:

$general: List with model metadata including model (model type), type (used for S3 dispatch), version (parsed model format version), and model-specific parameters.
Model-specific fields containing coefficients, tree structures, etc.

Parsed model versions

The $general$version field indicates the parsed model format:

Version 1: Original format. Linear models store coefficients in a data frame. Tree models use flat case_when() expressions where all leaf conditions are at the same level.
Version 2: Improved coefficient storage for linear models (lm, earth). Tree models still use flat case_when().
Version 3: Current format. Tree models (rpart, ranger, randomForest, xgboost, lightgbm, catboost, partykit, cubist) use nested case_when() expressions that mirror the tree structure. This produces more efficient SQL and R code because conditions are evaluated hierarchically rather than checking all leaf paths.

When loading a parsed model saved with an older version, tidypredict automatically uses the appropriate formula builder for backwards compatibility.

Model types

Each parsed model has a type that determines the S3 class used for dispatch:

pm_regression: Linear models (lm, glm, earth, glmnet)
pm_tree: Single trees and forests (rpart, partykit, ranger, randomForest, cubist)
pm_xgb: XGBoost gradient boosting models
pm_lgb: LightGBM gradient boosting models
pm_catboost: CatBoost gradient boosting models

Examples

library(dplyr)
#> 
#> Attaching package: ‘dplyr’
#> The following objects are masked from ‘package:stats’:
#> 
#>     filter, lag
#> The following objects are masked from ‘package:base’:
#> 
#>     intersect, setdiff, setequal, union
df <- mutate(mtcars, cyl = paste0("cyl", cyl))
model <- lm(mpg ~ wt + cyl * disp, offset = am, data = df)
parse_model(model)
#> $general
#> $general$model
#> [1] "lm"
#> 
#> $general$version
#> [1] 2
#> 
#> $general$type
#> [1] "regression"
#> 
#> $general$residual
#> [1] 25
#> 
#> $general$sigma2
#> [1] 5.414866
#> 
#> $general$offset
#> am
#> 
#> $general$is_glm
#> [1] 0
#> 
#> 
#> $terms
#> $terms[[1]]
#> $terms[[1]]$label
#> [1] "(Intercept)"
#> 
#> $terms[[1]]$coef
#> [1] 39.41278
#> 
#> $terms[[1]]$is_intercept
#> [1] 1
#> 
#> $terms[[1]]$fields
#> $terms[[1]]$fields[[1]]
#> $terms[[1]]$fields[[1]]$type
#> [1] "ordinary"
#> 
#> $terms[[1]]$fields[[1]]$col
#> [1] "(Intercept)"
#> 
