meteokid / python-rpn Goto Github PK

View Code? Open in Web Editor NEW

9.0 9.0 4.0 18.34 MB

Python RPN (RPNpy) is a collection of Python modules and scripts developed at RPN for scientific use.

License: GNU Lesser General Public License v2.1

Makefile 1.07% C 85.64% Shell 1.11% Fortran 1.43% Python 10.33% C++ 0.05% Pascal 0.38%

python-rpn's People

Contributors

Stargazers

Watchers

Forkers

guziy neishm wk1984 wobagi

python-rpn's Issues

Issue importing 'all'

While attempting

import rpnpy.librmn.all as rmn

I get the following traceback

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "~/git/python-rpn/lib/rpnpy/librmn/__init__.py", line 31, in <module>
    from rpnpy.version import *
ImportError: No module named version

Is this a missing package or have I set something up wrong?

Option to bypass fstprm when calling fstluk?

For the scripts I'm working on, I only need the data ('d') output from fstluk, not any of the record parameters from the header. Profiling indicates the call to fstprm takes 10-20% of the read time in my usage, a non-negligible amount of time. It would be useful (at least in my case) to have some optional argument to fstluk to bypass the fstprm call, e.g., to do something like

rec = fstluk(funit, key, fstprm=False)  # Skips fstprm, only has 'd' and 'shape' entries.

Is this something that could be feasibly added to the package?

As an aside: is this issue page the best place to put feature requests, or is the internal bugzilla the preferred medium?

Installation Outside CMC Intranet

To install python-rpn on a computer without access to the CMC intranet, I modifed installation steps. First, I executed http://scaweb.sca.uqam.ca/armnlib/repository/Linux_BOOTSTRAP-8.11-rmnlib15.2. Then I executed:

export SSM_DOMAIN_HOME=~/ssm-domains-base/setup/v_001
. ~/ECssm/etc/ssm.d/profile
. s.ssmuse.dot rmnlib-dev
export RPNPY_RMN_LIBPATH=~/ssm-domains-base/libs/rmnlib-dev/multi/lib/${EC_ARCH}/gfortran/
. ./.setenv.dot --dev

This allowed me to use fstopenall, fstinl and fstluk on a RPN Standard File. Should these steps be merged into https://github.com/meteokid/python-rpn/blob/master/README?

I also deleted these lines:
https://github.com/meteokid/python-rpn/blob/master/bin/.env_setup.dot#L68
https://github.com/meteokid/python-rpn/blob/master/ssmusedep-dev.bndl#L1

Problem with Bootstrap outside CMC

After installing the bootstrap suggested in a closed issue I get the following errors.

Global tag values:
ERROR: NO SHORTCUT FOUND FOR /ssm/net/hpcs/201402/02/base
ERROR: NO SHORTCUT FOUND FOR /ssm/net/hpcs/201402/02/intel13sp1u2
ERROR: NO SHORTCUT FOUND FOR /ssm/net/rpn/utils/15.2
ERROR: NO SHORTCUT FOUND FOR /ssm/net/rpn/libs/15.2
ERROR: NO SHORTCUT FOUND FOR /ssm/net/cmdn/vgrid/6.1.1/intel13sp1u2
ERROR: NO SHORTCUT FOUND FOR /ssm/net/hpcs/exp/aspgjdm/perftools
ERROR: NO SHORTCUT FOUND FOR /ssm/net/cmoi/base/20160901/
ERROR: NO SHORTCUT FOUND FOR ENV/rde/1.0.0
. .setenv.defaults.dot
./.setenv.dot: line 27: rdevar: command not found

Cannot write two records if only their date differ

When I write two records in sequence, and these records have the same metadata except for the date, the second write overwrites the first one.

