def test_search_field(field, molecular_definition, params, ref_ids):
query = rcsb.FieldQuery(field, molecular_definition, **params)
test_ids = rcsb.search(query)
test_count = rcsb.count(query)
assert set(test_ids) == set(ref_ids)
assert test_count == len(ref_ids)
# and count the number
for i, year in enumerate(years):
# A query that comprises one year
date_query = rcsb.FieldQuery(
"rcsb_accession_info.initial_release_date",
range_closed=(datetime.combine(datetime(year, 1, 1), time.min),
datetime.combine(datetime(year, 12, 31), time.max)))
xray_query = rcsb.FieldQuery("exptl.method",
exact_match="X-RAY DIFFRACTION")
nmr_query = rcsb.FieldQuery("exptl.method", exact_match="SOLUTION NMR")
em_query = rcsb.FieldQuery("exptl.method",
exact_match="ELECTRON MICROSCOPY")
# Get the amount of structures, that were released in that year
# AND were elucidated with the respective method
xray_count[i], nmr_count[i], em_count[i] = [
rcsb.count(date_query & method_query)
for method_query in [xray_query, nmr_query, em_query]
]
# Get the total amount of structures released in that year
tot_count[i] = rcsb.count(date_query)
fig, ax = plt.subplots(figsize=(8.0, 5.0))
ax.set_title("PDB release statistics")
ax.set_xlim(years[0] - 1, years[-1] + 1)
ax.set_xticks(years)
ax.set_xticklabels([str(y) for y in years], rotation=45)
ax.set_xlabel("Year")
ax.set_ylabel("Released structures per year")
ax.bar(years, xray_count, color=biotite.colors["darkorange"], label="X-Ray")
ax.bar(years,
nmr_count,
########################################################################
# In many cases you are not interested in a specific structure, but you
# want a set of structures that fits your desired criteria.
# For this purpose the *RCSB* search API can be used.
# At first you have to create :class:`Query` object for the property you
# want to filter.
# The :func:`search()` method takes the :class:`Query` and returns a
# list of PDB IDs, which itself can be used as input for
# :func:`fetch()`.
# Likewise, :func:`count()` is used to count the number of matching
# PDB IDs.
query = rcsb.BasicQuery("HCN1")
pdb_ids = rcsb.search(query)
print(pdb_ids)
print(rcsb.count(query))
files = rcsb.fetch(pdb_ids, "mmtf", gettempdir())
########################################################################
# This was a simple search for the occurrence of the search term in any
# field.
# You can also search for a value in a specific field with a
# :class:`FieldQuery`.
# A complete list of the available fields and its supported operators
# is documented
# `on this page <https://search.rcsb.org/search-attributes.html>`_.
# Query for 'lacA' gene
query1 = rcsb.FieldQuery("rcsb_entity_source_organism.rcsb_gene_name.value",
exact_match="lacA")
# Query for resolution below 1.5 Å
def test_search_basic():
query = rcsb.BasicQuery("tc5b")
assert rcsb.search(query) == ["1L2Y"]
assert rcsb.count(query) == 1
def test_search_invalid(field, params):
invalid_query = rcsb.FieldQuery(field, **params)
with pytest.raises(RequestError, match="400"):
rcsb.search(invalid_query)
with pytest.raises(RequestError, match="400"):
rcsb.count(invalid_query)
def test_search_empty():
query = rcsb.BasicQuery("This will not match any ID")
assert rcsb.search(query) == []
assert rcsb.count(query) == 0
def test_search_return_type(return_type, expected):
query = rcsb.BasicQuery("tc5b")
assert rcsb.search(query, return_type) == expected
assert rcsb.count(query, return_type) == len(expected)
def test_search_motif():
# motif is taken from official RCSB search API tutorial
MOTIF = "C-x(2,4)-C-x(3)-[LIVMFYWC]-x(8)-H-x(3,5)-H."
query = rcsb.MotifQuery(MOTIF, "prosite", "protein")
test_count = rcsb.count(query)
assert test_count == pytest.approx(503, rel=0.1)
