Don't force a datatype on widget values

[GeorgelPreput]

Exporting a sheet back to a Pandas dataframe will fail if the cell is represented through a custom ipywidget, if the data type of the underlying value is not float.

For example, when creating a sheet that has a column as such:

sample_sheet = sheet(columns=1)

sample_col = column(0, [
    widgets.Dropdown(
        options=["WARN", "DROP", "FAIL", ""],
        value="WARN"
    )
])

df = to_dataframe(sample_sheet)

You'll get the following stack trace:

---

ValueError Traceback (most recent call last) Cell In[8], line 10 1 sample_sheet = sheet(columns=1, rows=1) 3 sample_col = column(0, [ 4 widgets.Dropdown( 5 options=["WARN", "DROP", "FAIL", ""], 6 value="WARN" 7 ) 8 ]) ---> 10 df = to_dataframe(sample_sheet)

File ~/Source/sample/.venv/lib/python3.11/site-packages/ipysheet/pandas_loader.py:128, in to_dataframe(sheet) 124 else: 125 row_headers = list(sheet.row_headers) 127 return pd.DataFrame( --> 128 { 129 header: _extract_column(data, idx) 130 for idx, header in enumerate(column_headers) 131 }, 132 index=row_headers, 133 columns=column_headers 134 )

File ~/Source/sample/.venv/lib/python3.11/site-packages/ipysheet/pandas_loader.py:129, in (.0) 124 else: 125 row_headers = list(sheet.row_headers) 127 return pd.DataFrame( 128 { --> 129 header: _extract_column(data, idx) 130 for idx, header in enumerate(column_headers) 131 }, 132 index=row_headers, 133 columns=column_headers 134 )

File ~/Source/sample/.venv/lib/python3.11/site-packages/ipysheet/pandas_loader.py:86, in _extract_column(data, idx) 84 return np.array(d, dtype='M') 85 elif type == 'widget': ---> 86 return np.array([wid.value for wid in arr], dtype='f') 87 else: 88 return np.array(arr)

ValueError: could not convert string to float: 'WARN'

This commit removes the explicit dtype.