tda.conseq

rule format_gene_dep_csv:
    inputs:
        data={'type': 'dep-matrix', 'label' ~ 'Chronos_Combined|Chronos_Achilles|Chronos_Score|RNAi_merged|CERES_Combined|Sanger_CRISPR|Avana'},
        hdf5_utils=fileref("scripts/hdf5_utils.py", copy_to="hdf5_utils.py")
    outputs:
        {'type': 'dep-matrix-csv', 
        'dataset_id': '{{inputs.data.orig_dataset_id}}', 
        'label': '{{ inputs.data.label }}', 
        'confounders_label': '{{ inputs.data.confounders_label }}',
        'filename': {"$filename": 'deps.csv'} }
    run "python3" with """
        from hdf5_utils import read_hdf5

        df = read_hdf5("{{ inputs.data.filename }}")
        df = df.transpose()
        df.to_csv("deps.csv",index_label="Row.name")
    """

rule common_essentials:
    executor: dsub {"docker_image": "us.gcr.io/broad-achilles/depmap-pipeline-tda:v10", "min_ram": "10"}
    inputs:
        data={'type': 'dep-matrix-csv'},
        script=fileref('scripts/tda/CE_percentile_rank_analysis.py')
    outputs:
        {'type': 'common-essentials', 'label': '{{ inputs.data.label }}', 'filename': {'$filename': 'ce.csv'} }
    run "python3 {{ inputs.script.filename }} {{ inputs.data.filename }} ce.csv"

rule dep_moments:
    executor: dsub {"docker_image": "us.gcr.io/broad-achilles/depmap-pipeline-tda:v10", "min_ram": "10"}
    inputs:
        data={'type': 'dep-matrix-csv'},
        script=fileref('scripts/tda/moments.py')
    outputs:
        {'type': 'dep-moments', 'label': '{{ inputs.data.label }}', 'filename': {'$filename': 'moments.csv'} }
    run "python3 {{ inputs.script.filename }} {{ inputs.data.filename }} moments.csv"


####### LRT

let genes_per_lrt_task = '1000'


rule lrt_partition:
    inputs: 
        data={'type': 'dep-matrix-csv'},
        partition_lrt_input=fileref('scripts/tda/lrt/partition-lrt-input.py')
    outputs: 
        {'type': 'lrt-partitions', 
        'dataset_id': '{{inputs.data.dataset_id}}', 
        'label': '{{ inputs.data.label }}', 
        'filename': {'$filename': 'partitions.csv'} }
    run "python3 {{ inputs.partition_lrt_input.filename }} {{ inputs.data.filename }} partitions.csv {{ config.genes_per_lrt_task }}"

rule lrt:
    inputs:
        partitions={'type': 'lrt-partitions', 'dataset_id': dataset_id, 'label': label},
        data={'type': 'dep-matrix-csv', 'dataset_id': dataset_id, 'label': label},
        sparkles_config=fileref('sparkles-config'),
        lrt_sh=fileref('scripts/tda/lrt/lrt.sh')
    outputs: 
        {'type': 'lrt-scores', 
        'label': '{{ inputs.data.label }}', 
        'dataset_id': '{{ inputs.data.dataset_id }}', 
        'filename': { '$filename': 'out.csv' }}
    watch-regex: "task.*|Submitting job: .*" # print out the task status summary from sparkles
    run """bash {{ inputs.lrt_sh.filename }} \
        {{ config.SCRIPT_DIR }} \
        {{ config.sparkles_path }} \
        {{ inputs.sparkles_config.filename }} \
        {{ inputs.partitions.filename }} \
        {{ inputs.data.filename }} \
        out.csv
        """

####### Merge the results from above into a single dataset

rule summarize_gene_deps:
    inputs:
        common_essentials=all {'type':'common-essentials'},
        dep_moments=all {'type':'dep-moments'},
        lrt=all {'type':'lrt-scores'},
        probs=all {'type': 'raw-dep-prob-matrix'},
        count_dep_lines=fileref("scripts/count_dep_lines.py"),
        summarize_gene_deps=fileref("scripts/summarize_gene_deps.py")
    outputs: {'type': 'gene-dep-summary', 'filename': {"$filename": "deps.csv"}}

    run "python" with """
    import json
    with open("probs.json", "wt") as fd:
        fd.write(json.dumps( {{ inputs.probs }} ))
    """

    run "python3 {{ inputs.count_dep_lines.filename }} probs.json depcounts.csv"

    run "python3" with """
        import pandas as pd

        def read_artifacts(artifacts):
            dfs = []
            for a in artifacts:
                df = pd.read_csv(a['filename'])
                df['label'] = a['label']
                dfs.append(df)
            return pd.concat(dfs, ignore_index=True)

        common_essentials = read_artifacts( {{ inputs.common_essentials }} )
        common_essentials.to_csv("common_essentials.csv", index=False)

        dep_moments = read_artifacts( {{ inputs.dep_moments }} )
        dep_moments.to_csv("moments.csv", index=False)

        lrt = read_artifacts( {{ inputs.lrt }} )
        lrt.to_csv("lrt.csv", index=False)
    """

    run """
    python3 {{ inputs.summarize_gene_deps.filename }} \
        output=deps.csv \
        common_essentials=common_essentials.csv \
        moments=moments.csv \
        lrt=lrt.csv \
        depcounts=depcounts.csv
    """