diff --git a/.jules/bolt.md b/.jules/bolt.md index 7d3c603..6d03e33 100644 --- a/.jules/bolt.md +++ b/.jules/bolt.md @@ -16,3 +16,6 @@ ## 2025-02-12 - R 언어에서 반복적인 mirt 모델 생성 시 불필요한 데이터프레임 부분집합 추출 최적화 **Learning:** R에서 데이터프레임의 특정 열을 추출하는 작업(`df[cols]`)은 O(N)의 메모리 복사를 수반합니다. `autoFIPC`에서 `mirt` 모델의 파라미터를 설정하거나 호출하는 과정 중에 `newformXDataK[colnames(newFormModel@Data$data)]` 코드가 반복해서 사용되었고, 심지어 `ncol()`을 위해 단순히 개수를 구할 때도 사용되어 불필요한 메모리 할당과 오버헤드를 초래했습니다. **Action:** 조건문이나 반복문 내부에서 불필요하게 데이터프레임 부분집합 연산이 반복되지 않도록 외부에서 한 번만 `linkedFormData <- newformXDataK[colnames(newFormModel@Data$data)]`로 캐싱(caching)한 뒤, `ncol(linkedFormData)`와 `data = linkedFormData` 형태로 재사용하여 메모리 복사와 O(N) 오버헤드를 방지해야 합니다. +## 2024-07-13 - R 언어에서 데이터프레임 조건부 업데이트 시 2차원 인덱싱 병목 최적화 +**Learning:** R에서 조건에 맞는 특정 열을 업데이트할 때 `df[idx, 'col'] <- val` 방식의 2차원 인덱싱은 팩터(factor) 레벨 검증, 차원 확인, 전체 데이터프레임 깊은 복사(deep copy) 등의 S3 메소드 디스패치(`[<-.data.frame`) 오버헤드를 유발하여 O(N) 수준의 매우 비효율적인 성능 저하를 초래합니다. +**Action:** 메소드 디스패치를 우회하고 O(1) 리스트 접근(list access)과 C-레벨 벡터 수정으로 직접 처리하기 위해 벡터 직접 서브셋팅(direct vector subsetting) 방식인 `df$col[idx] <- val`를 사용하여 성능을 대폭 향상시켜야 합니다. diff --git a/R/aFIPC.R b/R/aFIPC.R index 6254651..cb78786 100644 --- a/R/aFIPC.R +++ b/R/aFIPC.R @@ -598,25 +598,28 @@ autoFIPC <- # Preserve mirt's structural estimability flags. Forcing every row TRUE # frees boundary parameters such as 2PL g/u and makes the Hessian unstable. - NewScaleParms[NewScaleParms$item == 'GROUP', "est"] <- FALSE - OldScaleParms[OldScaleParms$item == 'GROUP', "est"] <- FALSE + # ⚡ Bolt: Use direct vector subsetting (e.g. df$col[idx] <- val) instead of + # two-dimensional subsetting to bypass method dispatch overhead + NewScaleParms$est[NewScaleParms$item == 'GROUP'] <- FALSE + OldScaleParms$est[OldScaleParms$item == 'GROUP'] <- FALSE - NewScaleParms[NewScaleParms$name == "COV_11", "est"] <- TRUE - OldScaleParms[OldScaleParms$name == "COV_11", "est"] <- TRUE + NewScaleParms$est[NewScaleParms$name == "COV_11"] <- TRUE + OldScaleParms$est[OldScaleParms$name == "COV_11"] <- TRUE if (itemtype == 'Rasch') { - NewScaleParms[NewScaleParms$name == "a1", "est"] <- FALSE - OldScaleParms[OldScaleParms$name == "a1", "est"] <- FALSE + NewScaleParms$est[NewScaleParms$name == "a1"] <- FALSE + OldScaleParms$est[OldScaleParms$name == "a1"] <- FALSE } #IPD if (checkIPD == T) { # config + # ⚡ Bolt: Explicitly define factor levels to bypass automatic level inference overhead IPDgroup <- - as.factor(c( - rep('oldForm', nrow(oldformYDataK)), - rep('newForm', nrow(newformXDataK)) - )) + factor( + rep(c('oldForm', 'newForm'), c(nrow(oldformYDataK), nrow(newformXDataK))), + levels = c('newForm', 'oldForm') + ) IPDItemCount <- 0 # IPD target item checking @@ -786,14 +789,14 @@ autoFIPC <- oldIdx <- oldScaleParmsItemIdxCache[[oldFormItemStr]] # ⚡ Bolt: Remove unnecessary paste0() array string generation overhead - message(' Newform