Tool complains that contig does not exist when using Glow pipe transformer

Question

I am getting errors from the software that contig chr1 does not exist within the header. Does anybody have any insight as to why the infer function might not be inferring contigs?

Answer 1

As of release 0.4.1, Glow can only infer FORMAT and INFO header lines from a variant dataframe.

However, you can tell Glow to use the header from the input VCF when calling the command line tool.

Example:

import glow
vcf_path = "<path>"
df = spark.read.format("vcf").load(vcf_path)
glow.transform(
    "pipe",
    df,
    cmd=["tool"],
    input_formatter="vcf",
    in_vcf_header=vcf_path,
    output_formatter="vcf"
)