read_obo: default utf-8 encoding

closes #27

Author: dhimmel

obonet/io.py

@@ -1,3 +1,5 @@
+from __future__ import annotations
+
 import importlib
 import io
 import logging
@@ -6,12 +8,13 @@
 from urllib.request import urlopen
 
 
-def open_read_file(path):
+def open_read_file(path, encoding: str | None = None):
     """
     Return a file object from the path. Automatically detects and supports
     URLs and compression. If path is pathlike, it's converted to a string.
     If path is not a string nor pathlike, it's passed through without
-    modification.
+    modification. Use encoding to set the text character set encoding.
+    Use `encoding=None` to use the platform-dependent default locale encoding.
     """
     # Convert pathlike objects to string paths
     if hasattr(path, "__fspath__"):
@@ -29,16 +32,17 @@ def open_read_file(path):
         with urlopen(path) as response:
             content = response.read()
         if opener == io.open:
-            encoding = response.headers.get_content_charset(failobj="utf-8")
+            if not encoding:
+                encoding = response.headers.get_content_charset(failobj="utf-8")
             logging.info(f"Will decode content from {path} using {encoding} charset.")
             text = content.decode(encoding)
             return io.StringIO(text)
         else:
             compressed_bytes = io.BytesIO(content)
-            return opener(compressed_bytes, "rt")
+            return opener(compressed_bytes, "rt", encoding=encoding)
 
     # Read from file
-    return opener(path, "rt")
+    return opener(path, "rt", encoding=encoding)
 
 
 compression_to_module = {

obonet/read.py

@@ -1,3 +1,5 @@
+from __future__ import annotations
+
 import itertools
 import logging
 import re
@@ -9,7 +11,9 @@
 logger = logging.getLogger(__name__)
 
 
-def read_obo(path_or_file, ignore_obsolete=True):
+def read_obo(
+    path_or_file, ignore_obsolete: bool = True, encoding: str | None = "utf-8"
+):
     """
     Return a networkx.MultiDiGraph of the ontology serialized by the
     specified path or file.
@@ -25,8 +29,11 @@ def read_obo(path_or_file, ignore_obsolete=True):
     ignore_obsolete : boolean
         When true (default), terms that are marked 'is_obsolete' will
         not be added to the graph.
+    encoding : str of None
+        The character set encoding to use for path_or_file when path_or_file
+        is a path/URL. Set to None for platform-dependent locale default.
     """
-    obo_file = open_read_file(path_or_file)
+    obo_file = open_read_file(path_or_file, encoding=encoding)
     typedefs, terms, instances, header = get_sections(obo_file)
     obo_file.close()