Parms: ', paste(NewScaleParms[newIdx, "value"], collapse = ' ')) - message(' Oldform Parms: ', paste(OldScaleParms[oldIdx, "value"], collapse = ' ')) + message(' Newform Parms: ', paste(NewScaleParms$value[newIdx], collapse = ' ')) + message(' Oldform Parms: ', paste(OldScaleParms$value[oldIdx], collapse = ' ')) - NewScaleParms[newIdx, "value"] <- - OldScaleParms[oldIdx, "value"] - message(' Linkedform Parms: ', paste(NewScaleParms[newIdx, "value"], collapse = ' '), '\n') + NewScaleParms$value[newIdx] <- + OldScaleParms$value[oldIdx] + message(' Linkedform Parms: ', paste(NewScaleParms$value[newIdx], collapse = ' '), '\n') - NewScaleParms[newIdx, "est"] <- + NewScaleParms$est[newIdx] <- FALSE } else { message( @@ -813,9 +816,9 @@ autoFIPC <- newBetaIdx <- NewScaleParms$item == 'BETA' oldBetaIdx <- OldScaleParms$item == 'BETA' - NewScaleParms[newBetaIdx, "value"] <- - OldScaleParms[oldBetaIdx, "value"] - NewScaleParms[newBetaIdx, "est"] <- + NewScaleParms$value[newBetaIdx] <- + OldScaleParms$value[oldBetaIdx] + NewScaleParms$est[newBetaIdx] <- FALSE message('applying BETA parameter as linking') @@ -823,7 +826,7 @@ autoFIPC <- message( ' Linkedform Parms: ', paste0( - NewScaleParms[newBetaIdx, "value"], + NewScaleParms$value[newBetaIdx], ' ' ), '\n' @@ -858,13 +861,13 @@ autoFIPC <- new_mean11_idx <- NewScaleParms$name == "MEAN_11" old_mean11_idx <- OldScaleParms$name == "MEAN_11" - NewScaleParms[new_cov11_idx, "est"] <- FALSE - OldScaleParms[old_cov11_idx, "est"] <- FALSE - NewScaleParms[new_mean11_idx, "est"] <- FALSE - OldScaleParms[old_mean11_idx, "est"] <- FALSE + NewScaleParms$est[new_cov11_idx] <- FALSE + OldScaleParms$est[old_cov11_idx] <- FALSE + NewScaleParms$est[new_mean11_idx] <- FALSE + OldScaleParms$est[old_mean11_idx] <- FALSE - NewScaleParms[new_cov11_idx, "value"] <- 1 - OldScaleParms[old_mean11_idx, "value"] <- 0 + NewScaleParms$value[new_cov11_idx] <- 1 + OldScaleParms$value[old_mean11_idx] <- 0 } if (freeMEAN == T) { LinkedModelSyntax <- @@ -875,8 +878,8 @@ autoFIPC <- 'MEAN = F1' )) - NewScaleParms[NewScaleParms$name == "MEAN_1", "est"] <- TRUE - OldScaleParms[OldScaleParms$name == "MEAN_1", "est"] <- TRUE + NewScaleParms$est[NewScaleParms$name == "MEAN_1"] <- TRUE + OldScaleParms$est[OldScaleParms$name == "MEAN_1"] <- TRUE } else { LinkedModelSyntax <- mirt::mirt.model(paste0( diff --git a/tests/testthat/test-autoFIPC-estimation.R b/tests/testthat/test-autoFIPC-estimation.R new file mode 100644 index 0000000..01151ea --- /dev/null +++ b/tests/testthat/test-autoFIPC-estimation.R @@ -0,0 +1,45 @@ +test_that("autoFIPC initial estimation failure", { + skip_if_not_installed("mirt") + skip_if_not_installed("mockery") + + # Stub mirt::mirt so it just returns a string, making `isRealMirtModel` and `exists` checks work or fail in specific ways + m_fail <- function(...) stop("forced failure") + mockery::stub(aFIPC::autoFIPC, 