The assert fails on ppp2 with 2.1.b3

import os
import numpy as np
import rpnpy.librmn.all as rmn

TARGET_FILE = '/tmp/test.fst'

if __name__ == '__main__':
    if os.path.isfile(TARGET_FILE):
        os.unlink(TARGET_FILE)

    phony_data = np.zeros((10,10,1), order='FORTRAN')

    grid = rmn.encodeGrid({
        'grtyp': 'L',
        'ni': phony_data.shape[0],
        'nj': phony_data.shape[1],
        'lat0': 0.0,
        'lon0': 0.0,
        'dlat': 1.0,
        'dlon': 1.0,
    })

    deet = 3600
    npas = 2
    record = grid
    record.update({
        'd': phony_data,
        'etiket': 'TEST',
        'nomvar': 'XX',
        'datev': rmn.newdate(rmn.NEWDATE_PRINT2STAMP, 20180302, 20000000),
        'dateo': rmn.newdate(rmn.NEWDATE_PRINT2STAMP, 20180302, 18000000),
        'deet': deet,
        'npas': npas,
    })

    target_fst = rmn.fstopenall(TARGET_FILE, rmn.FST_RW)
    rmn.fstecr(target_fst, record)

    record['datev'] = rmn.newdate(rmn.NEWDATE_PRINT2STAMP, 20190302, 20000000)
    record['dateo'] = rmn.newdate(rmn.NEWDATE_PRINT2STAMP, 20190302, 18000000)
    rmn.fstecr(target_fst, record)

    n_records = rmn.fstnbr(target_fst)
    rmn.fstcloseall(target_fst)

    assert(n_records == 2)

fstopenall not linking files properly

When using fstopenall with an array of standard files, records beyond the first file are not accessible. I wrote a little reproduction script. Can you guys reproduce? I found this issue on PPP2 and my local machine.

import pathlib

import rpnpy.librmn.all as rmn

FST_PATH = pathlib.Path('/space/hall2/sitestore/eccc/cmod/prod/hubs/gridpt/dbase/caldas/hrdps_national/analysis/final/diag')

if __name__ == '__main__':
    fsts = list((str(x) for x in FST_PATH.iterdir() if rmn.isFST(str(x))))[0:2]

    print('Running test on two standard files: {}'.format(fsts))

    first_fst = rmn.fstopenall(fsts[0])
    first_fst_n_records = rmn.fstnbr(first_fst)
    rmn.fstcloseall(first_fst)

    second_fst = rmn.fstopenall(fsts[1])
    second_fst_n_records = rmn.fstnbr(second_fst)
    rmn.fstcloseall(second_fst)

    merged_fst = rmn.fstopenall(fsts)
    merged_fst_n_records = rmn.fstnbr(merged_fst)
    rmn.fstcloseall(merged_fst)

    # The following assert fails with rpnpy 2.1.b3 on ppp2
    assert(merged_fst_n_records == first_fst_n_records + second_fst_n_records)

librmn no longer has crc32 method

See: ECCC-ASTD-MRD/librmn@686e512

If Python-RPN is used with librmn 016.3, it immediately fails with an AttributeError from ctypes, due to a missing crc32 symbol.

cleanName

Salut,

dans la routines
bin/rpy.nml_set
bin/rpy.nml_get
Il y a un appel a "cleanname" or cleanname n est declarer nulle part.
Ca mene donc a une erreur.

Dans la cas de bin/rpy.nml_set, ne suffirai t'il pas de commenter la ligne (sa semble un oublie?). En tout cas, cette fonction redevient fonctionnel une fois cette ligne commentée.

Merci

Possibly useful utility functions?

