updated models and load functions for almeida

regen.sh

@@ -0,0 +1,3 @@
+rm alembic/versions/*.py
+alembic revision --autogenerate -m 'init'
+alembic upgrade head

scripts/load_amr.py

@@ -24,8 +24,8 @@ def main(args):
 
     parsed_amr = parsers.parse_amr_summary(args.input)
     #print(parsed_amr)
-    sample_run = parsers.parse_run_ids(args.locations)
-    created_amr_summary = crud.create_amr_summary(session, parsed_amr, sample_run)
+    #sample_run = parsers.parse_run_ids(args.locations)
+    created_amr_summary = crud.create_amr_summary(session, parsed_amr, args.runid)
 
     for amr in created_amr_summary:
         stmt = select(Library).where(Library.id == amr.library_id)
@@ -37,7 +37,7 @@ def main(args):
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
     parser.add_argument('input')
-    parser.add_argument('--locations')
+    parser.add_argument('--runid')
     parser.add_argument('-c', '--config', help="config file (JSON format))")
     args = parser.parse_args()
     main(args)

scripts/load_cgmlst.py

@@ -28,13 +28,13 @@ def main(args):
     cgmlst_profiles = list(cgmlst_by_sample_id.values())
     cgmlst_scheme = {'name':'Ridom cgMLST.org','version':'2.1','num_loci':2891} 
 
-    sample_run = parsers.parse_run_ids(args.locations)
+    #sample_run = parsers.parse_run_ids(args.locations)
 
     #print(sample_run)
 
 
 
-    created_profiles = crud.create_cgmlst_allele_profiles(session, cgmlst_scheme, cgmlst_profiles, sample_run)
+    created_profiles = crud.create_cgmlst_allele_profiles(session, cgmlst_scheme, cgmlst_profiles, args.runid)
 
     for profile in created_profiles:
         stmt = select(Library).where(Library.id == profile.library_id)
@@ -46,7 +46,7 @@ def main(args):
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
     parser.add_argument('input')
-    parser.add_argument('--locations')
+    parser.add_argument('--runid')
     parser.add_argument('-c', '--config', help="config file (JSON format))")
     args = parser.parse_args()
     main(args)

scripts/load_cgmlst_cluster.py

@@ -26,9 +26,9 @@ def main(args):
 
     cgmlst_cluster_by_sample = parsers.parse_cgmlst_cluster(args.input)
 
-    sample_run = parsers.parse_run_ids(args.locations)
+    #sample_run = parsers.parse_run_ids(args.locations)
 
-    created_cgmlst_clusters = crud.add_samples_to_cgmlst_clusters(session, cgmlst_cluster_by_sample,sample_run)
+    created_cgmlst_clusters = crud.add_samples_to_cgmlst_clusters(session, cgmlst_cluster_by_sample,args.runid)
 
     for sample in created_cgmlst_clusters:
         print("added cluster to sample: " + sample.samples.sample_id)
@@ -39,7 +39,7 @@ def main(args):
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
     parser.add_argument('input')
-    parser.add_argument('--locations')
+    parser.add_argument('--runid')
     parser.add_argument('-c', '--config', help="config file (JSON format))")
     args = parser.parse_args()
     main(args)

scripts/load_complex.py

@@ -24,8 +24,8 @@ def main(args):
 
     parsed_complex = parsers.parse_complex(args.input)
 
-    sample_run = parsers.parse_run_ids(args.locations)
-    created_complexes = crud.create_complexes(session, parsed_complex,sample_run)
+    #sample_run = parsers.parse_run_ids(args.locations)
+    created_complexes = crud.create_complexes(session, parsed_complex,args.runid)
 
     #created_species = crud.create_species(session,)
 
@@ -41,7 +41,7 @@ def main(args):
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
     parser.add_argument('input')
-    parser.add_argument('--locations')
+    parser.add_argument('--runid')
     parser.add_argument('-c', '--config', help="config file (JSON format))")
     args = parser.parse_args()
     main(args)

scripts/load_library.py

@@ -21,7 +21,7 @@ def main(args):
     session = Session()
 
 
-    parsed_libraries = parsers.parse_libraries(args.qc, args.locations)
+    parsed_libraries = parsers.parse_libraries(args.input, args.runid)
 
     created_libraries = crud.create_libraries(session,parsed_libraries)
 
@@ -36,8 +36,8 @@ def main(args):
 
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
-    parser.add_argument('--qc')
-    parser.add_argument('--locations')
+    parser.add_argument('input')
+    parser.add_argument('--runid')
     parser.add_argument('-c', '--config', help="config file (JSON format))")
     args = parser.parse_args()
     main(args)

scripts/load_snpit.py

@@ -0,0 +1,43 @@
+#!/usr/bin/env python
+
+import argparse
+import csv
+import json
+
+from sqlalchemy import create_engine, select
+from sqlalchemy.orm import Session, sessionmaker
+
+import tb_db.parsers as parsers
+import tb_db.crud as crud
+
+from tb_db.models import Sample
+from tb_db.models import Library
+
+def main(args):
+    with open(args.config, 'r') as f:
+        config = json.load(f)
+    connection_uri = config['connection_uri']
+    engine = create_engine(connection_uri)
+    Session = sessionmaker(bind=engine)
+    session = Session()
+
+
+    parsed_snpit = parsers.parse_snpit(args.input)
+    print(parsed_snpit)
+    #sample_run = parsers.parse_run_ids(args.locations)
+    created_snpit = crud.create_snpit(session, parsed_snpit, args.runid)
+
+    for snp in created_snpit:
+        stmt = select(Library).where(Library.id == snp.library_id)
+        sample = session.scalars(stmt).one()
+        print("Created snpit for: " + sample.samples.sample_id)
+
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('input')
+    parser.add_argument('--runid')
+    parser.add_argument('-c', '--config', help="config file (JSON format))")
+    args = parser.parse_args()
+    main(args)

scripts/load_taxon.py

@@ -25,8 +25,14 @@ def main(args):
     parsed_species = parsers.parse_species(args.input)
 
 
-    sample_run = parsers.parse_run_ids(args.locations)
-    created_species = crud.create_species(session, parsed_species,sample_run)
+    #sample_run = parsers.parse_run_ids(args.locations)
+    print(parsed_species)
+
+    #for row in parsed_species:
+        #print(row)
+    #    print(parsed_species[row])
+        #print(row.key())
+    created_species = crud.create_species(session, parsed_species,args.runid)
 
     for created_speci in created_species:
         stmt = select(Library).where(Library.id == created_speci.library_id)
@@ -40,7 +46,7 @@ def main(args):
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
     parser.add_argument('input')
-    parser.add_argument('--locations')
+    parser.add_argument('--runid')
     parser.add_argument('-c', '--config', help="config file (JSON format))")
     args = parser.parse_args()
     main(args)

setup.py

@@ -10,6 +10,7 @@
     install_requires=[
         "psycopg2-binary==2.9.3",
         "sqlalchemy==1.4.40",
+        "sqlalchemy-utils==0.40.0",
         "alembic==1.8.0"
     ],
     description="",