Dont load MACs (#91)

* reset limit to '' and fix config.deploy

* handle empty 'limit' parameter

* - Upgrade `py-ard` to 1.0.3
- Use pyard without loading MACs, saves time and disk space

* Update memory for ValidateBuildOutputFunction

* The imgt version can be 3 or 4 digits

* Set timeout to 900 seconds.

Author: pbashyal-nmdp

.gitignore

@@ -69,6 +69,7 @@ target/
 # pyenv python configuration file
 .python-version
 .venv*
+venv*
 
 # Profiling tools
 profile*

Makefile

@@ -249,7 +249,7 @@ pipeline.jobs.deploy:
 
 config.deploy:
 	$(MAKE) -C ${APP_NAME}/pipeline/ service.config.deploy
-	$(MAKE) -C ${APP_NAME}/database/ config.deploy
+	$(MAKE) -C ${APP_NAME}/database/ service.config.deploy
 
 monitoring.create-subscriptions:
 	$(MAKE) -C ${APP_NAME}/infrastructure service.monitoring.create-subscriptions

gfe-db/pipeline/config/pipeline-input.json

@@ -2,5 +2,5 @@
   "align": false,
   "kir": false,
   "mem_profile": false,
-  "limit": "1000"
+  "limit": ""
 }

gfe-db/pipeline/functions/invoke_pipeline/app.py

@@ -49,8 +49,10 @@ def lambda_handler(event, context):
 
             # limit is an integer
             try:
-                if not isinstance(int(event['limit']), int):
-                    raise ValueError('limit must be an integer')
+                # empty limit implies no limit
+                if event['limit']:
+                    if not isinstance(int(event['limit']), int):
+                        raise ValueError('limit must be an integer')
             except ValueError:
                 raise ValueError('limit must be an integer')
 
@@ -150,7 +152,7 @@ def get_branches(owner, repo):
     return [branch["name"] for branch in branches]
 
 
-def is_valid_release(branch, release_pattern=r"^\d{3}0$"):
+def is_valid_release(branch, release_pattern=r"^\d{2,3}0$"):
     """Returns True if the branch is a valid release, False if not"""
 
     # IMGT/HLA release format string

gfe-db/pipeline/jobs/build/Dockerfile

@@ -1,4 +1,4 @@
-FROM python:3.8-slim-buster
+FROM --platform=linux/amd64 python:3.8-slim-buster
 
 RUN apt update && \
     apt-get install -y \

gfe-db/pipeline/jobs/build/requirements.txt

@@ -1,4 +1,4 @@
-py-ard==1.0.1
+py-ard==1.0.3
 py-gfe==1.1.5
 lxml==4.6.3
 Pympler==0.9

gfe-db/pipeline/jobs/build/src/app.py

@@ -569,7 +569,7 @@ def process_allele(allele, alignments_dict, csv_path=None):
 
     alleles = parse_dat(data_dir, dbversion)
 
-    ard = pyard.init(dbversion)
+    ard = pyard.init(dbversion, data_dir="/tmp/gfe-pyard/", load_mac=False)
 
     gfe_maker = GFE(
         verbose=verbose, 

gfe-db/pipeline/template.yaml

@@ -506,8 +506,8 @@ Resources:
       CodeUri: functions/validate_build_output/
       Handler: app.lambda_handler
       Runtime: python3.10
-      MemorySize: 256
-      Timeout: 60
+      MemorySize: 1024
+      Timeout: 900
       Architectures:
         - x86_64
       Environment: