Fix unsafe loading from file buffer (#58)

* Fix unsafe loading from file buffer The numpy documentation states for numpy.frombuffer` that: > This function creates a view into the original object. > This should be safe in general, but it may make sense to copy > the result when the original object is mutable or untrusted. When the filebuffer is mutable a `ValueError: output array is read-only` is raised in: https://github.com/eyurtsev/fcsparser/blob/701000af178e36e0dedb53d409119c25675b9d16/fcsparser/api.py#L590 * Update pyproject.toml --------- Co-authored-by: Eugene Yurtsev <[email protected]>
eyurtsev · Sep 6, 2023 · 55235c7 · 55235c7
1 parent e3672ff
commit 55235c7
Showing 1 changed file with 6 additions and 1 deletion.
diff --git a/fcsparser/api.py b/fcsparser/api.py
@@ -68,13 +68,18 @@ def fromfile(file, dtype, count, *args, **kwargs):
             file, dtype=",".join(["u1"] * record_width), count=count, *args, **kwargs
         )
     except (TypeError, IOError):
-        ret = numpy.frombuffer(
+        _ret = numpy.frombuffer(
             file.read(count * record_width),
             dtype=",".join(["u1"] * record_width),
             count=count,
             *args,
             **kwargs
         )
+        # Create a copy of the file content as `numpy.frombuffer`
+        # returns a view into the original object which is not 
+        # safe for mutable file buffers.
+        # See https://numpy.org/doc/stable/reference/generated/numpy.frombuffer.html
+        ret = _ret.copy()
 
     # convert the DATA segment from a 1 x `count` array of records
     # (and remember, each record is composed of `record_width`