I have some optimized routines from the fstd2nc tool which I've found helpful when scanning through many (hundreds) of files on a routine basis, where even a small overhead can add up to a noticeable delay in the script. Maybe some of them might be useful within python-rpn as utility functions? Below is a brief description of them.

all_params

The all_params method extracts all the record parameters at once, and returns a vectorized dictionary of the result. It scrapes the information directly out of the librmn data structures, and avoids the overhead of repeatedly calling fstprm. Example:

In [1]: import os, sys

In [2]: import rpnpy.librmn.all as rmn

In [3]: from fstd2nc.extra import all_params

In [4]: ATM_MODEL_DFILES = os.getenv('ATM_MODEL_DFILES').strip()

In [5]: fileId = rmn.fstopenall(ATM_MODEL_DFILES+'/bcmk', rmn.FST_RO)

In [6]: p = all_params(fileId)

In [7]: print p
{'deet': array([900, 900, 900, ..., 900, 900, 900], dtype=int32), 'typvar': array(['P ', 'P ', 'P ', ..., 'P ', 'P ', 'X '], dtype='|S2'), 'lng': array([3762, 3762, 3762, ..., 3762, 3762,   13], dtype=int32), 'ni': array([200, 200, 200, ..., 200, 200,   1], dtype=int32), 'nj': array([100, 100, 100, ..., 100, 100,   1], dtype=int32), 'nbits': array([12, 12, 12, ..., 12, 12, 32], dtype=int8), 'swa': array([   2335,    6097,    9859, ..., 4040669, 4044431, 4048193],
      dtype=uint32), 'datyp': array([1, 1, 1, ..., 1, 1, 5], dtype=uint8), 'xtra1': array([354514400, 354514400, 354514400, ..., 354525200, 354525200,
       354525200], dtype=uint32), 'xtra2': array([0, 0, 0, ..., 0, 0, 0], dtype=uint32), 'xtra3': array([0, 0, 0, ..., 0, 0, 0], dtype=uint32), 'ip2': array([ 0,  0,  0, ..., 12, 12,  0], dtype=int32), 'ip3': array([0, 0, 0, ..., 0, 0, 0], dtype=int32), 'ip1': array([       0, 97642568, 97738568, ..., 93423264, 93423264, 44140192],
      dtype=int32), 'key': array([      1,    1025,    2049, ..., 2145288, 2146312, 2147336],
      dtype=int32), 'ubc': array([0, 0, 0, ..., 0, 0, 0], dtype=uint16), 'npas': array([ 0,  0,  0, ..., 48, 48, 48], dtype=int32), 'nk': array([1, 1, 1, ..., 1, 1, 1], dtype=int32), 'ig4': array([0, 0, 0, ..., 0, 0, 0], dtype=int32), 'ig3': array([0, 0, 0, ..., 0, 0, 0], dtype=int32), 'ig2': array([   0,    0,    0, ...,    0,    0, 1600], dtype=int32), 'ig1': array([  0,   0,   0, ...,   0,   0, 800], dtype=int32), 'nomvar': array(['P0  ', 'TT  ', 'TT  ', ..., 'UU  ', 'VV  ', 'HY  '], dtype='|S4'), 'datev': array([354514400, 354514400, 354514400, ..., 354525200, 354525200,
       354525200], dtype=int32), 'dateo': array([354514400, 354514400, 354514400, ..., 354514400, 354514400,
       354514400], dtype=int32), 'etiket': array(['G133K80P    ', 'G133K80P    ', 'G133K80P    ', ...,
       'G133K80P    ', 'G133K80P    ', 'G133K80P    '], dtype='|S12'), 'grtyp': array(['G', 'G', 'G', ..., 'G', 'G', 'X'], dtype='|S1'), 'dltf': array([0, 0, 0, ..., 0, 0, 0], dtype=uint8)}

This can be combined with pandas to get a convenient table of parameters:

In [8]: import pandas as pd

In [9]: p = pd.DataFrame(p)

