Error on windows 7 64bits

[C.Renaudo]

First, is nice to see this kind of software being developed on Python, i think is a great language to do the job.
I have all the necesary dependencies installed, just missing one optional because i couldnt find the webpage to download (the periodic table)

Now to the bug:
When i try to run the software i get:

C:\Users\Usuario\Downloads\pychemqt>python pychemqt.py
Traceback (most recent call last):
  File "pychemqt.py", line 12, in <module>
    conf_dir = os.environ["HOME"] + "/.pychemqt/"
  File "C:\Python27\lib\os.py", line 423, in __getitem__
    return self.data[key.upper()]
KeyError: 'HOME'

The issue comes from the use of:

conf_dir = os.environ["HOME"] + "/.pychemqt/"

Windows dont have the HOME variable, the one it have is USERPROFILE, checking a way to fix this i found the next link:

http://stackoverflow.com/questions/14742064/python-os-environhome-works-on-idle-but-not-in-a-script

This command is used on the files pychemqt.py and lib/config.py

Another issue is the next line on the file lib/config.py

os.environ["pychemqt"]="/home/jjgomera/Programacion/pychemqt/"

I will try to make it run under windows in this week, sadly i am learning python at the moment so i can make a lot of mistakes.

If my english is to bad, i can write in spanish, is my main language.

[jjgomera]

Hi,

excuse me i didn't answer soon, sourceforge don't alert me this comment.

The issue comes from the use of:
conf_dir = os.environ["HOME"] + "/.pychemqt/"

Thanks, i was running pychemqt in idle so i thought os.environ had that property

Another issue is the next line on the file lib/config.py

That was my fault in svn upload, these are debug line i must comment before upload. Solved in 73 svn commit

On the config dir when i try to run the script i get the issue that is missing CostIndex.dat, i just copy that file from the svn to fix that.

That CostIndex.dat file has not the correct format, is a "history" CostIndex file, the CostIndex file in config directory must be created in first run, extract the last line in that dat/CostIndex.dat file. Try to delete .pychemqt config dir and run pychemqt again.

Use last svn commit from today with some bug resolved

Thanks for feedback

[C.Renaudo]

thanks for the answer, yea this forum don't alert when someone write, idk why.

After rev 74 it can run on windows on the first try (i just did a simple run) if you have most of the dependency installed. It load all, but i had issues using the units. (I can't input data to do the simulation). The terminal is a feature that will work only on GNU/Linux system. (to do the same on windows, there is a need to make a emulator, totally not worth)

If you want i can take some pics to show how it run on Win 7 64bits.

Best regards

[jjgomera]

Hi,

did you try to load files in samples dir? I'm trying in a windows xp virtual machine and load fine, for example pump and pipe, and there, as the console say i don't have any optional dependences, only the mandatory.

Start a new project in windows is no problem for me, except add stream, because a have installed a old pyqt version.

I would like to know the error your program raise, and how long do pychemqt take start? In my VM with windows xp is terribly slow

About console, I don't know now how to do that in windows, if anyone don't know the code of pychemqt could be a weird tool, but for me are fairly useful

Cheers!

[C.Renaudo]

Ok, lets start with the load time:
The first time i load pychemqt it take a lot to load, more if i have been using windows in other stuff (for example, my pc was on for 9 hours, using web browser, some games, octave and LibreOffice Writer) and the load time is 19 secs (Yes, a lot). But if i close it, do other stuff, and then i open pychemqt again (without a reboot) the load time on the second run is 1 to 2 secs.

Probably the issue is more a Windows Memory management issue than your code.

About the samples I can open only the next samples:
flujo
Heatexchanger
pipe
pump
turbine

What i cant do is to add stream to conect the in or outs to the units.
About the terminal, if you only need a shell to write python code, it could be useful to look at the shell of Ninja-IDE.

My system:
Win 7 64bits
I dont have Microsoft Office installed.
These are the packages i have installed on Python 2.7:
CoolProp (4.1.2)
Cython (0.20.1)
ezodf (0.2.5)
lxml (3.3.4)
matplotlib (1.3.1)
numpy (1.8.1)
oasa (0.13.1)
openpyxl (1.8.5)
pip (1.5.4)
pyparsing (2.0.1)
python-dateutil (2.2)
python-graph-core (1.8.2)
pytz (2014.2)
pywin32 (218.5)
reportlab (3.0)
scipy (0.14.0c1)
setuptools (3.4.4)
six (1.6.1)
PyQT (4.10)

[jjgomera]

The first time i load pychemqt it take a lot to load, more if i have been using windows in other stuff (for example, my pc was on for 9 hours, using web browser, some games, octave and LibreOffice Writer) and the load time is 19 secs (Yes, a lot). But if i close it, do other stuff, and then i open pychemqt again (without a reboot) the load time on the second run is 1 to 2 secs.

That's normal, when python runs first time compile libraries, create the pyc files, in second and successive runs that compile are not necessary, so a 1-2 seconds i think it's a good normal load time.

About the samples I can open only the next samples:
flujo
Heatexchanger
pipe
pump
turbine

Yes, I broke the meos library in last svn, so projects with that option can't load, i will solve in next svn. If you have coolprop installed, with its python binding you can load solid separator equipment (ciclon, baghouse, ...), only change line #218 in file lib/project.py to use that thermal option:

 config.set("Thermo", "coolprop", "True")

What i cant do is to add stream to conect the in or outs to the units.

These was a nonsense, update in today svn commit

Cheers!

[C.Renaudo]

Well, after a few test i got the program to show the splash screen. I had to modify 3 files to get to this stage

/pychemqt.py
/lib/config.py
/lib/sql.py

In the 3 files i had to modify how the variable conf_dir is defined. The way to make it work was:

conf_dir = os.path.expanduser('~')+ os.sep+".pychemqt"+os.sep

On the config dir when i try to run the script i get the issue that is missing CostIndex.dat, i just copy that file from the svn to fix that.

After all of that the program start to load and i get the next issue:

from tools.costIndex import indiceBase, indiceActual
  File "C:\Users\Usuario\My Documents\LiClipse Workspace\pychemqt\tools\costIndex.py", line 20, in <module>
    indiceActual.append(float(archivo.readline()))
ValueError: could not convert string to float: Feb-2001 393.6 437.0 365.6 441.1 544.1 365.8 670 3 340.8 414.2 300.0 385.6 341.1

Checking the file, the issue is the first 8 chars (the date) because it isnt a value that can be converted to float. (letters) I dont know a way to make this work so i will stop for today. I hope that my test are usefull for something.