comparison XSeekerPreparator.R @ 12:bdb2878ee189 draft

" master branch Updating"

11:f4fc4a0f41e2

    context$groupidx <- rdata$xa@xcmsSet@groupidx
    xcms_set <- rdata$xa@xcmsSet
    singlefile <- rdata$singlefile
    process_arg_list <- rdata$listOFlistArguments
    var_meta <- rdata$variableMetadata
    ## We needed to get rid of the rdata, which is very big.
    ## So we gathered all variable assignment from rdata here, and got rid of it.
    rm(rdata)

    process_params <- list()
    for (list_name in names(process_arg_list)) {
        param_list <- list()
        for (param_name in names(process_arg_list[[list_name]])) {
            param_list[[param_name]] <- process_arg_list[[list_name]][[param_name]]
        }
        process_params[[length(process_params)+1]] <- param_list
    }
    message("Parameters from previous processes extracted.")


    indices <- as.numeric(unique(var_meta[, file_grouping_var]))
    smol_xcms_set <- orm$smol_xcms_set()

        env <- new.env()
        ms_file <- xcms::xcmsRaw(sample_path)
        env$tic <- ms_file@tic
        env$mz <- ms_file@env$mz
        env$scanindex <- ms_file@scanindex
        env$scantime <- ms_file@scantime
        env$intensity <- ms_file@env$intensity
        env$polarity <- as.character(ms_file@polarity[[1]])

        ## Again, ms file is huge, so we get rid of it quickly.
        rm(ms_file)

    }
    context$clusters <- list()
    context$show_percent <- show_percent
    context$cluster_mean_rt_abundance <- list()
    context$central_feature <- list()
    load_variable_metadata(orm, var_meta, context)
    clusters <- context$clusters
    rm(context)
    message("Features enrichment")
    complete_features(orm, clusters, show_percent)

            if (!is.na(max_o <- extract_peak_var(sample_peak_list, "maxo"))) {
                dummy_feature$set_max_o(max_o)
            }
        }
        create_associated_cluster(
            context$central_feature[[clusterID]],
            dummy_feature, clusterID,
            context, curent_var_meta, next_pc_group,
            next_align_group
        )

    }
    return (clusterID + next_cluster_id)
}

create_associated_cluster <- function(
    sample_no, feature, clusterID,
    context, curent_var_meta, next_pc_group, next_align_group
) {
    pcgroup <- as.numeric(curent_var_meta[["pcgroup"]])
    adduct <- as.character(curent_var_meta[["adduct"]])
    annotation <- curent_var_meta[["isotopes"]]
    clusterID <- as.character(clusterID)
    if (is.null(cluster <- context$clusters[[clusterID]])) {
        cluster <- context$clusters[[clusterID]] <- orm$cluster(
            pc_group=pcgroup + next_pc_group,
            adduct=adduct,
            align_group=next_align_group,
            # curent_group=curent_group,
            clusterID=context$clusterID,
            annotation=annotation
        )
        ## Crappy hack to assign sample id to cluster without loading the sample.
        ## Samples are too big (their sample$env) and slows the process, and eat all the menory
        ## so we dont't want to load them.
        cluster[["sample_id"]] <- context$samples[sample_no][[1]]
        cluster$modified__[["sample_id"]] <- cluster[["sample_id"]]

12:bdb2878ee189

    context$groupidx <- rdata$xa@xcmsSet@groupidx
    xcms_set <- rdata$xa@xcmsSet
    singlefile <- rdata$singlefile
    process_arg_list <- rdata$listOFlistArguments
    var_meta <- rdata$variableMetadata

