From 21ac66de2f92b7a433c6677d7526ee8d9639d999 Mon Sep 17 00:00:00 2001
From: Andrea Guarracino
Date: Mon, 22 Jun 2020 22:41:04 +0200
Subject: little fix for specimen_source

---
 scripts/from_genbank_to_fasta_and_yaml.py | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/scripts/from_genbank_to_fasta_and_yaml.py b/scripts/from_genbank_to_fasta_and_yaml.py
index 87e99d4..6216340 100755
--- a/scripts/from_genbank_to_fasta_and_yaml.py
+++ b/scripts/from_genbank_to_fasta_and_yaml.py
@@ -305,12 +305,14 @@ for path_metadata_xxx_xml in [os.path.join(dir_metadata, name_metadata_xxx_xml)
                         if GBQualifier_value_text in term_to_uri_dict:
                             info_for_yaml_dict['sample']['specimen_source'] = [term_to_uri_dict[GBQualifier_value_text]]
                         else:
-                            if GBQualifier_value_text.lower() in ['np/op', 'np/op swab', 'np/np swab', 'nasopharyngeal and oropharyngeal swab', 'nasopharyngeal/oropharyngeal swab']:
+                            if GBQualifier_value_text.lower() in ['np/op', 'np/op swab', 'np/np swab', 'nasopharyngeal and oropharyngeal swab', 'nasopharyngeal/oropharyngeal swab', 'combined nasopharyngeal and oropharyngeal swab']:
                                 info_for_yaml_dict['sample']['specimen_source'] = [term_to_uri_dict['nasopharyngeal swab'], term_to_uri_dict['oropharyngeal swab']]
-                            elif GBQualifier_value_text in ['nasopharyngeal swab/throat swab', 'nasopharyngeal/throat swab', 'nasopharyngeal swab and throat swab', 'nasal swab and throat swab']:
+                            elif GBQualifier_value_text.lower() in ['nasopharyngeal swab/throat swab', 'nasopharyngeal/throat swab', 'nasopharyngeal swab and throat swab', 'nasal swab and throat swab', 'nasopharyngeal aspirate/throat swab']:
                                 info_for_yaml_dict['sample']['specimen_source'] = [term_to_uri_dict['nasopharyngeal swab'], term_to_uri_dict['throat swab']]
-                            elif GBQualifier_value_text in ['nasopharyngeal aspirate/throat swab']:
-                                info_for_yaml_dict['sample']['specimen_source'] = [term_to_uri_dict['nasopharyngeal aspirate'], term_to_uri_dict['throat swab']]
+                            elif GBQualifier_value_text.lower() in ['nasal swab and throat swab']:
+                                info_for_yaml_dict['sample']['specimen_source'] = [term_to_uri_dict['nasal swab'], term_to_uri_dict['throat swab']]
+                            elif GBQualifier_value_text.lower() in ['nasal-swab and oro-pharyngeal swab']:
+                                info_for_yaml_dict['sample']['specimen_source'] = [term_to_uri_dict['nasal swab'], term_to_uri_dict['oropharyngeal swab']]
                             else:
                                 missing_value_list.append('\t'.join([accession_version, 'specimen_source', GBQualifier_value_text]))
                     elif GBQualifier_name_text == 'collection_date':
-- 
cgit v1.2.3