feat(bruker): handle encoding/decode errors resiliently during JCAMP reading

Author: harivyasi

nmrglue/fileio/bruker.py

@@ -2174,10 +2174,13 @@ def read_jcamp(filename, encoding=locale.getpreferredencoding()):
     """
     dic = {"_coreheader": [], "_comments": []}  # create empty dictionary
 
-    with open(filename, 'r', encoding=encoding) as f:
+    with open(filename, 'r', encoding=encoding, errors='replace') as f:
         while True:     # loop until end of file is found
-
-            line = f.readline().rstrip()    # read a line
+            try:
+                line = f.readline().rstrip()    # read a line
+            except Exception as e:
+                warn(f"Unable to read line in file {filename}: {e}. Treating it as comment.")
+                line = "$$"
             if line == '':      # end of file found
                 break
 

nmrglue/fileio/tests/test_bruker.py

@@ -80,3 +80,21 @@ def test_write_pdata():
     assert np.all(data == rdata)
     assert rdic['procs'].keys() == dic['procs'].keys()
     shutil.rmtree(td)
+
+
+def test_read_jcamp_decoding_error():
+    """Testing read_jcamp handling of decoding errors"""
+    fd, temp_path = tempfile.mkstemp()
+    try:
+        with os.fdopen(fd, 'wb') as f:
+            f.write(b"##$KEY=value\n")
+            f.write(b"##$BAD=\xff\xfe\xfd\n")  # Invalid UTF-8 sequence
+            f.write(b"##END=\n")
+        
+        # This should read it and treat the bad characters by replacing them, not crash
+        dic = ng.bruker.read_jcamp(temp_path, encoding='utf-8')
+        assert dic["KEY"] == "value"
+        assert "\ufffd" in dic["BAD"]
+    finally:
+        os.remove(temp_path)
+