Use persistent HTTP connection

src/biotite/database/afdb/download.py

@@ -87,6 +87,7 @@ def fetch(ids, format, target_path=None, overwrite=False, verbose=False):
         target_path.mkdir(parents=True, exist_ok=True)
 
     files = []
+    session = requests.Session()
     for i, id in enumerate(ids):
         # Verbose output
         if verbose:
@@ -98,7 +99,7 @@ def fetch(ids, format, target_path=None, overwrite=False, verbose=False):
             # 'file = None' -> store content in a file-like object
             file = None
         if file is None or not file.is_file() or file.stat().st_size == 0 or overwrite:
-            file_response = requests.get(_get_file_url(id, format))
+            file_response = session.get(_get_file_url(session, id, format))
             _assert_valid_file(file_response, id)
             if format in _BINARY_FORMATS:
                 content = file_response.content
@@ -128,12 +129,14 @@ def fetch(ids, format, target_path=None, overwrite=False, verbose=False):
         return files
 
 
-def _get_file_url(id, format):
+def _get_file_url(session, id, format):
     """
     Get the actual file URL for the given ID from the ``prediction`` API endpoint.
 
     Parameters
     ----------
+    session : requests.Session
+        The session to use for the request.
     id : str
         The ID of the file to be downloaded.
     format : str
@@ -146,7 +149,7 @@ def _get_file_url(id, format):
     """
     uniprot_id = _extract_id(id)
     try:
-        metadata = requests.get(f"{_METADATA_URL}/{uniprot_id}").json()
+        metadata = session.get(f"{_METADATA_URL}/{uniprot_id}").json()
     except requests.exceptions.JSONDecodeError:
         raise RequestError("Received malformed JSON response")
     if len(metadata) == 0:

src/biotite/database/entrez/download.py

@@ -106,6 +106,7 @@ def fetch(
     if target_path is not None and not isdir(target_path):
         os.makedirs(target_path)
     files = []
+    session = requests.Session()
     for i, id in enumerate(uids):
         # Verbose output
         if verbose:
@@ -127,7 +128,7 @@ def fetch(
             api_key = get_api_key()
             if api_key is not None:
                 param_dict["api_key"] = api_key
-            r = requests.get(_fetch_url, params=param_dict)
+            r = session.get(_fetch_url, params=param_dict)
             check_for_errors(r)
             content = r.text
             if file is None:

src/biotite/database/pubchem/download.py

@@ -107,6 +107,7 @@ def fetch(
         os.makedirs(target_path)
 
     files = []
+    session = requests.Session()
     for i, cid in enumerate(cids):
         # Prevent IDs as strings, this could be a common error, as other
         # database interfaces of Biotite use string IDs
@@ -125,7 +126,7 @@ def fetch(
 
         if file is None or not isfile(file) or getsize(file) == 0 or overwrite:
             record_type = "2d" if as_structural_formula else "3d"
-            r = requests.get(
+            r = session.get(
                 _base_url + f"compound/cid/{cid}/{format.upper()}",
                 params={"record_type": record_type},
             )

src/biotite/database/rcsb/download.py

@@ -102,6 +102,7 @@ def fetch(
         gz_suffix = ""
 
     files = []
+    session = requests.Session()
     for i, id in enumerate(pdb_ids):
         # Verbose output
         if verbose:
@@ -116,25 +117,25 @@ def fetch(
 
         if file is None or not isfile(file) or getsize(file) == 0 or overwrite:
             if format == "pdb":
-                r = requests.get(_standard_url + id + ".pdb" + gz_suffix)
+                r = session.get(_standard_url + id + ".pdb" + gz_suffix)
                 _assert_valid_file(r, id)
                 if gzip:
                     content = r.content
                 else:
                     content = r.text
             elif format in ["cif", "mmcif", "pdbx"]:
-                r = requests.get(_standard_url + id + ".cif" + gz_suffix)
+                r = session.get(_standard_url + id + ".cif" + gz_suffix)
                 _assert_valid_file(r, id)
                 if gzip:
                     content = r.content
                 else:
                     content = r.text
             elif format in ["bcif"]:
-                r = requests.get(_bcif_url + id + ".bcif" + gz_suffix)
+                r = session.get(_bcif_url + id + ".bcif" + gz_suffix)
                 _assert_valid_file(r, id)
                 content = r.content
             elif format == "fasta":
-                r = requests.get(_fasta_url + id)
+                r = session.get(_fasta_url + id)
                 _assert_valid_file(r, id)
                 content = r.text
             else:

src/biotite/database/uniprot/download.py

@@ -93,6 +93,7 @@ def fetch(ids, format, target_path=None, overwrite=False, verbose=False):
     if target_path is not None and not isdir(target_path):
         os.makedirs(target_path)
     files = []
+    session = requests.Session()
     for i, id in enumerate(ids):
         db_name = _get_database_name(id)
         # Verbose output
@@ -106,7 +107,7 @@ def fetch(ids, format, target_path=None, overwrite=False, verbose=False):
             file = None
         if file is None or not isfile(file) or getsize(file) == 0 or overwrite:
             if format in ["fasta", "gff", "txt", "xml", "rdf", "tab"]:
-                r = requests.get(_fetch_url + db_name + "/" + id + "." + format)
+                r = session.get(_fetch_url + db_name + "/" + id + "." + format)
                 content = r.text
                 assert_valid_response(r)
             else:

tests/database/test_afdb.py

@@ -98,7 +98,7 @@ def test_fetch_invalid(monkeypatch, format, invalid_id, bypass_metadata):
         monkeypatch.setattr(
             module,
             "_get_file_url",
-            lambda id, f: f"https://alphafold.ebi.ac.uk/files/AF-{id}-F1-model_v4.{f}",
+            lambda session, id, format: f"{AFDB_URL}files/AF-{id}-F1-model_v4.{format}",
         )
     with pytest.raises((RequestError, ValueError)):
         afdb.fetch(invalid_id, format)