Merge pull request #836 from padix-key/gallery-fixes

Hotfix: Fox documentation build

Author: padix-key

doc/examples/scripts/sequence/annotation/plasmid_map.py

@@ -25,8 +25,8 @@
 import biotite.sequence.io.genbank as gb
 
 PLASMID_URL = (
-    "https://media.addgene.org/snapgene-media/v2.0.0/sequences/12250/"
-    "9998fdbe-051f-4dc6-ba0f-24e65127a0c5/addgene-plasmid-26092-sequence-12250.gbk"
+    "https://media.addgene.org/snapgene-media/v3.0.0/sequences/466943/"
+    "17c6ce2c-cf6d-46e8-a4c9-58cd4a4760b6/addgene-plasmid-26092-sequence-466943.gbk"
 )
 
 

doc/tutorial/database/rcsb.rst

@@ -139,12 +139,12 @@ They can be included in search results by adding ``"computational"`` to the
 .. jupyter-execute::
 
     query = (
-        rcsb.FieldQuery("rcsb_polymer_entity.pdbx_description", contains_phrase="Hexokinase")
+        rcsb.FieldQuery("rcsb_polymer_entity.pdbx_description", contains_phrase="Lysozyme")
         & rcsb.FieldQuery(
             "rcsb_entity_source_organism.scientific_name", exact_match="Homo sapiens"
         )
     )
-    ids = rcsb.search(query, content_types=("experimental", "computational"))
+    ids = rcsb.search(query, content_types=("computational",))
     print(ids)
 
 The returned four-character IDs are the RCSB PDB IDs of experimental structures

src/biotite/database/afdb/download.py

@@ -18,7 +18,7 @@
 # Adopted from https://www.uniprot.org/help/accession_numbers
 # adding the optional 'AF-' prefix and '-F1' suffix used by RCSB
 _UNIPROT_PATTERN = (
-    r"^(?P<prefix>AF-)?"
+    r"^(?P<prefix>(AF-)|(AF_AF))?"
     r"(?P<id>[OPQ][0-9][A-Z0-9]{3}[0-9]|[A-NR-Z][0-9]([A-Z][A-Z0-9]{2}[0-9]){1,2})"
     r"(?P<suffix>-?F1)?$"
 )
@@ -34,8 +34,8 @@ def fetch(ids, format, target_path=None, overwrite=False, verbose=False):
     ----------
     ids : str or iterable object of str
         A single ID or a list of IDs of the file(s) to be downloaded.
-        They can be either UniProt IDs (e.g. ``P12345``) or AlphaFold DB IDs
-        (e.g. ``AF-P12345F1``).
+        They can be either UniProt IDs (e.g. ``P12345``), AlphaFold DB IDs
+        (e.g. ``AF-P12345-F1``) or computational RCSB IDs (e.g. ``AF_AFP12345F1``).
     format : {'pdb', 'pdbx', 'cif', 'mmcif', 'bcif', 'fasta'}
         The format of the files to be downloaded.
     target_path : str, optional
@@ -145,7 +145,10 @@ def _get_file_url(id, format):
         The URL of the file to be downloaded.
     """
     uniprot_id = _extract_id(id)
-    metadata = requests.get(f"{_METADATA_URL}/{uniprot_id}").json()
+    try:
+        metadata = requests.get(f"{_METADATA_URL}/{uniprot_id}").json()
+    except requests.exceptions.JSONDecodeError:
+        raise RequestError("Received malformed JSON response")
     if len(metadata) == 0:
         raise RequestError(f"ID {id} is invalid")
     # A list of length 1 is always returned, if the response is valid

src/biotite/structure/io/pdb/convert.py

@@ -16,6 +16,7 @@
     "list_assemblies",
     "get_assembly",
     "get_unit_cell",
+    "get_symmetry_mates",
 ]
 
 import warnings

tests/database/test_afdb.py

@@ -5,6 +5,7 @@
 import tempfile
 import pytest
 import biotite.database.afdb as afdb
+import biotite.database.rcsb as rcsb
 import biotite.structure.io.pdb as pdb
 import biotite.structure.io.pdbx as pdbx
 from biotite.database import RequestError
@@ -15,15 +16,20 @@
 
 @pytest.mark.skipif(cannot_connect_to(AFDB_URL), reason="AlphaFold DB is not available")
 @pytest.mark.parametrize("as_file_like", [False, True])
-@pytest.mark.parametrize("entry_id", ["P12345", "AF-P12345-F1", "AF-P12345F1"])
+@pytest.mark.parametrize(
+    "entry_id", ["P12345", "AF-P12345-F1", "AF-P12345F1", "AF_AFP12345F1"]
+)
 @pytest.mark.parametrize("format", ["pdb", "cif", "bcif"])
 def test_fetch(as_file_like, entry_id, format):
     """
     Check if files in different formats can be downloaded by being able to parse them.
     Also ensure that the downloaded file refers to the given input ID
     """
     path = None if as_file_like else tempfile.gettempdir()
-    file_path_or_obj = afdb.fetch(entry_id, format, path, overwrite=True)
+    try:
+        file_path_or_obj = afdb.fetch(entry_id, format, path, overwrite=True)
+    except RequestError:
+        pytest.skip("AFDB is probably busy")
     if format == "pdb":
         file = pdb.PDBFile.read(file_path_or_obj)
         pdb.get_structure(file)
@@ -37,6 +43,34 @@ def test_fetch(as_file_like, entry_id, format):
         assert file.block["struct_ref"]["pdbx_db_accession"].as_item() == "P12345"
 
 
+def test_fetch_cross_id():
+    """
+    Test if :func:`afdb.fetch()` works with AlphaFold DB cross-references
+    in search results from :func:`rcsb.search()`.
+    """
+    UNIPROT_ID = "P12345"
+
+    ids = rcsb.search(
+        rcsb.FieldQuery(
+            "rcsb_entry_info.structure_determination_methodology",
+            exact_match="computational",
+        )
+        & rcsb.FieldQuery(
+            "rcsb_polymer_entity_container_identifiers"
+            ".reference_sequence_identifiers"
+            ".database_accession",
+            exact_match=UNIPROT_ID,
+        ),
+        content_types=("computational",),
+    )
+    assert len(ids) == 1
+    try:
+        pdbx_file = pdbx.CIFFile.read(afdb.fetch(ids[0], "cif"))
+    except RequestError:
+        pytest.skip("AFDB is probably busy")
+    assert pdbx_file.block["struct_ref"]["pdbx_db_accession"].as_item() == UNIPROT_ID
+
+
 @pytest.mark.skipif(cannot_connect_to(AFDB_URL), reason="AlphaFold DB is not available")
 def test_fetch_multiple():
     """

tests/test_doctest.py

@@ -86,9 +86,13 @@
     pytest.param(
         "biotite.database.afdb",
         [],
-        marks=pytest.mark.skipif(
-            cannot_connect_to(AFDB_URL), reason="AlphaFold DB is not available"
-        ),
+        marks=[
+            pytest.mark.skipif(
+                cannot_connect_to(AFDB_URL), reason="AlphaFold DB is not available"
+            ),
+            # TODO: Check if AFDB tests in CI are still flaky
+            pytest.mark.skip("Currently, AFDB is regularly busy"),
+        ],
     ),
     pytest.param(
         "biotite.database.uniprot",