def test_exclusive_match():
f = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
f.exclusive_match("material.elements", "Al")
res1 = f.search()
assert check_field(res1, "material.elements", "Al") == 0
f.exclusive_match("material.elements", ["Al", "Cu"])
res2 = f.search()
assert check_field(res2, "material.elements", "Al") == 1
assert check_field(res2, "material.elements", "Cu") == 1
assert check_field(res2, "material.elements", "Cp") == -1
assert check_field(res2, "material.elements", "Fe") == -1
def test_add_sort_external():
f = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
# Sort ascending by atomic number
f.match_field("mdf.source_name", "oqmd")
f.add_sort('crystal_structure.number_of_atoms', True)
res = f.search(limit=1, reset_query=False)
assert res[0]['crystal_structure']['number_of_atoms'] == 1
# Sort descending by composition, with multi-sort
f.add_sort('material.composition', False)
res = f.search(limit=1)
assert res[0]['crystal_structure']['number_of_atoms'] == 1
assert res[0]['material']['composition'].startswith('Zr')
def test_anonymous(capsys):
f = SearchHelper(INDEX, anonymous=True)
# Test search
assert len(
f.search("mdf.source_name:ab_initio_solute_database",
advanced=True,
limit=300)) == 300
def test_match_field():
f = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
# Basic usage
f.match_field("mdf.source_name", "khazana_vasp")
res1 = f.search()
assert check_field(res1, "mdf.source_name", "khazana_vasp") == 0
# Check that query clears
assert f.current_query() == ""
# Also checking check_field and no-op
f.match_field("material.elements", "Al")
f.match_field("", "")
res2 = f.search() # Enough so that we'd find at least 1 non-Al example
assert check_field(res2, "material.elements", "Al") == 1
def test_reset_query():
f = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
# Term will return results
f.match_field("material.elements", "Al")
f.reset_query()
# Specifying no query will raise an error
with pytest.raises(ValueError):
assert f.search() == []
def test_exclude_field():
f = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
# Basic usage
f.exclude_field("material.elements", "Al")
f.exclude_field("", "")
f.match_field("mdf.source_name", "ab_initio_solute_database")
f.match_field("mdf.resource_type", "record")
res1 = f.search()
assert check_field(res1, "material.elements", "Al") == -1
def test_search(capsys):
# Error on no query
with pytest.raises(ValueError):
f = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
f.search()
# Return info if requested
res2 = f.search("Al", info=False)
assert isinstance(res2, list)
assert isinstance(res2[0], dict)
res3 = f.search("Al", info=True)
assert isinstance(res3, tuple)
assert isinstance(res3[0], list)
assert isinstance(res3[0][0], dict)
assert isinstance(res3[1], dict)
# Check limit
res4 = f.match_term("Al").search(limit=3)
assert len(res4) == 3
# Check reset_query
f.match_field("mdf.source_name", "ta_melting")
res5 = f.search(reset_query=False)
res6 = f.search()
assert all([r in res6 for r in res5]) and all([r in res5 for r in res6])
# Check default index
f2 = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
assert (f2.match_term("data").search(
limit=1,
info=True)[1]["index_uuid"] == mdf_toolbox.translate_index(INDEX))
def test_exclude_range():
# Single-value use
f = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
f.exclude_range("material.elements", "Am", "*")
f.exclude_range("material.elements", "*", "Ak")
f.match_field("material.elements", "*")
res1, info1 = f.search(info=True)
assert (check_field(res1, "material.elements", "Al") == 0
or check_field(res1, "material.elements", "Al") == 2)
res2, info2 = f.search("material.elements:Al", advanced=True, info=True)
assert info1["total_query_matches"] <= info2["total_query_matches"]
# Non-matching use, test inclusive
f.exclude_range("material.elements", "Am", "*")
f.exclude_range("material.elements", "*", "Ak")
f.exclude_range("material.elements", "Al", "Al", inclusive=False)
f.match_field("material.elements", "*")
res3, info3 = f.search(info=True)
assert info1["total_query_matches"] == info3["total_query_matches"]
# Nothing to match
assert f.exclude_range("field", start=None, stop=None) == f
def test_match_range():
# Single-value use
f = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
f.match_range("material.elements", "Al", "Al")
res1, info1 = f.search(info=True)
assert check_field(res1, "material.elements", "Al") == 1
res2, info2 = f.search("material.elements:Al", advanced=True, info=True)
assert info1["total_query_matches"] == info2["total_query_matches"]
# Non-matching use, test inclusive
f.match_range("material.elements", "Al", "Al", inclusive=False)
assert f.search() == []
# Actual range
f.match_range("material.elements", "Al", "Cu")
res4, info4 = f.search(info=True)
assert info1["total_query_matches"] < info4["total_query_matches"]
assert (check_field(res4, "material.elements", "Al") >= 0
or check_field(res4, "material.elements", "Cu") >= 0)
# Nothing to match
assert f.match_range("field", start=None, stop=None) == f
def test_chaining():
# Internal
q = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
q._field("source_name", "cip")
q._and_join()
q._field("elements", "Al")
res1 = q._ex_search(limit=10000)
res2 = (SearchHelper(INDEX, search_client=SEARCH_CLIENT)._field(
"source_name", "cip")._and_join()._field("elements",
"Al")._ex_search(limit=10000))
assert all([r in res2 for r in res1]) and all([r in res1 for r in res2])
# External
f = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
f.match_field("source_name", "cip")
f.match_field("material.elements", "Al")
res1 = f.search()
res2 = f.match_field("source_name",
"cip").match_field("material.elements",
"Al").search()
assert all([r in res2 for r in res1]) and all([r in res1 for r in res2])
def test_match_not_exists():
f = SearchHelper(INDEX, search_client=SEARCH_CLIENT)
# Basic usage
f.match_not_exists("services.citrine")
assert check_field(f.search(), "services.citrine", ".*") == -1