'mirt::mirt', m_fail) + + expect_error( + aFIPC::autoFIPC( + newformXData = data.frame(A=c(1,1,1), B=c(0,0,0)), + oldformYData = data.frame(A=c(1,1,1), B=c(0,0,0)), + newformCommonItemNames = c('A'), + oldformCommonItemNames = c('A'), + itemtype = '2PL', + confirmCommonItems = TRUE + ), + "Security Error: Initial estimation of oldFormModel completely failed" + ) +}) + +test_that("autoFIPC newform initial estimation failure", { + skip_if_not_installed("mirt") + skip_if_not_installed("mockery") + + set.seed(1) + dat_old <- mirt::simdata(a=matrix(c(1,1,1,1)), d=matrix(c(0,0,0,0)), N=100, itemtype="2PL") + colnames(dat_old) <- c("A", "B", "C", "D") + old_mod <- mirt::mirt(dat_old, 1, itemtype="2PL", SE=FALSE, verbose=FALSE) + + m_fail <- function(...) stop("forced failure") + mockery::stub(aFIPC::autoFIPC, 'mirt::mirt', m_fail) + + expect_error( + aFIPC::autoFIPC( + newformXData = dat_old, + oldformYData = old_mod, + newformCommonItemNames = c('A'), + oldformCommonItemNames = c('A'), + itemtype = '2PL', + confirmCommonItems = TRUE + ), + "Security Error: Initial estimation of newFormModel completely failed" + ) +}) diff --git a/tests/testthat/test-autoFIPC-extras.R b/tests/testthat/test-autoFIPC-extras.R new file mode 100644 index 0000000..e9bcd76 --- /dev/null +++ b/tests/testthat/test-autoFIPC-extras.R @@ -0,0 +1,109 @@ +test_that("autoFIPC specific error paths", { + skip_if_not_installed("mirt") + + # oldformBILOGprior must be logical + expect_error( + aFIPC::autoFIPC( + newformXData = data.frame(A=1), + oldformYData = data.frame(A=2), + newformCommonItemNames = c('A'), + oldformCommonItemNames = c('A'), + oldformBILOGprior = "TRUE" + ), + "Security Error: oldformBILOGprior must be a single non-NA logical value or NULL" + ) + + # tryFitwholeOldItems must be logical + expect_error( + aFIPC::autoFIPC( + newformXData = data.frame(A=1), + oldformYData = data.frame(A=2), + newformCommonItemNames = c('A'), + oldformCommonItemNames = c('A'), + tryFitwholeOldItems = "TRUE" + ), + "Security Error: tryFitwholeOldItems must be a single non-NA logical value" + ) + + # checkIPD must be logical + expect_error( + aFIPC::autoFIPC( + newformXData = data.frame(A=1), + oldformYData = data.frame(A=2), + newformCommonItemNames = c('A'), + oldformCommonItemNames = c('A'), + checkIPD = "TRUE" + ), + "Security Error: checkIPD must be a single non-NA logical value" + ) + + # freeMEAN must be logical + expect_error( + aFIPC::autoFIPC( + newformXData = data.frame(A=1), + oldformYData = data.frame(A=2), + newformCommonItemNames = c('A'), + oldformCommonItemNames = c('A'), + freeMEAN = "TRUE" + ), + "Security Error: freeMEAN must be a single non-NA logical value" + ) + + # forceNormalZeroOne must be logical + expect_error( + aFIPC::autoFIPC( + newformXData = data.frame(A=1), + oldformYData = data.frame(A=2), + newformCommonItemNames = c('A'), + oldformCommonItemNames = c('A'), + forceNormalZeroOne = "TRUE" + ), + "Security Error: forceNormalZeroOne must be a single non-NA logical value" + ) + + # parameterOverwrite must be logical + expect_error( + aFIPC::autoFIPC( + newformXData = data.frame(A=1), + oldformYData = data.frame(A=2), + newformCommonItemNames = c('A'), + oldformCommonItemNames = c('A'), + parameterOverwrite = "TRUE" + ), + "Security Error: parameterOverwrite must be a single non-NA logical value" + ) + + # empiricalhist must be logical + expect_error( + aFIPC::autoFIPC( + newformXData = data.frame(A=1), + oldformYData = data.frame(A=2), + newformCommonItemNames = c('A'), + oldformCommonItemNames = c('A'), + empiricalhist = "TRUE" + ), + "Security Error: empiricalhist must be a single non-NA logical value" + ) + + # Unequal common items + expect_error( + aFIPC::autoFIPC( + newformXData = data.frame(A=1), + oldformYData = data.frame(A=2), + newformCommonItemNames = c('A', 'B'), + oldformCommonItemNames = c('A') + ), + "Common Items are not equal" + ) + + # 0 length common items + expect_error( + aFIPC::autoFIPC( + newformXData = data.frame(A=1), + oldformYData = data.frame(A=2), + newformCommonItemNames = character(0), + oldformCommonItemNames = character(0) + ), + "Please provide common item names" + ) +}) diff --git a/tests/testthat/test-surveyFA-extras.R b/tests/testthat/test-surveyFA-extras.R new file mode 100644 index 0000000..c3d533e --- /dev/null +++ b/tests/testthat/test-surveyFA-extras.R @@ -0,0 +1,19 @@ +test_that("surveyFA specific branches", { + skip_if_not_installed("mirt") + + # 1. Invalid itemtype + expect_error(aFIPC::surveyFA(data.frame(a=c(0,1), b=c(1,0)), itemtype=c("2PL", "3PL")), "surveyFA requires itemtype to be a single non-NA character value.") + expect_error(aFIPC::surveyFA(data.frame(a=c(0,1), b=c(1,0)), itemtype=123), "surveyFA requires itemtype to be a single non-NA character value.") + + # 2. Invalid maxItemRemovals + expect_error(aFIPC::surveyFA(data.frame(a=c(0,1), b=c(1,0)), maxItemRemovals="3"), "surveyFA requires maxItemRemovals to be a non-negative numeric scalar.") + expect_error(aFIPC::surveyFA(data.frame(a=c(0,1), b=c(1,0)), maxItemRemovals=-1), "surveyFA requires maxItemRemovals to be a non-negative numeric scalar.") + + # 3. Invalid pThreshold + expect_error(aFIPC::surveyFA(data.frame(a=c(0,1), b=c(1,0)), pThreshold="0.05"), "surveyFA requires pThreshold to be in \\(0, 1\\]\\.") + expect_error(aFIPC::surveyFA(data.frame(a=c(0,1), b=c(1,0)), pThreshold=-0.1), "surveyFA requires pThreshold to be in \\(0, 1\\]\\.") + expect_error(aFIPC::surveyFA(data.frame(a=c(0,1), b=c(1,0)), pThreshold=1.5), "surveyFA requires pThreshold to be in \\(0, 1\\]\\.") + + # 4. forceUIRT = FALSE (triggers stop) + expect_error(aFIPC::surveyFA(data.frame(a=c(0,1), b=c(1,0)), forceUIRT = FALSE), "surveyFA requires forceUIRT=TRUE") +})