pip install https://github.com/modflowpy/pymake/zipball/master gives an error in a clean conda env.

Needs numpy and requests to call from pymake.pymake import version in setup.py

Please find attached a hacked version of the pymake.py file to use the mingw64 toolchain for compiling a windows x64.exe version of MODFLOW using Linux. Here is the general reference for how the mingw64 gfortran differs from standard gfortran (it also has options for the 32-bit versions):

https://sourceforge.net/p/mingw-w64/wiki2/UsingLinuxBinaries/#using-the-cross-compilers-to-generate-windows-32-bit-and-64-bit-binary-executables

Basically, in addition to the prefix for both the gfortran and gcc, in the linker phase needs the -static option, as well as replace the -lc option with either -m64 or -m32.

Don't currently have a clean patch for you.

pymake_w64.py.txt

Hi. I've compiled MODFLOW 2005 using pymake succesfully, but when I run modflow (using the example model files) I get the following message:

/MF2005.1_11u/test-run/mf2005dbl ; exit;

                              MODFLOW-2005     
U.S. GEOLOGICAL SURVEY MODULAR FINITE-DIFFERENCE GROUND-WATER FLOW MODEL
                         Version 1.11.00 8/8/2013                        

Enter the name of the NAME FILE:
mf2005.1_11u/test-run/bcf2ss
Using NAME file: mf2005.1_11u/test-run/bcf2ss.nam
Run start date and time (yyyy/mm/dd hh:mm:ss): 2015/09/22 22:36:36

*** ERROR OPENING FILE "bcf2ss.ba6" ON UNIT 8
SPECIFIED FILE STATUS: OLD
SPECIFIED FILE FORMAT: FORMATTED
SPECIFIED FILE ACCESS: SEQUENTIAL
SPECIFIED FILE ACTION: READ
-- STOP EXECUTION (SGWF2BAS7OPEN)
logout

[Process completed]

Does anyone have any suggestions for what causes this error message? The compile process ran with 6 warnings, similar to:

./src_temp/mf2kgmg.c:164:29: warning: cast to 'MF2KGMG_operator ' (aka 'struct
MF2KGMG_operator *') from smaller integer type 'int'
[-Wint-to-pointer-cast]
MF2KGMG_operator GMG_ptr=(MF2KGMG_operator_)_GMGID;

I've also tried compiling using the makefile supplied with the source code, but I end up with the same error message from the compiled MODFLOW executable. I'm using a macbook pro running OSX10.10.

Many thanks.

I successfully compiled mfnwt on the mac using pymake. I only encountered one issue halting the compilation process:

./src_temp/gwf2swr7_NWT.f:14361.22:

    dt = SECNDS(T1)                                               
                          1

Error: 'x' argument of 'secnds' intrinsic at (1) must be of kind 4

I fixed this by changing the following code in gwf2swr7_NWT.f:

DOUBLEPRECISION, INTENT(INOUT) :: T1
DOUBLEPRECISION, INTENT(INOUT) :: T2
DOUBLEPRECISION, INTENT(INOUT) :: Ts

to

REAL(4), INTENT(INOUT) :: T1
REAL(4), INTENT(INOUT) :: T2
REAL(4), INTENT(INOUT) :: Ts

This issue appears to be exclusive of gfortran. Compiling other MODFLOW packages may generate other errors. To debug the .f files and achieve compilation, check warnings and errors appearing on the terminal window—they do not appear in the ipython notebook. This would be a nice addition to pymake.

Thanks Chris!