In [10]: print p
          dateo      datev  datyp  deet  dltf  ...   typvar ubc      xtra1  xtra2  xtra3
0     354514400  354514400      1   900     0  ...       P    0  354514400      0      0
1     354514400  354514400      1   900     0  ...       P    0  354514400      0      0
2     354514400  354514400      1   900     0  ...       P    0  354514400      0      0
3     354514400  354514400      1   900     0  ...       P    0  354514400      0      0
4     354514400  354514400      1   900     0  ...       P    0  354514400      0      0
...         ...        ...    ...   ...   ...  ...      ...  ..        ...    ...    ...
2783  354514400  354525200      1   900     0  ...       P    0  354525200      0      0
2784  354514400  354525200      1   900     0  ...       P    0  354525200      0      0
2785  354514400  354525200      1   900     0  ...       P    0  354525200      0      0
2786  354514400  354525200      1   900     0  ...       P    0  354525200      0      0
2787  354514400  354525200      5   900     0  ...       X    0  354525200      0      0

[2788 rows x 28 columns]

Having the parameters in this pandas.DataFrame structure provides a more powerful tool for analysing the data. For instance, the pivot method could be used to quickly organize the records into multidimensional time/level structures.
The method is also about 20x faster than looping over fstprm:

In [11]: %timeit map(rmn.fstprm, rmn.fstinl(fileId))
10 loops, best of 3: 84.3 ms per loop

In [12]: %timeit pd.DataFrame(all_params(fileId))
100 loops, best of 3: 4.46 ms per loop

80ms isn't much, but it adds up if you're scanning over hundreds (or thousands) of files.

maybeFST

The maybeFST function is a more compact version of isFST which avoids any librmn calls such as c_wkoffit. That library function can incur some overhead since it's testing for many different formats, not just FST.

When combined together, the maybeFST and all_params functions can allow a user to very quickly scan over many hundreds of files and get a snapshot of all the records inside.

stamp2datetime

The stamp2datetime function converts an array of RPN date stamps into datetime objects. Useful in conjunction with all_params to get date information quickly, e.g.:

In [13]: from fstd2nc.mixins.dates import stamp2datetime

In [14]: print stamp2datetime(p['datev'])
['2009-04-27T00:00:00.000000000' '2009-04-27T00:00:00.000000000'
 '2009-04-27T00:00:00.000000000' ... '2009-04-27T12:00:00.000000000'
 '2009-04-27T12:00:00.000000000' '2009-04-27T12:00:00.000000000']

In [15]: p['date'] = stamp2datetime(p['datev'])

In [16]: print p
          dateo      datev  datyp  deet         ...              xtra1 xtra2 xtra3                date
0     354514400  354514400      1   900         ...          354514400     0     0 2009-04-27 00:00:00
1     354514400  354514400      1   900         ...          354514400     0     0 2009-04-27 00:00:00
2     354514400  354514400      1   900         ...          354514400     0     0 2009-04-27 00:00:00
3     354514400  354514400      1   900         ...          354514400     0     0 2009-04-27 00:00:00
4     354514400  354514400      1   900         ...          354514400     0     0 2009-04-27 00:00:00
...         ...        ...    ...   ...         ...                ...   ...   ...                 ...
2783  354514400  354525200      1   900         ...          354525200     0     0 2009-04-27 12:00:00
2784  354514400  354525200      1   900         ...          354525200     0     0 2009-04-27 12:00:00
2785  354514400  354525200      1   900         ...          354525200     0     0 2009-04-27 12:00:00
2786  354514400  354525200      1   900         ...          354525200     0     0 2009-04-27 12:00:00
2787  354514400  354525200      5   900         ...          354525200     0     0 2009-04-27 12:00:00

[2788 rows x 29 columns]

decode_ip1

The decode_ip1 function quickly decodes the levels from an array of ip1 values. For example:

In [18]: import numpy as np

In [19]: from fstd2nc.mixins.vcoords import decode_ip1

In [20]: print decode_ip1(p['ip1'])
[array([(2, 0.)], dtype=[('kind', '<i4'), ('level', '<f4')])
 array([(5, 0.000125)], dtype=[('kind', '<i4'), ('level', '<f4')])
 array([(5, 0.000221)], dtype=[('kind', '<i4'), ('level', '<f4')]) ...
 array([(5, 1.)], dtype=[('kind', '<i4'), ('level', '<f4')])
 array([(5, 1.)], dtype=[('kind', '<i4'), ('level', '<f4')])
 array([(2, 0.1)], dtype=[('kind', '<i4'), ('level', '<f4')])]

