aboutsummaryrefslogtreecommitdiff
path: root/gn3
diff options
context:
space:
mode:
authorBonfaceKilz2022-03-14 18:04:27 +0300
committerBonfaceKilz2022-03-14 18:04:27 +0300
commita8c3e35cc80bf14ba40a136bb54e8850d23c0065 (patch)
treec82ecd2f2b322f20425c49e55d153495ab9a504a /gn3
parenta4ffdf5caced76f720b44f2b3ccd13a9be6f7040 (diff)
downloadgenenetwork3-a8c3e35cc80bf14ba40a136bb54e8850d23c0065.tar.gz
Given a csv text and permissible headers, extract invalid headers
* gn3/csvcmp.py (extract_invalid_csv_headers): New function. * tests/unit/test_csvcmp.py: Import "extract_invalid_csv_headers". (test_extract_invalid_csv_headers_with_some_wrong_headers): Test case for the above.
Diffstat (limited to 'gn3')
-rw-r--r--gn3/csvcmp.py14
1 files changed, 14 insertions, 0 deletions
diff --git a/gn3/csvcmp.py b/gn3/csvcmp.py
index aa057b7..dd3f72b 100644
--- a/gn3/csvcmp.py
+++ b/gn3/csvcmp.py
@@ -118,3 +118,17 @@ def get_allowable_sampledata_headers(conn: Any) -> List:
attributes += [attributes[0] for attributes in
cursor.fetchall()]
return attributes
+
+
+def extract_invalid_csv_headers(allowed_headers: List, csv_text: str) -> List:
+ """Check whether a csv text's columns contains valid headers"""
+ csv_header = []
+ for line in csv_text.split("\n"):
+ if line.startswith("Strain Name"):
+ csv_header = [_l.strip() for _l in line.split(",")]
+ break
+ invalid_headers = []
+ for header in csv_header:
+ if header not in allowed_headers:
+ invalid_headers.append(header)
+ return invalid_headers