Hello, as of last year I could make gflow using the make_gsflow.py script, which seems to have left the repository now. I tried using the make-program utility and get the following:

make-program gsflow -fc=gfortran --fflags='-O3'
Traceback (most recent call last):
File "/Users/payton.gardner/soft/FlopyTothDemo/.conda/bin/make-program", line 10, in
sys.exit(main())
^^^^^^
File "/Users/payton.gardner/soft/FlopyTothDemo/.conda/lib/python3.11/site-packages/pymake/cmds/build.py", line 140, in main
build_apps(**args)
File "/Users/payton.gardner/soft/FlopyTothDemo/.conda/lib/python3.11/site-packages/pymake/pymake_build_apps.py", line 146, in build_apps
code_dict[target] = usgs_program_data.get_target(target)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/payton.gardner/soft/FlopyTothDemo/.conda/lib/python3.11/site-packages/pymake/utils/usgsprograms.py", line 198, in get_target
return usgs_program_data()._target_data(key)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/payton.gardner/soft/FlopyTothDemo/.conda/lib/python3.11/site-packages/pymake/utils/usgsprograms.py", line 144, in _target_data
raise KeyError(msg)
KeyError: '"gsflow" key does not exist. Available keys: "mf6", "zbud6", "libmf6", "mp7", "mt3dms", "mt3dusgs", "vs2dt", "triangle", "gridgen", "crt", "sutra", "mf2000", "mf2005", "mfusg", "zonbudusg", "swtv4", "mp6", "mflgr", "zonbud3", "mfnwt1.1.4", "mfnwt", "mfusg_gsi"'

Looks, like gsflow is no longer a make target? Is there any reason? If not, can we get the make_gflow.py script back?

Thanks,
Payton

I am possibly confused. Does is this pymake specific to building modflow and modflow packages, or is it a version of the standard PyPi pymake? I am getting an error that is, in one form or another, most likely version related:

AttributeError: module 'pymake' has no attribute 'build_apps'

So I guess my question is, what is the relationship between modflowpy/pymake and PyPI/pymake (ie, pip install py-make )

Thanks!

A Pymake() class was recently introduced. The class makes it possible to download and build multiple MODFLOW-based programs with the same class. The class is implemented and used in the autotests and by the build_apps method that is used in the examples included in the repository.

Although the Pymake() is being used and tested, feedback on the implementation is welcomed. Pull Requests to improve the class would also be welcomed.

Both Intel Fortran (ifort) and Portland Group Fortran (pgfortran) use -module dir to describe the directory where module files files should be placed when created and where they should be searched for.

This is in contrast to GNU Fortran, which uses the -Jdir directive to describe the same feature, and is currently hard-coded to pymake.py

I can think of two ways to address this difference: one in the Makefile using variables, and the other by modifying pymake.py to produce different output depending on the options. (On a related note, I've also been thinking that a PyMake class should be created in pymake.py to provide OOP conveniences in creating objects for (e.g.) GNU and Intel compilers on POSIX or Win32 platforms, or create target rules only if certain source files are found.)

Hi,

I have a problem with mf6.2 on my mac:

I compiled it with pymake_mf6.py, using gfortran 5.2.0

When trying to run a program with the new mf6 I get the message:

mf6(21895,0x7fff77256300) malloc: *** error for object 0x7faeb1e01458: incorrect checksum for freed object - object was probably modified after being freed.
*** set a breakpoint in malloc_error_break to debug

Program received signal SIGABRT: Process abort signal.

Backtrace for this error:
#0 0x103b10092
#1 0x103b0f3b0
#2 0x7fff8ba19f19

Can you please help me.

Thanks
Barbara Fleck ([email protected]

I just tried to install pymake, but unfortunately it didn't work.

I think the line in pymake.py
from flopy import is_exe ...
should be
from flopy.mbase import is_exe ...

When doing a download and build of SEAWAT or MT3D on linux, would be good to enforce a lowercase naming convention for filespec.inc. This will require programmatically changing the BTN source file to use a lowercase "include filespec.inc" and undoing the upper case filespec.inc change that Joe just made to pymake.

Hey guys - Im not sure if Im doing this right, but im trying to get makefile from pymake that I can use to build the shared lib on linux. I added pm.makefile=True in mf6/autotest/build_exes.py.build_mf6_so() and I get a makefile and a makedefault, but they dont make it thru the build (haha - make) - all the objects get compiled, but the linker dies:

/usr/lib/gcc/x86_64-linux-gnu/7/../../../x86_64-linux-gnu/Scrt1.o: In function `_start':
(.text+0x20): undefined reference to `main'

pymake itself has no trouble building. Im sure this is user error - any idea what Im doing wrong?

