Use all endpoints, explicit startup of hail runtime (#5)

Dockerfile

@@ -1,7 +1,7 @@
 FROM python:3.10-bullseye
 
 # take as a command line argument, or
-ARG RELEASE=${RELEASE:-1.2.0}
+ARG RELEASE=${RELEASE:-1.3.0}
 
 RUN apt update && apt install -y \
         apt-transport-https \

clinvarbitration/clinvar_by_codon.py

@@ -108,6 +108,9 @@ def main(input_vcf: str, output_root: str):
     # now set a schema to read that into a table... if you want hail
     schema = hl.dtype('struct{newkey:str,clinvar_alleles:str}')
 
+    # start the local hail runtime
+    hl.context.init_local(default_reference='GRCh38')
+
     # import the table, and transmute to top-level attributes
     ht = hl.import_table(json_out_path, no_header=True, types={'f0': schema})
     ht = ht.transmute(**ht.f0)

clinvarbitration/resummarise_clinvar.py

@@ -388,8 +388,7 @@ def parse_into_table(json_path: str, out_path: str) -> hl.Table:
     """
 
     # start a hail runtime
-    hl.init(default_reference='GRCh38')
-    # hl.context.init_local(default_reference='GRCh38')
+    hl.context.init_local(default_reference='GRCh38')
 
     # define the schema for each written line
     schema = hl.dtype(

example_script.sh

@@ -26,8 +26,8 @@ docker run -v "$(pwd)/data":/data clinvarbitration:example \
 ## This is where you should run VEP on data/clinvar_summary.vcf.bgz, with protein consequence annotation per transcript
 ## Let's imagine you did that, and the result is in data/pathogenic_annotated.vcf.bgz
 ## I've enclosed a 10-variant example of this, as annotated by https://www.ensembl.org/Homo_sapiens/Tools/VEP
-#docker run --platform linux/amd64 -v "$(pwd)/data":/data clinvarbitration:example \
-#    /bin/bash -c "python3 /clinvarbitration/clinvar_by_codon.py -i /data/pathogenic_annotated.vcf.bgz -o /data/pm5"
+docker run --platform linux/amd64 -v "$(pwd)/data":/data clinvarbitration:example \
+    pm5_table -i /data/pathogenic_annotated.vcf.bgz -o /data/pm5
 
 # upon completion, this will generate files in the data directory:
 # - data/pm5.json - a JSON file containing the PM5 results, one JSON object per line