-
Notifications
You must be signed in to change notification settings - Fork 10
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Added implementation of 'combination cohorts'. #193
base: main
Are you sure you want to change the base?
Changes from all commits
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
Original file line number | Diff line number | Diff line change |
---|---|---|
|
@@ -106,24 +106,28 @@ generateCohortSet <- function(connectionDetails = NULL, | |
if (incremental) { | ||
recordKeepingFile <- file.path(incrementalFolder, "GeneratedCohorts.csv") | ||
|
||
if (isTRUE(attr(cohortDefinitionSet, "hasSubsetDefinitions"))) { | ||
cohortDefinitionSet$checksum <- "" | ||
for (i in 1:nrow(cohortDefinitionSet)) { | ||
cohortDefinitionSet$checksum <- "" | ||
for (i in 1:nrow(cohortDefinitionSet)) { | ||
if (isTRUE(attr(cohortDefinitionSet, "hasSubsetDefinitions"))) { | ||
# This implementation supports recursive definitions (subsetting subsets) because the subsets have to be added in order | ||
if (cohortDefinitionSet$subsetParent[i] != cohortDefinitionSet$cohortId[i]) { | ||
if (cohortDefinitionSet$isSubset[i] && cohortDefinitionSet$subsetParent[i] != cohortDefinitionSet$cohortId[i]) { | ||
j <- which(cohortDefinitionSet$cohortId == cohortDefinitionSet$subsetParent[i]) | ||
cohortDefinitionSet$checksum[i] <- computeChecksum(paste( | ||
cohortDefinitionSet$sql[j], | ||
cohortDefinitionSet$sql[i] | ||
)) | ||
} else { | ||
cohortDefinitionSet$checksum[i] <- computeChecksum(cohortDefinitionSet$sql[i]) | ||
} | ||
} else if (isTRUE(attr(cohortDefinitionSet, "hasCombinedCohorts"))) { | ||
dependantCohortIds <- as.integer(strsplit(cohortDefinitionSet$dependentCohorts[i])) | ||
dependentCohortIdx <- which(cohortDefinitionSet$cohortId %in% dependantCohortIds) | ||
cohortDefinitionSet$checksum[i] <- | ||
Comment on lines
+120
to
+123
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. The loop logic is now: if it is a subset cohort, calc checksum one way, if it is combined cohort, do it another way, else do it the simple 'by sql' way. |
||
computeChecksum(paste0(c(cohortDefinitionSet$sql[dependentCohortIdx], cohortDefinitionSet$sql[i]), collapse = "")) | ||
} else { | ||
cohortDefinitionSet$checksum <- computeChecksum(cohortDefinitionSet$sql) | ||
} | ||
} else { | ||
cohortDefinitionSet$checksum <- computeChecksum(cohortDefinitionSet$sql) | ||
} | ||
} | ||
|
||
# Create the cluster | ||
# DEV NOTE :: running subsets in a multiprocess setup will not work with subsets that subset other subsets | ||
# To resolve this issue we need to execute the dependency tree. | ||
|
@@ -145,7 +149,7 @@ generateCohortSet <- function(connectionDetails = NULL, | |
dplyr::select("cohortId") %>% | ||
dplyr::pull() | ||
} | ||
|
||
# Apply the generation operation to the cluster | ||
cohortsGenerated <- ParallelLogger::clusterApply( | ||
cluster, | ||
|
@@ -253,7 +257,15 @@ generateCohort <- function(cohortId = NULL, | |
rlang::inform(paste0(i, "/", nrow(cohortDefinitionSet), "- Generating cohort: ", cohortName, " (id = ", cohortId, ")")) | ||
sql <- cohortDefinitionSet$sql[i] | ||
|
||
if (!isSubset) { | ||
if (isSubset) { | ||
sql <- SqlRender::render( | ||
sql = sql, | ||
cdm_database_schema = cdmDatabaseSchema, | ||
cohort_table = cohortTableNames$cohortTable, | ||
cohort_database_schema = cohortDatabaseSchema, | ||
warnOnMissingParameters = FALSE | ||
) | ||
} else { # combined cohorts apply same paramaters as standard cohort generation | ||
Comment on lines
+260
to
+268
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Same sort of re-organization here: before, the assumption was if it's not a subset, then it must be a standard cohort generation (with sql). But now that there's another choice, it is better to condition on positive identification |