Hello,

I am trying to use pymake to compile seawat v4 (Ubuntu 14.04.2). Unfortunately I have run into some errors. One I have not been able to resolve yet:

gfortran -O2 -c ./src_temp/parallel/para-mpi.f -o ./obj_temp/para-mpi.o -Iobj_temp -Jmod_temp
./src_temp/parallel/para-mpi.f:15: Error: Can't open included file 'mpif.h'
Traceback (most recent call last):
File "", line 1, in
File "pymake.py", line 342, in main
expedite, dryrun)
File "pymake.py", line 232, in compile_with_gnu
subprocess.check_call(cmdlist)
File "/usr/lib/python2.7/subprocess.py", line 540, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['gfortran', '-O2', '-c', './src_temp/parallel/para-mpi.f', '-o', './obj_temp/para-mpi.o', '-Iobj_temp', '-Jmod_temp']' returned non-zero exit status 1

Any ideas?

Much appreciated.

Regards,
Arjen

I'm trying to compile modflow-nwt on debian jessie, but I can't get it work.
I have tried with pymake, but thats not possible (like written in the documentation).

A makefile is not available with the source, so is there a known workaround?

I am using a MacBook pro intel architecture and tried to compile mt3dms. I am using gcc10.3. Unfortunatelly, I am getting the following error:

File "/Users/markuszechner/opt/anaconda3/lib/python3.7/site-packages/pymake/pymake.py", line 799, in build
raise FileNotFoundError("could not build {}".format(self.target))
FileNotFoundError: could not build mt3dms

The strange thing is that I can compile mp6, mp7, or mf2005 without any problems - it seems that only mt3dms is causing that issue.

When I used my old MacBook pro (also intel), I was able to install it without any issues - at that time, I used gcc 9.2.0. I tried to go back to 9.2 but homebrew only had 9.3 available. I tried 9.3 but that did not work. I don't know if this is important but I used homebrew to install the gcc versions (brew install gcc). There was one strange thing I realized when installing gcc: Ater I installed version 10 I was able to get the gfortran version (gfortran --version) but when reinstalling 9.3 that command did not work anymore - not sure if that is important.

Thanks a lot for your help!

Stay Safe!

Markus

I had to rename some .h files to get code compiled, so after creating FILESPEC.inc I added the following lines:

Replace #include "solvers.h" with #include "SOLVERS.h"

fn = os.path.join(srcdir, 'SOLVERS.h')
nfn = os.path.join(srcdir, 'solvers.h')
os.rename(fn,nfn)

fn = os.path.join(srcdir, 'MF2KGMG.h')
nfn = os.path.join(srcdir, 'mf2kgmg.h')
os.rename(fn,nfn)

fn = os.path.join(srcdir, 'CCFD.h')
nfn = os.path.join(srcdir, 'ccfd.h')
os.rename(fn,nfn)

The pymake installation on python 2.7 on Mac OS X brings the following error:

$ pip install https://github.com/modflowpy/pymake/zipball/master
Collecting https://github.com/modflowpy/pymake/zipball/master
  Downloading https://github.com/modflowpy/pymake/zipball/master
    Complete output from command python setup.py egg_info:
    pymake graphing capabilities not available.
    running egg_info
    error: Invalid distribution name or version syntax: -main--1.1.0

    ----------------------------------------
Command "python setup.py egg_info" failed with error code 1 in /var/folders/l0/k1r1vbqx1hbglfgs2w7z38c40000gn/T/pip-R9so8g-build/

With pip3 it's workign well.

