---------------------------
note: these test files are not redistributed with this library
so a todo for myself is to create a script outside this python
module to automatically test on all these files 
---------------------------
Todo: see if it is usefull to add a requirements.txt  file.
See: https://caremad.io/blog/setup-vs-requirement/ for details.
---------------------------
# tested ./bufr_count_msgs.py on these bufr files:
# ascat/* [works fine]
# qscat/* [works fine]
# ers12/* [works fine]
# MSS*    [works fine]
# synop/* [works fine]
---------------------------
# tested ./bufr_to_ascii.py on these bufr files:

see my private ./BUFR_test_all_files.py script.

results for files in pybufr_ecmwf/ecmwf_bufr_lib/000401/data:
Totals:
count_passed =  140
count_failed =  0

results for files in BUFR_test_files:
Totals:
count_passed =  78
count_failed =  19

problematic files are the sn.00*.bin files from noaa (15 files)
because I don't have the matching BUFR tables
BUFR_test_files/noaa_mos/sn.0001.bin
exit_status =  1
['KeyError: 2200', '']

and a few rader BUFR files, fail for the same reason

Finally a qscat file fails with a key error:
BUFR_test_files/qscat/QS_D07060_S1821_E2002_B4009191_025.genws
exit_status =  1
['KeyError: 312221', '']

---------------------------
todo: add a little script to allow easy compilation and use of the
      synop2bufr tool

This sequence of commands works for me:

csh
cd pybufr_ecmwf/ecmwf_bufrtables
setenv BUFR_TABLES `pwd`
ln -s B_default.TXT B0000000000000014000.TXT
ln -s D_default.TXT D0000000000000014000.TXT
cd ../ecmwf_bufr_lib/bufr_000385/synop2bufr/
setenv PP_BASE `pwd`
./synop2bufr -i data/greek.gts -o data/greek.gts.bufr -c 96

this generates an example file greek.gts.bufr
from the provided greek.gts example input file.

Note that the provided file greek.gts.bufr.gts
is the same data converted to BUFR, but now the gts header is still 
present in front of each BUFr message, which makes the file a bit larger.
---------------------------

