ValueError When Running pymsis.calculate with Documentation Example: Cannot Convert Empty String to int32

[ssmong]

SW-All.csv
Issue Summary:
Executing the example code from the PyMSIS documentation results in a ValueError. The error message indicates that an empty string cannot be converted to an int32 at a specific row and column. This issue prevents users from successfully using the pymsis.calculate function as demonstrated in the documentation.

Steps to Reproduce:

1. Use the following code (identical to the documentation example):

   import pymsis
   import numpy as np
   
   dates = np.arange(
       np.datetime64("2003-10-28T00:00"),
       np.datetime64("2003-11-04T00:00"),
       np.timedelta64(30, "m")
   )
   # geomagnetic_activity=-1 is a storm-time run
   data = pymsis.calculate(dates, 0, 0, 400, geomagnetic_activity=-1)

2. Run the code in your Python environment.

Expected Behavior:
The pymsis.calculate function should execute without errors and return the calculated atmospheric data based on the provided dates and parameters.

Actual Behavior:
A ValueError is raised during execution:

ValueError: could not convert string '' to int32 at row 24531, column 13.

Error Traceback:

/usr/local/lib/python3.10/dist-packages/pymsis/utils.py:45: UserWarning: Downloading ap and F10.7 data from https://celestrak.org/SpaceData/SW-All.csv
  warnings.warn(f"Downloading ap and F10.7 data from {_F107_AP_URL}")
/usr/local/lib/python3.10/dist-packages/pymsis/utils.py:45: UserWarning: Downloading ap and F10.7 data from https://celestrak.org/SpaceData/SW-All.csv
  warnings.warn(f"Downloading ap and F10.7 data from {_F107_AP_URL}")
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-41-161cc04f4b20> in <cell line: 4>()
      2 dates = np.arange(np.datetime64("2003-10-28T00:00"), np.datetime64("2003-11-04T00:00"), np.timedelta64(30, "m"))
      3 # geomagnetic_activity=-1 is a storm-time run
----> 4 data = pymsis.calculate(dates, 0, 0, 400, geomagnetic_activity=-1)

5 frames
/usr/local/lib/python3.10/dist-packages/numpy/lib/npyio.py in _read(fname, delimiter, comment, quote, imaginary_unit, usecols, skiplines, max_rows, converters, ndmin, unpack, dtype, encoding)
   1014 
   1015         if read_dtype_via_object_chunks is None:
-> 1016             arr = _load_from_filelike(
   1017                 data, delimiter=delimiter, comment=comment, quote=quote,
   1018                 imaginary_unit=imaginary_unit,

ValueError: could not convert string '' to int32 at row 24531, column 13.

Environment Details:

• Python Version: 3.11.9 (VSCode), 3.10.12 (Colab)
• PyMSIS Version: 0.10.0
• Operating System: Window 11 - VSCode and Colab

Attached file is the one

[mananapr]

This is an issue at Celestrak's end where they have published empty data under observed values.
I suppose pymsis can add a step where it discards rows with empty values if found under observed data.

EDIT: Looks like celestrak has fixed their CSV. @ssmong - Try re-running your code with the latest CSV

[ssmong]

Thanks for your support!

I just re-ran it, and it worked perfectly fine
Always appreciate your awesome library

[greglucas]

I think we can handle this in the library a little more gracefully too. See #72 for an attempt at skipping those bad lines.

[mananapr]

I think we can handle this in the library a little more gracefully too. See #72 for an attempt at skipping those bad lines.

Hi @greglucas, is there a reason we haven't merged the PR yet?