#> 
#> 
#> $terms[[1]]$qr
#> $terms[[1]]$qr$qr_1
#> [1] -0.1767767
#> 
#> $terms[[1]]$qr$qr_2
#> [1] -0.5905573
#> 
#> $terms[[1]]$qr$qr_3
#> [1] -0.1262157
#> 
#> $terms[[1]]$qr$qr_4
#> [1] 0.3862155
#> 
#> $terms[[1]]$qr$qr_5
#> [1] -0.1536563
#> 
#> $terms[[1]]$qr$qr_6
#> [1] -0.02235892
#> 
#> $terms[[1]]$qr$qr_7
#> [1] 1.040622
#> 
#> 
#> 
#> $terms[[2]]
#> $terms[[2]]$label
#> [1] "wt"
#> 
#> $terms[[2]]$coef
#> [1] -1.619173
#> 
#> $terms[[2]]$is_intercept
#> [1] 0
#> 
#> $terms[[2]]$fields
#> $terms[[2]]$fields[[1]]
#> $terms[[2]]$fields[[1]]$type
#> [1] "ordinary"
#> 
#> $terms[[2]]$fields[[1]]$col
#> [1] "wt"
#> 
#> 
#> 
#> $terms[[2]]$qr
#> $terms[[2]]$qr$qr_1
#> [1] 0
#> 
#> $terms[[2]]$qr$qr_2
#> [1] 0.1835596
#> 
#> $terms[[2]]$qr$qr_3
#> [1] 0.01011187
#> 
#> $terms[[2]]$qr$qr_4
#> [1] -0.2305162
#> 
#> $terms[[2]]$qr$qr_5
#> [1] 0.3054412
#> 
#> $terms[[2]]$qr$qr_6
#> [1] -0.09551174
#> 
#> $terms[[2]]$qr$qr_7
#> [1] 0.156527
#> 
#> 
#> 
#> $terms[[3]]
#> $terms[[3]]$label
#> [1] "cylcyl6"
#> 
#> $terms[[3]]$coef
#> [1] -18.41701
#> 
#> $terms[[3]]$is_intercept
#> [1] 0
#> 
#> $terms[[3]]$fields
#> $terms[[3]]$fields[[1]]
#> $terms[[3]]$fields[[1]]$type
#> [1] "conditional"
#> 
#> $terms[[3]]$fields[[1]]$col
#> [1] "cyl"
#> 
#> $terms[[3]]$fields[[1]]$val
#> [1] "cyl6"
#> 
#> $terms[[3]]$fields[[1]]$op
#> [1] "equal"
#> 
#> 
#> 
#> $terms[[3]]$qr
#> $terms[[3]]$qr$qr_1
#> [1] 0
#> 
#> $terms[[3]]$qr$qr_2
#> [1] 0
#> 
#> $terms[[3]]$qr$qr_3
#> [1] 0.4282663
#> 
#> $terms[[3]]$qr$qr_4
#> [1] 0.3323365
#> 
#> $terms[[3]]$qr$qr_5
#> [1] 0.1509332
#> 
#> $terms[[3]]$qr$qr_6
#> [1] 1.88383
#> 
#> $terms[[3]]$qr$qr_7
#> [1] -1.412292
#> 
#> 
#> 
#> $terms[[4]]
#> $terms[[4]]$label
#> [1] "cylcyl8"
#> 
#> $terms[[4]]$coef
#> [1] -16.20665
#> 
#> $terms[[4]]$is_intercept
#> [1] 0
#> 
#> $terms[[4]]$fields
#> $terms[[4]]$fields[[1]]
#> $terms[[4]]$fields[[1]]$type
#> [1] "conditional"
#> 
#> $terms[[4]]$fields[[1]]$col
#> [1] "cyl"
#> 
#> $terms[[4]]$fields[[1]]$val
#> [1] "cyl8"
#> 
#> $terms[[4]]$fields[[1]]$op
#> [1] "equal"
#> 
#> 
#> 
#> $terms[[4]]$qr
#> $terms[[4]]$qr$qr_1
#> [1] 0
#> 
#> $terms[[4]]$qr$qr_2
#> [1] 0
#> 
#> $terms[[4]]$qr$qr_3
#> [1] 0
#> 
#> $terms[[4]]$qr$qr_4
#> [1] 0.6462039
#> 
#> $terms[[4]]$qr$qr_5
#> [1] 0.7608288
#> 
#> $terms[[4]]$qr$qr_6
#> [1] -0.4039674
#> 
#> $terms[[4]]$qr$qr_7