||
sql <- SqlRender::render( | ||
sql = sql, | ||
cdm_database_schema = cdmDatabaseSchema, | ||
|
@@ -269,14 +281,6 @@ generateCohort <- function(cohortId = NULL, | |
results_database_schema.cohort_censor_stats = paste(cohortDatabaseSchema, cohortTableNames$cohortCensorStatsTable, sep = "."), | ||
warnOnMissingParameters = FALSE | ||
) | ||
} else { | ||
sql <- SqlRender::render( | ||
sql = sql, | ||
cdm_database_schema = cdmDatabaseSchema, | ||
cohort_table = cohortTableNames$cohortTable, | ||
cohort_database_schema = cohortDatabaseSchema, | ||
warnOnMissingParameters = FALSE | ||
) | ||
} | ||
sql <- SqlRender::translate( | ||
sql = sql, | ||
|
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,249 @@ | ||
# CombinedCohortOp ------------- | ||
#' @title A Combined cohort operation used to UNION or INTERSECT cohorts. Note: only UNION supported. | ||
#' @export | ||
#' @description | ||
#' Defines a UNION or INTERSECT on a set of cohorts. | ||
CombinedCohortOp <- R6::R6Class( | ||
classname = "CombinedCohortOp", | ||
private = list( | ||
.targetCohortIds = NULL, | ||
.opType = "union" | ||
), | ||
active = list( | ||
#' @field targetCohortIds The list of cohorts to apply in this group. | ||
targetCohortIds = function(targetCohortIds) { | ||
if (missing(targetCohortIds)) { | ||
return(private$.targetCohortIds) | ||
} | ||
checkmate::assertList(as.list(targetCohortIds), types="numeric", min.len = 0) | ||
private$.targetCohortIds <- targetCohortIds | ||
return(self) | ||
}, | ||
#' @field opType The group operation, either 'union' or 'intersect' | ||
opType = function(opType) { | ||
if (missing(opType)) { | ||
return(private$.opType) | ||
} | ||
checkmate::assertChoice(opType, c("union")) | ||
private$.opType <- opType | ||
return(self) | ||
} | ||
), | ||
public = list( | ||
#' @description | ||
#' creates a new instance, using the provided data param if provided. | ||
#' @param data the data (as a json string or list) to initialize with | ||
initialize = function(data = list()) { | ||
dataList <- .convertJSON(data) | ||
|
||
if ("targetCohortIds" %in% names (dataList)) self$targetCohortIds <- dataList$targetCohortIds | ||
if ("opType" %in% names (dataList)) self$opType <- dataList$opType | ||
}, | ||
#' @description List representation of object | ||
toList = function() { | ||
.removeEmpty(list( | ||
targetCohortIds = .toJsonArray(private$.targetCohortIds), | ||
opType = jsonlite::unbox(private$.opType) | ||
)) | ||
}, | ||
#' To JSON | ||
#' @description json serialized representation of object | ||
toJSON = function() { | ||
.toJSON(self$toList()) | ||
}, | ||
|
||
#' Is Equal to | ||
#' @description Compare CombinedCohortGroup to another | ||
#' @param other CombinedCohortGroup instance | ||
isEqualTo = function(other) { | ||
checkmate::assertR6(other, "CombinedCohortGroup") | ||
return(other$toJSON() == self$toJSON()) | ||
}, | ||
#' getDependentCohortIds | ||
#' @description Gets the dependent cohortIds from this operation | ||
getDependentCohortIds = function() { | ||
if (length(private$.targetCohortIds) == 0) return(c()) | ||
return(private$.targetCohortIds) | ||
} | ||
) | ||
) | ||
|
||
# CombinedCohortDef ------------- | ||
#' @title A CombinedCohortDef | ||
#' @export | ||
#' @description | ||
#' Defines the ID, name and operation to produce a new cohort. | ||
CombinedCohortDef <- R6::R6Class( | ||
classname = "CombinedCohortDef", | ||
private = list( | ||
.cohortId = NA, | ||
.cohortName = NA, | ||
.expression = NULL | ||
), | ||
active = list( | ||
#' @field cohortId The list of cohorts to apply in this group. | ||