$ pip3 install https://github.com/modflowpy/pymake/zipball/master
Collecting https://github.com/modflowpy/pymake/zipball/master
  Downloading https://github.com/modflowpy/pymake/zipball/master
Installing collected packages: -main-
  Running setup.py install for -main- ... done
Successfully installed -main--1.1.0

Looking for help with the following error with compiling mf2k_gwt, compiling on Ubuntu 16.04

here is the script:

from __future__ import print_function
import os
import shutil
import pymake
from pymake.download import download_and_unzip


def make_mf2k_gwt():

    # get current directory
    dstpth = os.path.join('temp')
    if os.path.exists(dstpth):
        shutil.rmtree(dstpth)
    os.makedirs(dstpth)
    os.chdir(dstpth)

    # Download the MODFLOW-2000 GWT distribution
    url = "https://water.usgs.gov/nrp/gwsoftware/mf2k_gwt/UNIX/mf2k_gwt.src.tar.gz"

    download_and_unzip(url)

    # Set directory name and source directory
    dirname = 'mf2k_gwt'
    srcdir = os.path.join(dirname, 'src')

    # Replace filespec with standard fortran
    l = '''
          CHARACTER*20 ACCESS,FORM,ACTION(2)
          DATA ACCESS/'STREAM'/
          DATA FORM/'UNFORMATTED'/
          DATA (ACTION(I),I=1,2)/'READ','READWRITE'/
    '''
    fn = os.path.join(srcdir, 'openspec.inc')
    f = open(fn, 'w')
    f.write(l)
    f.close()

    fn = os.path.join(srcdir, 'MF2KGMG.h')
    nfn = os.path.join(srcdir, 'mf2kgmg.h')
    os.rename(fn,nfn)

    fn = os.path.join(srcdir, 'CCFD.h')
    nfn = os.path.join(srcdir, 'ccfd.h')
    os.rename(fn,nfn)

    fn = os.path.join(srcdir, 'SOLVERS.h')
    nfn = os.path.join(srcdir, 'solvers.h')
    os.rename(fn,nfn)

    target = 'mf2k_gwt'
    pymake.main(srcdir, target, 'gfortran', 'gcc', makeclean=True,
                expedite=False, dryrun=False, double=True, debug=False)

if __name__ == "__main__":
    make_mf2k_gwt()

In python 2.X, sys.platform returns "linux2" for Linux machines (tested on ubuntu 14.04 LTS).

Current version of pymake.f in line 986:

if platform.lower() == 'darwin' or platform.lower() == 'linux'

returns False for ifort option when using python 2.X on Linux

Hope this helps for your project.

Thanks,
Harry

The new MacOS Monterey broke well running old USGS programs now leaving incomplete output files upon normal termination. This proved to be due to not flushing the file buffers upon ending the program. This is also the case when recompiled with the newest gfortran for MacOS Monterey, hence the problem would lay in the dynamic system libraries that come with the OS. This is solved by adding CALL FLUSH() just before the last END statement in the main program.
Next to this, Apple also broke compiling with Pymake. This is due to 1) the stdio.h and stdlib.h no longer on normal /usr/local/bin or so. It is solved by adding an INCLUDEDIR directive pointing to the right folder. Second, Apple removed the standard libraries like libc.a, libc.so from the file system. They are not win the "dynamic link cache". This turns out to be solved by adding --sysroot compiler directive.

I add the adapted makefiles for both mf2005 and seawat that show the correct arguments. Notice that seawat has more issues, due to the new fortran compiler being more seviere than previous ones, which are also described in the makefile.
makefile_mac_monterey_swt.txt
makefile_mf2005_MacMontery.txt
Compling_issue_analysis.txt

When I run make_mf2005.py I get the following error and the target file 'mf2005dbl' is never created.

Traceback (most recent call last):
File "make_mf2005.py", line 42, in
make_mf2005()
File "make_mf2005.py", line 35, in make_mf2005
assert os.path.isfile(target), 'Target does not exist.'
AssertionError: Target does not exist.

Can you think of any reason(s) why there is no mf2005dbl file? MF2005.1_12.u gets downloaded and placed in a folder 'temp' but I can't think of what is wrong to cause the file never to be written to the path (sorry, newbie)

On Ubuntu cannot find file 'filespec.inc'

change with:

fn = os.path.join(srcdir, 'FILESPEC.INC')

and compiled perfectly

A clean pymake install no longer downloads the required package subdirectories (.plot/ and .utils/). Probably due to setup.py changes in #98. This is causing flopy ci to fail.

