nonquarterly-processed.conseq

include "xrefs-nonquarterly-unprocessed.conseq"

rule process_taiga_pulled_artifact:
   inputs:
      download={"type": "download_from_taiga"},
      hdf5_utils=fileref("scripts/hdf5_utils.py")
   run "python" with """
      import shutil
      import taigapy
      import os

      tc = taigapy.TaigaClient() # TODO: Need to update to Taiga Client V3 but don't know what the requested format should be for download_to_cache

      ### The if blocks are here to convert the new oncref format to the old oncref format so that 
      ### the oncref datasets can be ingested in the pipeline without much additional changes.
      if {{ inputs.download.label == "Prism_oncology_AUC" or inputs.download.label == "Prism_oncology_IC50"}}:
         prism_oncref_df = tc.get("{{ inputs.download.dataset_id }}")
         prism_oncref_df = prism_oncref_df.transpose()
         prism_oncref_df.to_csv("out.csv", index=True)
      elif {{ inputs.download.label == "Prism_oncology_per_curve" }}:
         prism_oncref_curves_df = tc.get("{{ inputs.download.dataset_id }}")
         prism_oncref_curves_df = prism_oncref_curves_df.rename(columns={
            'ModelID': 'cell_line_name',
            'SampleID': 'compound_name',
            'EC50': 'ec50',
            'LowerAsymptote': 'lower_asymptote',
            'UpperAsymptote': 'upper_asymptote',
            'Slope': 'slope'
         })
         prism_oncref_curves_df.to_csv("out.csv", index=False)
      else:
         cached = tc.download_to_cache("{{ inputs.download.dataset_id }}")
         shutil.copy2(cached, "out.csv")
      
      assert os.path.exists("out.csv"), "Output file 'out.csv' not generated"
	"""
   # Convert files to HDF5 if needed
   run "bash" with """
      if [[ "{{ inputs.download.format }}" == "hdf5" ]]; then
         python {{ inputs.hdf5_utils.filename }} to_hdf5 out.csv csv out.hdf5
      fi
   """
   # The input is polymorphic and we want the output to have (mostly) the same keys
   run "python" with """
      import json

      artifact = {{inputs.download | quoted}}
      artifact["filename"] = {"$filename": "out.{{ inputs.download.format }}"}
      artifact["type"] = artifact["target_type"]

      del artifact["target_type"]
      del artifact["$manually-added"]

      with open("results.json", "w") as f:
         json.dump({"outputs": [artifact]}, f)
   """
# TODO: Need to update to Taiga Client V3 but don't know what the requested format should be for download_to_cache
rule process_taiga_pulled_aggregated_dose_artifact:
   inputs:
      download={"type": "download_aggregated_dose_from_taiga"},
      hdf5_utils=fileref("scripts/hdf5_utils.py"),
      script=fileref("scripts/process_taiga_pulled_aggregated_dose_artifact.py")
   outputs: {
      "type": "aggregated-dose-replicate-level",
      "label": "{{ inputs.download.label }}",
      "dataset_id": "{{ inputs.download.dataset_id }}",
      "orig_dataset_id": "{{ inputs.download.orig_dataset_id }}",
      "cell_lines_dataset_id": "{{ inputs.download.cell_lines_dataset_id }}",
      "perturbations_dataset_id": "{{ inputs.download.perturbations_dataset_id }}",
      "hdf5_filename": {"$filename": "out.hdf5"},
      "cell_lines_filename": {"$filename": "cell_lines.csv"},
      "perturbations": {"$filename": "perturbations.csv"}
   }
   run "python {{ inputs.script.filename }} {{ inputs.download.label }} {{ inputs.download.cell_lines_dataset_id }} {{ inputs.download.perturbations_dataset_id }} {{ inputs.download.dataset_id }}"
   run "python {{ inputs.hdf5_utils.filename }} to_hdf5 out.csv csv out.hdf5"