cohortId = function(cohortId) { | ||
if (missing(cohortId)) { | ||
return(private$.cohortId) | ||
} | ||
checkmate::assertInt(cohortId) | ||
private$.cohortId <- cohortId | ||
return(self) | ||
}, | ||
#' @field cohortName the name given to this outcome definition | ||
cohortName = function(cohortName) { | ||
if (missing(cohortName)) { | ||
private$.cohortName | ||
} else { | ||
# check type | ||
checkmate::assertCharacter(cohortName) | ||
private$.cohortName <- cohortName | ||
self | ||
} | ||
}, | ||
#' @field expression The combo operator that is the root of the definition. | ||
expression = function(expression) { | ||
if (missing(expression)) { | ||
return(private$.expression) | ||
} | ||
checkmate::assertClass(expression, classes="CombinedCohortOp") | ||
private$.expression <- expression | ||
return(self) | ||
} | ||
), | ||
public = list( | ||
#' @description | ||
#' creates a new instance, using the provided data param if provided. | ||
#' @param data the data (as a json string or list) to initialize with | ||
initialize = function(data = list()) { | ||
dataList <- .convertJSON(data) | ||
|
||
if ("cohortId" %in% names (dataList)) self$targetCohortIds <- dataList$cohortId | ||
if ("cohortName" %in% names (dataList)) self$cohortName <- dataList$cohortName | ||
if ("expression" %in% names (dataList)) self$expression<- CohortGenerator::CombinedCohortOp$new(dataList$expression) | ||
}, | ||
#' @description List representation of object | ||
toList = function() { | ||
.removeEmpty(list( | ||
cohortId = jsonlite::unbox(private$.cohortId), | ||
cohortName = jsonlite::unbox(private$.cohortName), | ||
expression = .r6ToListOrNA(private$.expression) | ||
)) | ||
}, | ||
#' To JSON | ||
#' @description json serialized representation of object | ||
toJSON = function() { | ||
.toJSON(self$toList()) | ||
}, | ||
|
||
#' Is Equal to | ||
#' @description Compare CombinedCohortDef to another | ||
#' @param other CombinedCohortDef instance | ||
isEqualTo = function(other) { | ||
checkmate::assertR6(other, "CombinedCohortDef") | ||
return(other$toJSON() == self$toJSON()) | ||
} | ||
) | ||
) | ||
|
||
### Factory Functions | ||
|
||
#' Create CombinedCohortOp instance | ||
#' @description | ||
#' A factory function to create CombinedCohortOp | ||
#' @export | ||
#' @param targetCohortIds list of target cohort IDs to combine in this operation | ||
#' @param opType The op type of this cohort combination, can only be 'union' | ||
createCombinedCohortOp <- function(targetCohortIds, opType) { | ||
|
||
cohortOp <- CombinedCohortOp$new() | ||
if (!missing(targetCohortIds)) cohortOp$targetCohortIds <- targetCohortIds | ||
if (!missing(opType)) cohortOp$opType <- opType | ||
|
||
return (cohortOp); | ||
} | ||
|
||
#' Defines a combined cohort using combined cohort operations | ||
#' @description | ||
#' Creates an instance of CombinedCohortDef with the provided cohortId, cohortName and the combine operator expression | ||
#' @export | ||
#' @param cohortId The output cohort id from applying the combine expression. | ||
#' @param cohortName The output cohort name | ||
#' @param expression The combine operator that will yield the final cohort. | ||
createCombinedCohortDef <- function(cohortId, cohortName, expression) { | ||
|
||
cohortDef <- CombinedCohortDef$new(); | ||
if (!missing(cohortId)) cohortDef$cohortId <- cohortId; | ||
if (!missing(cohortName)) cohortDef$cohortName <- cohortName; | ||
if (!missing(expression)) cohortDef$expression <- expression; | ||
|
||
return (cohortDef); | ||
} | ||
|
||
|
||
|
||
#' Add cohort subset definition to a cohort definition set | ||
#' @description | ||