In [21]: levels = np.concatenate(decode_ip1(p['ip1']))

In [22]: print levels
[(2, 0.00e+00) (5, 1.25e-04) (5, 2.21e-04) ... (5, 1.00e+00) (5, 1.00e+00)
 (2, 1.00e-01)]

In [23]: print levels.dtype
[('kind', '<i4'), ('level', '<f4')]

In [24]: p['level'] = levels['level']

In [25]: p['kind'] = levels['kind']

In [26]: print p
          dateo      datev  datyp  deet  dltf  ...  xtra2 xtra3                date     level  kind
0     354514400  354514400      1   900     0  ...      0     0 2009-04-27 00:00:00  0.000000     2
1     354514400  354514400      1   900     0  ...      0     0 2009-04-27 00:00:00  0.000125     5
2     354514400  354514400      1   900     0  ...      0     0 2009-04-27 00:00:00  0.000221     5
3     354514400  354514400      1   900     0  ...      0     0 2009-04-27 00:00:00  0.000382     5
4     354514400  354514400      1   900     0  ...      0     0 2009-04-27 00:00:00  0.000635     5
...         ...        ...    ...   ...   ...  ...    ...   ...                 ...       ...   ...
2783  354514400  354525200      1   900     0  ...      0     0 2009-04-27 12:00:00  0.995000     5
2784  354514400  354525200      1   900     0  ...      0     0 2009-04-27 12:00:00  0.995000     5
2785  354514400  354525200      1   900     0  ...      0     0 2009-04-27 12:00:00  1.000000     5
2786  354514400  354525200      1   900     0  ...      0     0 2009-04-27 12:00:00  1.000000     5
2787  354514400  354525200      5   900     0  ...      0     0 2009-04-27 12:00:00  0.100000     2

[2788 rows x 31 columns]

Example

For completion, here's an example using the information from all the above steps to get multidimensional structures for a field. First, get the list of variables:

In [32]: print pd.unique(p.nomvar)
['P0  ' 'TT  ' 'ES  ' 'HU  ' 'MX  ' 'LA  ' 'LO  ' 'ME  ' 'WE  ' 'GZ  '
 'HR  ' 'WW  ' 'PN  ' 'TD  ' 'QC  ' 'FC  ' 'FQ  ' 'IO  ' 'OL  ' 'UE  '
 'EN  ' 'GL  ' 'LH  ' 'MF  ' 'MG  ' 'TM  ' 'VG  ' 'RC  ' 'AL  ' 'I0  '
 'I1  ' 'SD  ' 'I6  ' 'I7  ' 'I8  ' 'I9  ' 'AB  ' 'AG  ' 'AH  ' 'AI  '
 'AP  ' 'AR  ' 'AS  ' 'AU  ' 'AV  ' 'AW  ' 'EV  ' 'FS  ' 'FV  ' 'IC  '
 'IE  ' 'IH  ' 'IV  ' 'NF  ' 'SI  ' 'I2  ' 'I3  ' 'I4  ' 'I5  ' 'DN  '
 'NR  ' 'HF  ' 'FL  ' 'N0  ' 'O1  ' 'RT  ' 'RY  ' 'PR  ' 'PE  ' 'FR  '
 'RN  ' 'SN  ' 'PC  ' 'FB  ' 'N4  ' 'CX  ' 'FI  ' 'AD  ' 'FN  ' 'EI  '
 'H   ' 'J9  ' 'TG  ' 'Z0  ' 'NT  ' 'K6  ' 'K4  ' 'U9  ' 'AE  ' 'RZ  '
 'RR  ' 'U4  ' 'L7  ' 'L8  ' 'IY  ' 'UU  ' 'VV  ' 'HY  ' '>>  ' '^^  '
 'VF  ' 'GA  ' 'J1  ' 'J2  ' 'Y7  ' 'Y8  ' 'Y9  ' 'ZP  ' 'META' 'TS  '
 'TP  ' 'HS  ' 'LG  ' 'ICEL' 'EMIB']