My personal computer is Windows, and I hope to gain an executable file to run my Seawat model on HPC which is based on Linux system.

When I run the example file (make_swtv4.py) on my personal computer, there comes an error, and I don't know how to deal with it. So can someone offer me the swtv4 file that can be run on the Linux system?
Thanks for your help!

Is there some trick to using pymake without having to first run setup.py install? I've got it plumbed up in site-package pth file, but when I import pymake, I get an "ImportError python attempted relative import with no known parent package" coming from the from .dag line in __init__.py.

The autotest functionality in pymake will be moved to a separate repository over the next month and will be deprecated in pymake and eventually removed.

The new autotest repository mftestpy will be established the week of 7/27/2020 and no new features will be added once the mftestpy repo is being used in the MODFLOW 6 CI workflow.

It takes several tries for the modflow2000 example, make_mf2000.py, script to download the associated tar.gz file. Link seems fine if downloading in a web-browser though. Took me about 6 tries before I got a successful download. Not sure where the issue is. Pic is of full error code. Thanks

The following code in pymake.py (starting at line 373):

    # syslibs
    syslibs = []
    if sys.platform != 'win32':
        syslibs.append('-lc')

Should probably default to:

    syslibs = []
    if sys.platform != 'win32':
        syslibs.append('-static')
        syslibs.append('-lc')

The difference is that without the -static option, the resulting executable does not bundle the "c library", rather it links to where it is installed on the system.

Without the static it means, that in order to run the executable, you have to install the gfortran development files. Which is fine, but adds a step that requires internet access if you are running a live USB version of Linux.

The difference in executable size between the linked (~2.5MB) and static (~5MB) is IMHO not worth it as a default.

@mwtoews, I think some of the recent changes you made for capturing the compile messages is causing a problem on windows. Our builds are stalling with SEAWAT and MF2000, for example. I can run compile.bat just fine from a windows terminal, but python seems to get hung up. We've verified that it happens on a couple of different windows machines. I'm guessing the issue is somewhere in here:

        proc = Popen([batchfile, ], stdout=PIPE, stderr=PIPE)
        while True:
            line = proc.stdout.readline()
            c = line.decode('utf-8')
            if c != '':
                c = c.rstrip('\r\n')
                print('{}'.format(c))
            else:
                break

Could you take a look?

The default is not "binary" or STREAM ACCESS. To make mf2000 consistent with the other executables in this repo, the openspec file should be changed so that FORM = 'UNFORMATTED' should be replaced with FORM = 'BINARY' to be consistent with mfnwt and mf2005.

Problem shows up when trying to run the flopy3_MapExample.ipynb notebook using a mp6 binary created with make_modpath6.py.

It appears that after a certain version, -ffpe-summary compiler flag isn't available in gfortran.

mfusg appears to build with gfortran 4.6.3 on ubuntu precise x86_64 with the removal of this flag. I'm not sure what the implications are though.

I get the following error message (I wonder why autotest pops up here at all):

Collecting mfpymake
  Using cached mfpymake-1.2.4.tar.gz (75 kB)
  Preparing metadata (setup.py) ... error
  error: subprocess-exited-with-error
  
  × python setup.py egg_info did not run successfully.
  │ exit code: 1
  ╰─> [55 lines of output]
      Traceback (most recent call last):
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\site-packages\setuptools\config.py", line 419, in _parse_attr
          return getattr(StaticModule(module_name), attr_name)
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\site-packages\setuptools\config.py", line 26, in __init__
          spec = importlib.util.find_spec(name)
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\importlib\util.py", line 94, in find_spec
          parent = __import__(parent_name, fromlist=['__path__'])
        File "C:\Users\hofer_jn\AppData\Local\Temp\pip-install-ivzcy3zu\mfpymake_a60fcc056ca94562ba05233ab202b5e5\pymake\__init__.py", line 9, in <module>
          from .autotest.autotest import (
        File "C:\Users\hofer_jn\AppData\Local\Temp\pip-install-ivzcy3zu\mfpymake_a60fcc056ca94562ba05233ab202b5e5\pymake\autotest\autotest.py", line 54, in <module>
          import numpy as np
      ModuleNotFoundError: No module named 'numpy'
     