#> [1] -1.645469
#> 
#> 
#> 
#> $terms[[5]]
#> $terms[[5]]$label
#> [1] "disp"
#> 
#> $terms[[5]]$coef
#> [1] -0.09297857
#> 
#> $terms[[5]]$is_intercept
#> [1] 0
#> 
#> $terms[[5]]$fields
#> $terms[[5]]$fields[[1]]
#> $terms[[5]]$fields[[1]]$type
#> [1] "ordinary"
#> 
#> $terms[[5]]$fields[[1]]$col
#> [1] "disp"
#> 
#> 
#> 
#> $terms[[5]]$qr
#> $terms[[5]]$qr$qr_1
#> [1] 0
#> 
#> $terms[[5]]$qr$qr_2
#> [1] 0
#> 
#> $terms[[5]]$qr$qr_3
#> [1] 0
#> 
#> $terms[[5]]$qr$qr_4
#> [1] 0
#> 
#> $terms[[5]]$qr$qr_5
#> [1] -0.005178979
#> 
#> $terms[[5]]$qr$qr_6
#> [1] 0.002289148
#> 
#> $terms[[5]]$qr$qr_7
#> [1] -0.01330082
#> 
#> 
#> 
#> $terms[[6]]
#> $terms[[6]]$label
#> [1] "cylcyl6:disp"
#> 
#> $terms[[6]]$coef
#> [1] 0.1113389
#> 
#> $terms[[6]]$is_intercept
#> [1] 0
#> 
#> $terms[[6]]$fields
#> $terms[[6]]$fields[[1]]
#> $terms[[6]]$fields[[1]]$type
#> [1] "conditional"
#> 
#> $terms[[6]]$fields[[1]]$col
#> [1] "cyl"
#> 
#> $terms[[6]]$fields[[1]]$val
#> [1] "cyl6"
#> 
#> $terms[[6]]$fields[[1]]$op
#> [1] "equal"
#> 
#> 
#> $terms[[6]]$fields[[2]]
#> $terms[[6]]$fields[[2]]$type
#> [1] "ordinary"
#> 
#> $terms[[6]]$fields[[2]]$col
#> [1] "disp"
#> 
#> 
#> 
#> $terms[[6]]$qr
#> $terms[[6]]$qr$qr_1
#> [1] 0
#> 
#> $terms[[6]]$qr$qr_2
#> [1] 0
#> 
#> $terms[[6]]$qr$qr_3
#> [1] 0
#> 
#> $terms[[6]]$qr$qr_4
#> [1] 0
#> 
#> $terms[[6]]$qr$qr_5
#> [1] 0
#> 
#> $terms[[6]]$qr$qr_6
#> [1] -0.01081956
#> 
#> $terms[[6]]$qr$qr_7
#> [1] 0.01266668
#> 
#> 
#> 
#> $terms[[7]]
#> $terms[[7]]$label
#> [1] "cylcyl8:disp"
#> 
#> $terms[[7]]$coef
#> [1] 0.08795571
#> 
#> $terms[[7]]$is_intercept
#> [1] 0
#> 
#> $terms[[7]]$fields
#> $terms[[7]]$fields[[1]]
#> $terms[[7]]$fields[[1]]$type
#> [1] "conditional"
#> 
#> $terms[[7]]$fields[[1]]$col
#> [1] "cyl"
#> 
#> $terms[[7]]$fields[[1]]$val
#> [1] "cyl8"
#> 
#> $terms[[7]]$fields[[1]]$op
#> [1] "equal"
#> 
#> 
#> $terms[[7]]$fields[[2]]
#> $terms[[7]]$fields[[2]]$type
#> [1] "ordinary"
#> 
#> $terms[[7]]$fields[[2]]$col
#> [1] "disp"
#> 
#> 
#> 
#> $terms[[7]]$qr
#> $terms[[7]]$qr$qr_1
#> [1] 0
#> 
#> $terms[[7]]$qr$qr_2
#> [1] 0
#> 
#> $terms[[7]]$qr$qr_3
#> [1] 0
#> 
#> $terms[[7]]$qr$qr_4
#> [1] 0
#> 
#> $terms[[7]]$qr$qr_5
#> [1] 0
#> 
#> $terms[[7]]$qr$qr_6
#> [1] 0
#> 
#> $terms[[7]]$qr$qr_7
#> [1] 0.01324096
#> 
#> 
#> 
#> 
#> attr(,"class")
#> [1] "parsed_model"  "pm_regression" "list"