yaml.scanner.ScannerError when initializing PEP downloaded with geofetch

[nleroy917]

I was trying to initialize a PEP I got from GEO using geofetch and peppy threw this error: yaml.scanner.ScannerError: mapping values are not allowed here

Here is the project config:

# Autogenerated by geofetch

name: GSE96155
pep_version: 2.1.0
sample_table: GSE96155_annotation.csv
subsample_table: null

looper:
  output_dir: GSE96155
  pipeline_interfaces: {pipeline_interfaces}

sample_modifiers:
  append:
    Sample_extract_protocol_ch1: general protocol: https://www.encodeproject.org/documents/be2a0f12-af38-430c-8f2d-57953baab5f5/@@download/attachment/Epigenomics_Alternative_Mag_Bead_ChIP_Protocol_v1.1_exp.pdf
    Sample_characteristics_ch1: link: ENCODE dbxrefs Cellosaurus CVCL_B260; http://web.expasy.org/cellosaurus/CVCL_B260
    Sample_data_processing: See GSM*_README.txt supplementary file linked below
    SRR_files: SRA
    
  derive:
    attributes: [read1, read2, SRR_files]
    sources:
      SRA: "${SRABAM}/{SRR}.bam"
      FQ: "${SRAFQ}/{SRR}.fastq.gz"
      FQ1: "${SRAFQ}/{SRR}_1.fastq.gz"
      FQ2: "${SRAFQ}/{SRR}_2.fastq.gz"      
  imply:
    - if: 
        organism: "Mus musculus"
      then:
        genome: mm10
    - if: 
        organism: "Homo sapiens"
      then:
        genome: hg38          
    - if: 
        read_type: "PAIRED"
      then:
        read1: FQ1
        read2: FQ2          
    - if: 
        read_type: "SINGLE"
      then:
        read1: FQ1

project_modifiers:
  amend:
    sra_convert:
      looper:
        results_subdir: sra_convert_results
      sample_modifiers:
        append:
          SRR_files: SRA
          pipeline_interfaces: ${CODE}/geofetch/pipeline_interface_convert.yaml
        derive:
          attributes: [read1, read2, SRR_files]
          sources:
            SRA: "${SRARAW}/{SRR}.sra"
            FQ: "${SRAFQ}/{SRR}.fastq.gz"
            FQ1: "${SRAFQ}/{SRR}_1.fastq.gz"
            FQ2: "${SRAFQ}/{SRR}_2.fastq.gz"

I think this is an issue with the sample_modifiers section. If I wrap the append modifiers in a string with double quotes, the issue resolves. However, I get a second error.

This time peppy throws a TypeError: Provided argument has to be a List[str] or a str, got 'NoneType'. I was able to solve this by removing entirely the subsample_table: null line. Then the project gets initialized properly.

I think there are two "bugs" here.

1. geofetch doesn't account for colons messing with the key-value syntax of yaml files.
2. A subsample_table key of null breaks a peppy.Project() instantiation. I am wondering if we should either not include the subsample table path in the config file if it doesn't exist or change peppy to permit null subsample tables.

[nsheff]

yes, I think you are exactly right on both of these.

For your second point -- I think you should do both:

1. raise an issue with peppy to allow null subsample tables.
2. Change geofetch to not create null subsample tables if not needed.