    process_params <- list()
    if (is.null(process_arg_list)) {
        histories <- list()
        for (history in xcms_set@.processHistory) {
            if (
                class(history@param) == "CentWaveParam"
                && history@type == "Peak detection"
            ) {
                params <- history@param
                process_params <- list(list(
                    xfunction="annotatediff",
                    ppm=params@ppm,
                    peakwidth=sprintf("%s - %s", params@peakwidth[[1]], params@peakwidth[[2]]),
                    snthresh=params@snthresh,
                    prefilterStep=params@prefilter[[1]],
                    prefilterLevel=params@prefilter[[2]],
                    mzdiff=params@mzdiff,
                    fitgauss=params@fitgauss,
                    noise=params@noise,
                    mzCenterFun=params@mzCenterFun,
                    integrate=params@integrate,
                    firstBaselineCheck=params@firstBaselineCheck,
                    snthreshIsoROIs=!identical(params@roiScales, numeric(0))
                ))
                break
            }
        }
    } else {
        for (list_name in names(process_arg_list)) {
            param_list <- list()
            for (param_name in names(process_arg_list[[list_name]])) {
                param_list[[param_name]] <- process_arg_list[[list_name]][[param_name]]
            }
            process_params[[length(process_params)+1]] <- param_list
        }
    }

    message("Parameters from previous processes extracted.")


    indices <- as.numeric(unique(var_meta[, file_grouping_var]))
    smol_xcms_set <- orm$smol_xcms_set()

        env <- new.env()
        ms_file <- xcms::xcmsRaw(sample_path)
        env$tic <- ms_file@tic
        env$mz <- ms_file@env$mz
        env$scanindex <- ms_file@scanindex
        env$scantime <- ms_file@scantime * 60
        env$intensity <- ms_file@env$intensity
        env$polarity <- as.character(ms_file@polarity[[1]])

        ## Again, ms file is huge, so we get rid of it quickly.
        rm(ms_file)

    }
    context$clusters <- list()
    context$show_percent <- show_percent
    context$cluster_mean_rt_abundance <- list()
    context$central_feature <- list()
    context$adducts <- list()
    load_variable_metadata(orm, var_meta, context)
    clusters <- context$clusters
    rm(context)
    message("Features enrichment")
    complete_features(orm, clusters, show_percent)

            if (!is.na(max_o <- extract_peak_var(sample_peak_list, "maxo"))) {
                dummy_feature$set_max_o(max_o)
            }
        }
        create_associated_cluster(
            orm,
            context$central_feature[[clusterID]],
            dummy_feature, clusterID,
            context, curent_var_meta, next_pc_group,
            next_align_group
        )

    }
    return (clusterID + next_cluster_id)
}

create_associated_cluster <- function(
    orm,
    sample_no, feature, clusterID,
    context, curent_var_meta, next_pc_group, next_align_group
) {
    clusterID <- as.character(clusterID)
    if (is.null(cluster <- context$clusters[[clusterID]])) {
        pcgroup <- as.numeric(curent_var_meta[["pcgroup"]])
        adduct_name <- as.character(curent_var_meta[["adduct"]])
        annotation <- curent_var_meta[["isotopes"]]
        cluster <- context$clusters[[clusterID]] <- orm$cluster(
            pc_group=pcgroup + next_pc_group,
            # adduct=adduct,
            align_group=next_align_group,
            # curent_group=curent_group,
            clusterID=context$clusterID,
            annotation=annotation
        )
        if (is.null(adduct <- context$adducts[[adduct_name]])) {
            context$adducts[[adduct_name]] <- orm$adduct()$load_by(name=adduct_name)$first()
            if (is.null(adduct <- context$adducts[[adduct_name]])) {
                adduct <- context$adducts[[adduct_name]] <- orm$adduct(name=adduct_name, charge=0)
                adduct$save()
            }
        }
        cluster$set_adduct(adduct)
        ## Crappy hack to assign sample id to cluster without loading the sample.
        ## Samples are too big (their sample$env) and slows the process, and eat all the menory
        ## so we dont't want to load them.
        cluster[["sample_id"]] <- context$samples[sample_no][[1]]
        cluster$modified__[["sample_id"]] <- cluster[["sample_id"]]