v3.reads_py() never reads last cell

I am using the following to convert a python script to a notebook:

```python
from nbformat import v3, v4

with open("test.py") as fpin:
nbook = v3.reads_py(fpin.read())

nbook = v4.upgrade(nbook)

with open("test.ipynb", "w") as fpout:
fpout.write(v4.writes(nbook) + "\n")

```

However using the following `test.py` file

```python
#
# Title of the notebook

#
import numpy as np
import matplotlib.pyplot as plt

#
plt.plot()

#
#
# This is some markdown documentation text, bla
#
# - with a
# - list
#
#
#
# The last cell is never imported, so we need with this pointless cell

```

I realized that the last `<*cell>` is never read and never ends up in `test.ipynb`, so I have to add a dummy `` at the end to convert everything preceding it.

Start at the v3.reads_py() entry point and reproduce the issue with the test.py content shown in the report. The conversion should include the final marked cell in test.ipynb without requiring a dummy cell; run the example after making the change to verify the output.

Written by the indexing model from the issue text.