def test_add_single_synonym(self):
synonym_key = "synonym_key_1"
TaxonomyManager.register_synonym_key("synonym_key_1")
data = [
{
'id': 0,
'full_name': 'Family One',
'rank_name': 'One',
'rank': niamoto_db_meta.TaxonRankEnum.FAMILIA,
'parent_id': None,
'synonyms': {},
'mptt_left': 0,
'mptt_right': 0,
'mptt_tree_id': 0,
'mptt_depth': 0,
},
]
ins = niamoto_db_meta.taxon.insert().values(data)
with Connector.get_connection() as connection:
connection.execute(ins)
TaxonomyManager.add_synonym_for_single_taxon(0, synonym_key, 1)
df1 = TaxonomyManager.get_raw_taxon_dataframe()
self.assertEqual(df1.loc[0]['synonyms'], {synonym_key: 1})
TaxonomyManager.add_synonym_for_single_taxon(0, synonym_key, 2)
df2 = TaxonomyManager.get_raw_taxon_dataframe()
self.assertEqual(df2.loc[0]['synonyms'], {
synonym_key: 1,
synonym_key: 2,
})
def test_get_not_empty_raw_taxon_dataset(self):
data = [
{
'id': 0,
'full_name': 'Family One',
'rank_name': 'One',
'rank': niamoto_db_meta.TaxonRankEnum.FAMILIA,
'parent_id': None,
'synonyms': {},
'mptt_left': 0,
'mptt_right': 0,
'mptt_tree_id': 0,
'mptt_depth': 0,
},
{
'id': 1,
'full_name': 'Genus Two',
'rank_name': 'Two',
'rank': niamoto_db_meta.TaxonRankEnum.GENUS,
'parent_id': 0,
'synonyms': {},
'mptt_left': 0,
'mptt_right': 0,
'mptt_tree_id': 0,
'mptt_depth': 0,
},
{
'id': 2,
'full_name': 'Species Three',
'rank_name': 'Three',
'rank': niamoto_db_meta.TaxonRankEnum.SPECIES,
'parent_id': None,
'synonyms': {},
'mptt_left': 1,
'mptt_right': 0,
'mptt_tree_id': 0,
'mptt_depth': 0,
},
]
ins = niamoto_db_meta.taxon.insert().values(data)
with Connector.get_connection() as connection:
connection.execute(ins)
df1 = TaxonomyManager.get_raw_taxon_dataframe()
self.assertEqual(len(df1), 3)
def test_construct_mptt_less_simple_tree(self):
tree = [[1, [1, [
1,
]]], [
1,
], [
1,
], [
1,
], [1, [3, [1, [1, [
1,
]]]]]]
data, last_id = make_taxon_tree(tree)
self.assertEqual(len(data), 24)
ins = niamoto_db_meta.taxon.insert().values(data)
with Connector.get_connection() as connection:
connection.execute(ins)
TaxonomyManager.make_mptt()
mptt = TaxonomyManager.get_raw_taxon_dataframe()
self.assertEqual(list(mptt['mptt_tree_id']), [
1, 1, 1, 1, 1, 1, 7, 7, 9, 9, 11, 11, 13, 13, 13, 13, 13, 13, 13,
13, 13, 13, 13, 13
])
self.assertEqual(list(mptt['mptt_left']), [
1, 2, 4, 5, 8, 9, 1, 2, 1, 2, 1, 2, 1, 2, 4, 5, 7, 9, 12, 13, 15,
16, 19, 20
])
self.assertEqual(list(mptt['mptt_right']), [
12, 3, 7, 6, 11, 10, 4, 3, 4, 3, 4, 3, 24, 3, 11, 6, 8, 10, 23, 14,
18, 17, 22, 21
])
self.assertEqual(list(mptt['mptt_depth']), [
0, 1, 1, 2, 1, 2, 0, 1, 0, 1, 0, 1, 0, 1, 1, 2, 2, 2, 1, 2, 2, 3,
2, 3
])
self.assertEqual([
int(i) if pd.notnull(i) else None
for i in mptt['parent_id'].tolist()
], [
None, 1, 1, 3, 1, 5, None, 7, None, 9, None, 11, None, 13, 13, 15,
15, 15, 13, 19, 19, 21, 19, 23
])
def test_construct_mptt_simple_tree(self):
tree = [[1, [
1,
]], [
1,
]]
data, last_id = make_taxon_tree(tree)
ins = niamoto_db_meta.taxon.insert().values(data)
with Connector.get_connection() as connection:
connection.execute(ins)
TaxonomyManager.make_mptt()
mptt = TaxonomyManager.get_raw_taxon_dataframe()
self.assertEqual(list(mptt['mptt_tree_id']), [1, 1, 1, 1, 5, 5])
self.assertEqual(list(mptt['mptt_left']), [1, 2, 4, 5, 1, 2])
self.assertEqual(list(mptt['mptt_right']), [8, 3, 7, 6, 4, 3])
self.assertEqual(list(mptt['mptt_depth']), [0, 1, 1, 2, 0, 1])
self.assertEqual([
int(i) if pd.notnull(i) else None
for i in mptt['parent_id'].tolist()
], [None, 1, 1, 3, None, 5])
def test_get_empty_raw_taxon_dataset(self):
df1 = TaxonomyManager.get_raw_taxon_dataframe()
self.assertEqual(len(df1), 0)
def test_set_taxonomy(self):
result, synonyms = TaxonomyManager.set_taxonomy(
pd.DataFrame(
columns=['full_name', 'rank_name', 'rank', 'parent_id'], ))
self.assertEqual(result, 0)
data = pd.DataFrame.from_records([
{
'id': 0,
'full_name': 'Family One',
'rank_name': 'One',
'rank': niamoto_db_meta.TaxonRankEnum.FAMILIA,
'parent_id': None,
},
],
index='id')
result, synonyms = TaxonomyManager.set_taxonomy(data)
self.assertEqual(result, 1)
data = pd.DataFrame.from_records([
{
'id': 0,
'full_name': 'Family One',
'rank_name': 'One',
'rank': niamoto_db_meta.TaxonRankEnum.FAMILIA,
'parent_id': None,
'gbif': 5,
'taxref': 1,
},
{
'id': 1,
'full_name': 'Genus Two',
'rank_name': 'Two',
'rank': niamoto_db_meta.TaxonRankEnum.GENUS,
'parent_id': 0,
'gbif': 10,
'taxref': 2,
},
{
'id': 2,
'full_name': 'Species Three',
'rank_name': 'Three',
'rank': niamoto_db_meta.TaxonRankEnum.SPECIES,
'parent_id': None,
'gbif': 7,
'taxref': 3,
},
],
index='id')
result, synonyms = TaxonomyManager.set_taxonomy(data)
self.assertEqual(result, 3)
df = TaxonomyManager.get_raw_taxon_dataframe()
self.assertEqual(len(df), 3)
synonym_keys = TaxonomyManager.get_synonym_keys()
self.assertEqual(len(synonym_keys), 2)
identity_synonyms = TaxonomyManager.get_synonyms_for_key(
TaxonomyManager.IDENTITY_SYNONYM_KEY)
self.assertEqual(len(identity_synonyms), 3)
null_synonyms = TaxonomyManager.get_synonyms_for_key(None)
self.assertEqual(len(null_synonyms), 0)
gbif_synonyms = TaxonomyManager.get_synonyms_for_key("gbif")
self.assertEqual(len(gbif_synonyms), 3)
taxref_synonyms = TaxonomyManager.get_synonyms_for_key('taxref')
self.assertEqual(len(taxref_synonyms), 3)