#' Given a subset definition and cohort definition set, this function returns a modified cohortDefinitionSet | ||
#' That contains cohorts that's have parent's contained within the base cohortDefinitionSet | ||
#' | ||
#' Also adds the columns subsetParent and isSubset that denote if the cohort is a subset and what the parent definition | ||
#' is. | ||
#' @export | ||
#' @param cohortDefinitionSet data.frame that conforms to CohortDefinitionSet | ||
#' @param combinedCohortDefiniton CombinedCohortDefinition instance | ||
#' @param overwriteExisting Overwrite existing subset definition of the same definitionId if present | ||
addCombinedCohort <- function(cohortDefinitionSet, | ||
combinedCohortDefiniton, | ||
overwriteExisting = FALSE) { | ||
checkmate::assertTRUE(isCohortDefinitionSet(cohortDefinitionSet)) | ||
checkmate::assertR6(combinedCohortDefiniton, "CombinedCohortDef") | ||
checkmate::assertTRUE(!is.null(combinedCohortDefiniton$expression)) | ||
|
||
if (!"dependentCohorts" %in% colnames(cohortDefinitionSet)) { | ||
cohortDefinitionSet$dependentCohorts <- "" | ||
} | ||
|
||
if (!"isCombinedCohort" %in% colnames(cohortDefinitionSet)) { | ||
cohortDefinitionSet$isCombinedCohort <- FALSE | ||
} | ||
|
||
dependentCohortIds <- combinedCohortDefiniton$expression$getDependentCohortIds() | ||
if (length(dependentCohortIds) > 0) { | ||
checkmate::assertSubset(dependentCohortIds, cohortDefinitionSet$cohortId) | ||
} | ||
|
||
if (!overwriteExisting && nrow(cohortDefinitionSet %>% dplyr::filter(.data$cohortId == combinedCohortDefiniton$cohortId)) > 0) { | ||
stop("The specified cohortId for this combined cohort already exists in the cohort definition set") | ||
} else { | ||
# remove this definition from the set | ||
cohortDefinitionSet <- cohortDefinitionSet %>% dplyr::filter(.data$cohortId != combinedCohortDefiniton$cohortId) | ||
} | ||
|
||
defSql <- SqlRender::readSql(system.file("sql", "sql_server", "combinedCohorts", "CombinedCohortDefinition.sql", package = "CohortGenerator")) | ||
|
||
queryBuilder <- CombinedCohortQueryBuilder$new() | ||
combinationQuery <- queryBuilder$buildQuery(combinedCohortDefiniton) | ||
|
||
defSql <- SqlRender::render(defSql, | ||
output_cohort_id = combinedCohortDefiniton$cohortId, | ||
combined_cohort_query = combinationQuery) | ||
|
||
cohortDefinitionSet <- | ||
dplyr::bind_rows( | ||
cohortDefinitionSet, | ||
data.frame( | ||
cohortId = combinedCohortDefiniton$cohortId, | ||
cohortName = combinedCohortDefiniton$cohortName, | ||
sql = defSql, | ||
json = as.character(combinedCohortDefiniton$toJSON()), | ||
isCombinedCohort = TRUE, | ||
dependentCohorts = paste0(dependentCohortIds, collapse = ",") | ||
) | ||
) | ||
|
||
attr(cohortDefinitionSet, "hasCombinedCohorts") <- TRUE | ||
|
||
return(cohortDefinitionSet) | ||
} | ||
|
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,16 @@ | ||
CombinedCohortQueryBuilder <- R6::R6Class( | ||
classname = "CombinedCohortQueryBuilder", | ||
private = list( | ||
), | ||
public = list( | ||
buildQuery = function(combinedCohortDefiniton) { | ||
checkmate::assertR6(combinedCohortDefiniton, "CombinedCohortDef") | ||
sql <- SqlRender::readSql(system.file("sql", "sql_server", "combinedCohorts", "unionCombine.sql", package = "CohortGenerator")) | ||
sql <- SqlRender::render(sql, | ||
target_cohort_ids = combinedCohortDefiniton$expression$targetCohortIds, | ||
warnOnMissingParameters = TRUE | ||
) | ||
return(sql) | ||
} | ||
) | ||
) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This was moved around to better work with > 2 cohort types. When it was just the two, it was simpler to say 'either or' in this loop, but arranged this way we can apply different styles of generated cohorts to CohortGenerator.