Run process after all other processes have finished

[jdlamstein]

Hey,

I want to run a process after all processes have finished. In the workflow below, the user runs Intensity, Crop, and a CNN with boolean params. The channels are set earlier as strings or integers. I want to know how to consolidate the output of the processes so the GETCSVS process runs after all have finished.

    if (params.DO_INTENSITY) {
        int_ch = INTENSITY(track_result, experiment_ch, norm_ch, morphology_ch, target_channel_ch, well_ch, tp_ch, well_toggle_ch, tp_toggle_ch)
        intensity_result = INTENSITY.out
        int_ch.view { it }
    }
    else {
        intensity_result = true
    }
    if (params.DO_CROP) {
        crop_ch = CROP(track_result, experiment_ch, target_channel_crop_ch, morphology_ch, crop_size_ch, well_ch, tp_ch, well_toggle_ch, tp_toggle_ch)
        crop_ch.view { it }
        crop_result = CROP.out
    }
    else {
        crop_result = true
    }
    if (params.DO_CNN) {
        cnn_ch = CNN(crop_result, experiment_ch, label_type_ch, label_name_ch, classes_ch, img_norn_name_cnn_ch, filters_ch, num_channels_ch, n_samples_ch,
        epochs_ch, batch_size_ch, learning_rate_ch, momentum_ch, optimizer_ch, well_ch, tp_ch, chosen_channels_for_cnn_ch,
        well_toggle_ch, tp_toggle_ch, channel_toggle_ch)
        cnn_result = CNN.out
    }
    else {
        cnn_result = true
    }
// How do I set csv_ready?
    if (params.DO_GET_CSVS) {
        csv_ch = GETCSVS(csv_ready, experiment_ch)
        csv_ch.view { it }
    }

where one of the processes is

process CNN {
    containerOptions "--gpus all --mount type=bind,src=/data/,target=/data/"
    maxForks = 1

    input:
    val ready
    val exp
    val label_type
    val label_name
    val classes
    val img_norm_name
    val filters
    val num_channels
    val n_samples
    val epochs
    val batch_size
    val learning_rate
    val momentum
    val optimizer
    val chosen_wells
    val chosen_timepoints
    val chosen_channels
    val wells_toggle
    val timepoints_toggle
    val channels_toggle

    output:
    val true

    script:
    """
    cnn.py --experiment ${exp} --label_type ${label_type} --label_name ${label_name} --classes ${classes} \
    --img_norm_name ${img_norm_name} --filters ${filters} --num_channels ${num_channels} --n_samples ${n_samples} \
    --epochs ${epochs} --batch_size ${batch_size} --learning_rate ${learning_rate} \
    --momentum ${momentum} --optimizer ${optimizer} \
    --chosen_wells ${chosen_wells} --chosen_timepoints ${chosen_timepoints} --chosen_channels ${chosen_channels}\
    --wells_toggle ${wells_toggle} --timepoints_toggle ${timepoints_toggle} --channels_toggle ${channels_toggle}
    """
}

Thanks!

[mribeirodantas]

Hello, @jdlamstein. If you want a process only to start when some subset of other processes have finished, you can use the collect channel operator on their outputs. Check this link for more details.