Retrive and display generic bundle errors

Implement the first QC check for generic errors e.g. missing files in
bundle, etc. Display retrieved errors on UI.

1 files changed, 36 insertions, 4 deletions

diff --git a/scripts/qc_on_rqtl2_bundle.py b/scripts/qc_on_rqtl2_bundle.py
index 5802f3a..8dd6b7e 100644
--- a/scripts/qc_on_rqtl2_bundle.py
+++ b/scripts/qc_on_rqtl2_bundle.py
@@ -1,6 +1,10 @@
 """Run Quality Control checks on R/qtl2 bundle."""
 import sys
-from logging import getLogger, StreamHandler
+import json
+from typing import Sequence
+from zipfile import ZipFile
+from argparse import Namespace
+from logging import Logger, getLogger, StreamHandler
 
 from redis import Redis
 
@@ -8,12 +12,40 @@ from qc_app import jobs
 from qc_app.db_utils import database_connection
 from qc_app.check_connections import check_db, check_redis
 
+from r_qtl import r_qtl2_qc as rqc
+from r_qtl import fileerrors as rqfe
+
 from scripts.cli_parser import init_cli_parser
+from scripts.process_rqtl2_bundle import parse_job
 from scripts.redis_logger import setup_redis_logger
 
-def run_qc(_args):
+def add_to_errors(rconn: Redis, fqjobid: str, key: str, errors: Sequence[rqfe.MissingFile]):
+    """Add `errors` to a given list of errors"""
+    errs = tuple(set(
+        json.loads(rconn.hget(fqjobid, key) or "[]") +
+        [error.message for error in errors]))
+    rconn.hset(fqjobid, key, json.dumps(errs))
+
+def run_qc(rconn: Redis, args: Namespace, logger: Logger) -> int:
     """Run the QC programs."""
-    return 1
+    fqjobid = jobs.job_key(args.redisprefix, args.jobid)
+    thejob = parse_job(rconn, args.redisprefix, args.jobid)
+    meta = thejob["job-metadata"]
+    with ZipFile(meta["rqtl2-bundle-file"], "r") as zfile:
+        missing = rqc.missing_files(zfile)
+        add_to_errors(rconn, fqjobid, "errors-generic", tuple(
+            rqfe.MissingFile(
+                mfile[0], mfile[1], (
+                    f"File '{mfile[1]}' is listed in the control file under "
+                    f"the '{mfile[0]}' key, but it does not actually exist in "
+                    "the bundle."))
+            for mfile in missing))
+
+    if len(missing) > 0:
+        logger.error("Missing files in the bundle!")
+        return 1
+
+    return 0
 
 if __name__ == "__main__":
     def main():
@@ -34,7 +66,7 @@ if __name__ == "__main__":
                 rconn, fqjobid, f"{fqjobid}:log-messages",
                 args.redisexpiry))
 
-            exitcode = run_qc(args)
+            exitcode = run_qc(rconn, args, logger)
             rconn.hset(
                 jobs.job_key(args.redisprefix, args.jobid), "exitcode", exitcode)
             return exitcode