Pick UU:

In [46]: p = p.loc[p.dltf==0]   # Ignore deleted records

In [47]: uu = p.loc[p.nomvar=='UU  ']  # Select UU records

In [48]: print uu
          dateo      datev  datyp  deet  dltf  ...  xtra2 xtra3                date     level  kind
922   354514400  354514400      1   900     0  ...      0     0 2009-04-27 00:00:00  0.000125     5
924   354514400  354514400      1   900     0  ...      0     0 2009-04-27 00:00:00  0.000221     5
926   354514400  354514400      1   900     0  ...      0     0 2009-04-27 00:00:00  0.000382     5
928   354514400  354514400      1   900     0  ...      0     0 2009-04-27 00:00:00  0.000635     5
930   354514400  354514400      1   900     0  ...      0     0 2009-04-27 00:00:00  0.001010     5
...         ...        ...    ...   ...   ...  ...    ...   ...                 ...       ...   ...
2777  354514400  354525200      1   900     0  ...      0     0 2009-04-27 12:00:00  0.961000     5
2779  354514400  354525200      1   900     0  ...      0     0 2009-04-27 12:00:00  0.974000     5
2781  354514400  354525200      1   900     0  ...      0     0 2009-04-27 12:00:00  0.985000     5
2783  354514400  354525200      1   900     0  ...      0     0 2009-04-27 12:00:00  0.995000     5
2785  354514400  354525200      1   900     0  ...      0     0 2009-04-27 12:00:00  1.000000     5

[160 rows x 31 columns]

Organize the records by date/level:

In [49]: uu = uu.pivot(index='date',columns='level')  # Organize by date/level

In [51]: print uu['ip1']
level                0.000125  0.000221  0.000382    ...     0.985000  0.995000  1.000000
date                                                 ...                                 
2009-04-27 00:00:00  97642568  97738568  97899568    ...     95356840  95366840  93423264
2009-04-27 12:00:00  97642568  97738568  97899568    ...     95356840  95366840  93423264

[2 rows x 80 columns]

In [52]: print uu['ip2']
level                0.000125  0.000221  0.000382    ...     0.985000  0.995000  1.000000
date                                                 ...                                 
2009-04-27 00:00:00         0         0         0    ...            0         0         0
2009-04-27 12:00:00        12        12        12    ...           12        12        12

[2 rows x 80 columns]

In [53]: print uu['key']  # The keys/handles for these record
level                0.000125  0.000221  0.000382    ...     0.985000  0.995000  1.000000
date                                                 ...                                 
2009-04-27 00:00:00   1730561   1732609   1734657    ...      2150401   2152449   2154497
2009-04-27 12:00:00   1721352   1723400   1725448    ...      2141192   2143240   2145288

[2 rows x 80 columns]

Librmn Undefined Symbol

I am trying to use Python-RPN outside the CMC Intranet, again. To produce librmnshared_015.2.so, I download and execute http://scaweb.sca.uqam.ca/armnlib/repository/Linux_BOOTSTRAP-8.11-rmnlib15.4.

When I try import rpnpy.librmn.all as rmn, Python prints:
OSError: /home/konu/ssm-domains-base/libs/rmnlib-dev/rmnlib-git_1.0_multi/lib/Linux_x86-64/gfortran/librmnshared_015.2.so: undefined symbol: get_gossip_dir

The complete output of the Linux_BOOTSTRAP-8.11-rmnlib15.4 is here: https://gist.github.com/jeixav/ca93310d9b23d4c20a05ce5cdb1d3c97. While the script appears to run to completion, there are a few errors. Does anyone have suggestions to troubleshoot?

The operating system is Debian Stretch

Some elements not being decoded in rpnpy.burpc?