      During handling of the above exception, another exception occurred:
     
      Traceback (most recent call last):
        File "<string>", line 2, in <module>
        File "<pip-setuptools-caller>", line 34, in <module>
        File "C:\Users\hofer_jn\AppData\Local\Temp\pip-install-ivzcy3zu\mfpymake_a60fcc056ca94562ba05233ab202b5e5\setup.py", line 3, in <module>
          setup()
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\site-packages\setuptools\__init__.py", line 155, in setup
          return distutils.core.setup(**attrs)
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\site-packages\setuptools\_distutils\core.py", line 122, in setup
          dist.parse_config_files()
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\site-packages\setuptools\dist.py", line 818, in parse_config_files
          parse_configuration(
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\site-packages\setuptools\config.py", line 158, in parse_configuration
          meta.parse()
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\site-packages\setuptools\config.py", line 498, in parse
          section_parser_method(section_options)
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\site-packages\setuptools\config.py", line 469, in parse_section
          self[name] = value
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\site-packages\setuptools\config.py", line 222, in __setitem__
          value = parser(value)
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\site-packages\setuptools\config.py", line 599, in _parse_version
          version = self._parse_attr(value, self.package_dir)
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\site-packages\setuptools\config.py", line 422, in _parse_attr
          module = importlib.import_module(module_name)
        File "C:\Users\hofer_jn\Anaconda3\envs\xmipy\lib\importlib\__init__.py", line 126, in import_module
          return _bootstrap._gcd_import(name[level:], package, level)
        File "<frozen importlib._bootstrap>", line 1050, in _gcd_import
        File "<frozen importlib._bootstrap>", line 1027, in _find_and_load
        File "<frozen importlib._bootstrap>", line 992, in _find_and_load_unlocked
        File "<frozen importlib._bootstrap>", line 241, in _call_with_frames_removed
        File "<frozen importlib._bootstrap>", line 1050, in _gcd_import
        File "<frozen importlib._bootstrap>", line 1027, in _find_and_load
        File "<frozen importlib._bootstrap>", line 1006, in _find_and_load_unlocked
        File "<frozen importlib._bootstrap>", line 688, in _load_unlocked
        File "<frozen importlib._bootstrap_external>", line 883, in exec_module
        File "<frozen importlib._bootstrap>", line 241, in _call_with_frames_removed
        File "C:\Users\hofer_jn\AppData\Local\Temp\pip-install-ivzcy3zu\mfpymake_a60fcc056ca94562ba05233ab202b5e5\pymake\__init__.py", line 9, in <module>
          from .autotest.autotest import (
        File "C:\Users\hofer_jn\AppData\Local\Temp\pip-install-ivzcy3zu\mfpymake_a60fcc056ca94562ba05233ab202b5e5\pymake\autotest\autotest.py", line 54, in <module>
          import numpy as np
      ModuleNotFoundError: No module named 'numpy'
      [end of output]
  
  note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed

× Encountered error while generating package metadata.
╰─> See above for output.

When I try to use the makebin.bat file included with Modpath-7 to run run makebin.py I get the error:

args = pymake.pymake.parser()
AttributeError: module 'pymake.pymake' has no attribute 'parser'

When I run pip list it shows that pymake 1.2.1 is installed (it previously wasn't and the output from makebin.py instructed me to do so). Any suggestions for what I might be doing wrong? I was hoping to recompile with this commit included so that I can use MP7 on a grid that didn't meet the unnecessarily strict criteria.

Sorry, don't know how I ended up here. I thought I was in flopy.