agatha.ml.util.test_embedding_lookup module¶
-
agatha.ml.util.test_embedding_lookup.
assert_table_contains_embeddings
(actual=typing.Dict[str, typing.List[int]], expected=<class 'agatha.ml.util.embedding_lookup.EmbeddingLookupTable'>)¶ - Return type
None
-
agatha.ml.util.test_embedding_lookup.
assert_writable
(path)¶ - Return type
None
-
agatha.ml.util.test_embedding_lookup.
make_embedding_hdf5s
(part2embs, embedding_dir)¶ This function creates an embedding hdf5 file for test purposes.
- Return type
None
-
agatha.ml.util.test_embedding_lookup.
make_entity_lookup_table
(part2names, test_dir)¶ Writes embedding location database
- Return type
Path
-
agatha.ml.util.test_embedding_lookup.
setup_embedding_lookup_data
(name2vec, test_name, num_parts, test_root_dir=PosixPath('/tmp'))¶ Creates an embedding hdf5 file and an entity sqlite3 database for testing
- Parameters
name2vec (
Dict
[str
,List
[int
]]) – name2vec[x] = embedding of xtest_name (
str
) – A unique name for this testnum_parts (
int
) – The number of partitions to split this dataset among.
- Return type
Tuple
[Path
,Path
]- Returns
embedding_dir, entity_db_path You can run EmbeddingLookupTable(*setup_embedding_lookup_data(…))
-
agatha.ml.util.test_embedding_lookup.
test_embedding_keys
()¶
-
agatha.ml.util.test_embedding_lookup.
test_setup_lookup_data
()¶
-
agatha.ml.util.test_embedding_lookup.
test_setup_lookup_data_two_parts
()¶
-
agatha.ml.util.test_embedding_lookup.
test_typical_embedding_lookup
()¶