sagemathgh-38833: Respect sparse=True in vector constructor when passed ndarray
    
Self-explanatory.

(Of course user would not normally do this but… just in case.)

### 📝 Checklist


- [x] The title is concise and informative.
- [x] The description explains in detail what this PR is about.
- [x] I have linked a relevant issue or discussion.
- [x] I have created tests covering the changes.
- [x] I have updated the documentation and checked the documentation
preview.
    
URL: sagemath#38833
Reported by: user202729
Reviewer(s): Kwankyu Lee

Author: Release Manager

src/sage/modules/free_module_element.pyx

@@ -486,6 +486,16 @@ def vector(arg0, arg1=None, arg2=None, sparse=None, immutable=False):
         sage: vector(S, 3)                                                              # needs sage.rings.finite_rings
         ...
         (0, 0, 0)
+
+    We check that ``sparse`` is respected for numpy arrays::
+
+        sage: # needs numpy
+        sage: import numpy
+        sage: a = numpy.array([1,2,3], dtype=numpy.float64)
+        sage: v = vector(a, sparse=True); v
+        (1.0, 2.0, 3.0)
+        sage: v.is_sparse()
+        True
     """
     from sage.modules.free_module import FreeModule
     # We first efficiently handle the important special case of the zero vector
@@ -563,7 +573,7 @@ def vector(arg0, arg1=None, arg2=None, sparse=None, immutable=False):
     except ImportError:
         pass
     else:
-        if isinstance(v, ndarray):
+        if isinstance(v, ndarray) and not sparse:
             if len(v.shape) != 1:
                 raise TypeError("cannot convert %r-dimensional array to a vector" % len(v.shape))
             from sage.modules.free_module import VectorSpace