It appears that some types of burp elements aren't being decoded in rpnpy.burpc. Instead, the e_rval array holds some random values (from uninitialized memory?).

Example of triggering this problem:

import rpnpy.librmn.all as rmn
import rpnpy.burpc.all as brp
import os
mypath = os.path.join(os.environ['ATM_MODEL_DFILES'],'bcmk_burp','2007021900.brp')
with brp.BurpcFile(mypath) as bfile:
  rpt = bfile[0]
  blk = rpt[1]
  for i, ele in enumerate(blk):
    print 'ELE', i
    print ele

When I run this test on the science network, the "ELE 6" e_rval changes every time I run the script. The only differences I can see with this element is it has e_cvt set to 0 (no conversion?). This issue is present in both 2.1.b3 and on the latest master branch.

This was causing sporadic failures for me in the test_brp_blk_getele_iter unit test - if the garbage values happen to have a nan in there, then the == comparison would fail. This only happened once in a while, so it took a while to figure out why tests that were working suddenly failed... fun times.

Unable to open capital letter files with fstopenall

If I have some standard file XYZ, I am unable to open it with fstopenall. Sample code:

>>> f = rmn.fstopenall('XYZ', verbose=True)
(fstopenall) Not a RPNSTD file: XYZ
(fstopenall) Problem Opening: XYZ
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "rpnpy/librmn/fstd98.py", line 304, in fstopenall
    .format(str(paths)))
rpnpy.librmn.fstd98.FSTDError: fstopenall: unable to open any file in path ['XYZ']

Tested with 2.1.b2.

Installation outside CMC intranet, again

Running units tests external to CMC

So I'm back to trying to install python-rpn on a CMC external computer. After following the steps outlined in Issue #2, with the following changes:

removed all lines in https://github.com/meteokid/python-rpn/blob/master/ssmusedep-dev.bndl
removed line 71 rather than 68 https://github.com/meteokid/python-rpn/blob/master/bin/.env_setup.dot#L71

When running the tests I encounter the following issues:

The directories are not accessible externally so the tests are not portable:

/home/ordenv/ssm-domains9/release/gem-data_4.2.0/gem-data_4.2.0_all/...
/users/dor/armn/env/SsmBundles/GEM/d/gem-data/gem-data_4.2.0/gem-data_4.2.0_all/...

The following three tests fail independantly (I think) of the missing files.

======================================================================
ERROR: test_14 (__main__.rpnpyCookbook)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "share/tests/test_cookbook.py", line 342, in test_14
    import rpnpy.vgd.all as vgd
  File "/home/doyle/git/rpnpy/lib/rpnpy/vgd/__init__.py", line 102, in <module>
    (VGD_VERSION, VGD_LIBPATH, libvgd) = loadVGDlib()
  File "/home/doyle/git/rpnpy/lib/rpnpy/vgd/__init__.py", line 90, in loadVGDlib
    raise IOError(-1, 'Failed to find libdescrip.so: ', vgd_libfile)
IOError: [Errno -1] Failed to find libdescrip.so: : 'libdescripshared_6.0.0.so'

======================================================================
ERROR: test_22 (__main__.rpnpyCookbook)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "share/tests/test_cookbook.py", line 608, in test_22
    sys.stderr.write("Problem opening the files: %s, %s\n" % (fileNameIn, fileNameOut))
NameError: global name 'sys' is not defined

======================================================================
ERROR: testfnomfclosKnownValues (__main__.LibrmnFilesKnownValues)
fnomfclos should give known result with known input
----------------------------------------------------------------------
Traceback (most recent call last):
  File "share/tests/test_librmn_base.py", line 33, in testfnomfclosKnownValues
    iout  = rmn.fnom(mypath,rmn.FST_RO)
  File "/home/doyle/git/rpnpy/lib/rpnpy/librmn/base.py", line 147, in fnom
    raise RMNBaseError()
RMNBaseError