Dear author,

I have some questions about the ARL reader. After reading the ARL data, could we change it and draw new trajectory?

And before running, should we install numpy first? and install the ARL reader? Looking forward to your reply! Thanks!

Best Regards,
Chung

Hi Martin,

Bug Report

The 2-D data has very limited length after data extraction from certain level. It can be 2x2 if certain latitude and longitude is given (see below).

But the indexing is still going under the index of original laitude and longitude, which can be values between [0,180] and [0-360]. This will throw error.

Bug Fixes

Change line 931

val = data[0, 0]

I don't know whether this fix will trigger other issues (Please check it).
For my purpose of extraction GDAS1 profile, it works OK.

Greetings from China!

Solution:
Adding new lines from After currentDate, line 234
week_no_start = ((currentDate.day - 1) // 7) + 1
currentday_start = (week_no_start - 1) * 7 + 1

Firstly, I would like to thank you for such a useful library.

I have not been able to use the fast_funcs but after working with the library, I came with a performance upgrade in the "standard" unpack_data function that could be of your interest:

def unpack_data2(binarray, nx, ny, initval, exp, prec, stopat=(-1, -1)):
    """
    unpack the binary data with the option to stop at a given index
    """

    # Reshape the array
    binarray  = np.reshape(binarray,(ny,nx))

    # Calculate the array difference matrix
    e = (2 ** float(7 - exp))
    binarray = (binarray - 127) / e

    # Calculate the first row using the initial value
    first_row = np.cumsum(binarray[:, 0]) + initval

    # Asign the values to the array
    binarray[:, 0] = first_row

    # Calculate all the rows using the first row. Transpose in order to have an (nx,ny) array
    data = np.cumsum(binarray, axis=1).T

    data[np.abs(data) < prec] = 0

    if stopat != (-1, -1):
        nx_min = max(stopat[1] - 1, 0)
        nx_max = min(stopat[1] + 1, nx)
        ny_min = max(stopat[0] - 1, 0)
        ny_max = min(stopat[0] + 1, ny)
    else:
        nx_min = 0
        nx_max = nx
        ny_min = 0
        ny_max = ny

    return data[nx_min:nx_max, ny_min:ny_max]

The following screenshot shows the difference in the time required to load a single variable from a 0.25 gfs model:

Once again, thanks.

I think here should be
coord = (coord[0], coord[1]+360)

to change the range of longitude to (-180, 180)

Hi, thank you for this package.

I am using ARLreader to parse GFS 0.25 files that I downloaded from ARL's FTP server (ftp://arlftp.arlhq.noaa.gov/archives/gfs0p25/). When I try to load the profile for a location, I get an IndexError. Here's a simple example and the resultant error:

import ARLreader as Ar
gfs = Ar.reader('/path/to/file/20180222_gfs0p25')
profile, sfcdata, indexinfo, ind = gfs.load_profile(22, 3, (28.62, 77.22)) # Delhi

Error log:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\Users\username\Anaconda3\envs\arlenv\lib\site-packages\ARLreader\__init__.py", line 940, in load_profile
    calc_p_from_sigma(profile["SIGMA"], sfc_pres)
IndexError: tuple index out of range

On some digging, I find that the string formatter in logger.info() expects two return values from calc_p_from_sigma() which returns only one. An easy fix is to remove the second set of empty braces ({}) from line 938 in the snippet below.

Hey @ZPYin,

Is there a specific reason for the requirement of certain version numbers of the required packages?

I wish to use it with an existing python env I am already running (e.g. python > 3.8).

These are the installation steps I tried:

conda install altgraph atomicwrites attrs colorama Cython future importlib-metadata more-itertools packaging pefile pluggy py pyparsing pytest pywin32-ctypes six wcwidth wincertstore zipp numpy
pip install git+https://github.com/martin-rdz/ARLreader.git
However, the last command failed due to dependencies.

Then I solved it as follows:

1. Unhighlighting the versions number written in the requirments. E.g. altgraph==0.16.1 ---> altgraph # ==0.16.1
2. git clone https://github.com/martin-rdz/ARLreader.git ARLreader
3. cd ARLreader
4. python setup.py install

Currently, it seems that the installation worked and it's possible to import the package.

For future users, I guess that it will be better to drop the versions number requirements (unless you find it important for package functionality)

tnx

Adi