search

UVAdMIST / swmm_mpc

Python package for model predictive control (MPC) for EPASWMM5 models
MIT License
23 stars 9 forks source link

running with no results #6

Open Jiadalee opened 5 years ago

Jiadalee commented 5 years ago

Hi :

I can run your codes but I didn't get any results. Have you ever encounter this problem?

jsadler2 commented 5 years ago

Hi Jiadalee,

Thanks for reaching out. The main result that the current version of the code produces is a csv file in the results_dir (an input to the run_swmm_mpc function). The csv file is titled [start datetime]_ctl_results_[run_suffix] where run_suffix is another parameter of the run_swmm_mpc function. After you run your code do you see that file?

Jiadalee commented 5 years ago

Hi Jeff:

No. Basically, I wrote the 'my_swmm_mpc.py'file and created a 'result' folder. After that, I run my_swmm_mpc.py in Linux system. This shows running is successful but there is nothing in the 'results 'folder. Can you help me check the 'my_swmm_mpc.py' file? I can share it with you.

Jiadalee commented 5 years ago

Hi Jeff:

What does 'To add to the path, add this line to your .bashrc' mean? Is .bashrc a file?

jsadler2 commented 5 years ago

Hi @Jiadalee , I can definitely look at your my_swmm_mpc.py file. What is output to the shell when you run it? Do you see some information about the the genetic algorithm generations? And yeah. The .bashrc is a file. It's should be in your home directory if you are running Linux.

Jiadalee commented 5 years ago

Thanks!. I added 'print(MPC is great)'at the end of the my_swmm_mpc.py file. So the output to my shell is only'MPC is great'. I didn't see any info about GA generators. I'm wondering if there are problems in the setting-up before running.

Below is the my_swmm_mpc codes:

from swmm_mpc import run_swmm_mpc

input_file = "/uufs/chpc.utah.edu/common/home/u1147017/RL/flux_version_system_run/swmm_mpc/swmm_mpc/sample.inp" control_horizon = 1. #hr control_time_step = 900. #sec control_str_ids = ["ORFICE R1", "ORIFICE R2"] results_dir = "/uufs/chpc.utah.edu/common/home/u1147017/RL/flux_version_system_run/swmm_mpc/swmm_mpc/results/" work_dir = "/uufs/chpc.utah.edu/common/home/u1147017/RL/flux_version_system_run/swmm_mpc/swmm_mpc/" ngen = 4 nindividuals = 300

target_depth_dict = {'St1':{'target':1, 'weight':0.1}, 'St2':{'target':1.5, 'weight':0.1}}

def main(): run_swmm_mpc(inp_file, control_horizon, control_time_step, work_dir, target_depth_dict = target_depth_dict, ngen = ngen, nindividuals = nindividuals ) if name == "main": main()

Jiadalee commented 5 years ago

mpc problem

jsadler2 commented 5 years ago

How long does it take to run that?

Jiadalee commented 5 years ago

just two seconds. Very short

jsadler2 commented 5 years ago

Hmm. When I run it, it usually takes at least 10 minutes. Where in the above code do you have your print statement? I don't see it. Also, do you mind using the markdown code block feature so it's easier to read?

Jiadalee commented 5 years ago

Hi:

No problem！ Sorry, I put the code in the swmm_mpc.py. Did you mean create an MD file?

Jiada

Jiadalee commented 5 years ago

Can I email this code to you?

jsadler2 commented 5 years ago

Sure. You can email me OR maybe it would be easier to send on Research Gate. I just replied to your thread about sharing the paper. Let me know if that works. If not, I can give you my email address.

jsadler2 commented 5 years ago

Just got your code file. Which version of Python are you running? Also, I think one of your problems is in the last part of your script. You have 

if _name_ == "_main_":

What I think you need to have is 

if __name__ == "__main__":

(add another underscore before and after name and main) Also, although you have the results_dir variable at the top of your script, you do not include that as an input parameter when you are actually calling the run_swmm_mpc method.

Jiadalee commented 5 years ago

Hi Jeff:

I use Python 3.6.7 to run my codes.

Oh! I see. I have changed the codes and also added the results_dir to the input variable(see the screenshot). Still, same problem.... 2

Jiadalee commented 5 years ago

underscore to 'nam'e was also added ..

Jiadalee commented 5 years ago 
from swmm_mpc import run_swmm_mpc

input_file = "/uufs/chpc.utah.edu/common/home/u1147017/MPC/swmm_mpc/swmm_mpc/sample.inp"
control_horizon = 1. #hr
control_time_step = 900. #sec
control_str_ids = ["ORFICE R1", "ORIFICE R2"]
results_dir = "/uufs/chpc.utah.edu/common/home/u1147017/MPC/swmm_mpc/swmm_mpc/results/"
work_dir = "/uufs/chpc.utah.edu/common/home/u1147017/MPC/swmm_mpc/swmm_mpc/"
ngen = 4
nindividuals = 300

target_depth_dict = {'St1':{'target':1, 'weight':0.1}, 'St2':{'target':1.5, 'weight':0.1}}

def main():
    run_swmm_mpc(input_file,
                 control_horizon,
                 control_time_step,
                 work_dir,
                 results_dir,
                 target_depth_dict = target_depth_dict,
                 ngen = ngen,
                 nindividuals = nindividuals
                 )
    if __name__ == "__main__":
        main()
jsadler2 commented 5 years ago

Okay. I've only been running this on Python 2.7. I haven't tried on Python 3. So I would first suggest to try running on Python 2.7. After that, I'm still a little unsure that it is running at all. Could you put a print statement in your code? like this: 


def main():
    run_swmm_mpc(input_file,
    control_horizon,
    control_time_step,
    work_dir,
    results_dir,
    target_depth_dict = target_depth_dict,
    ngen = ngen,
    nindividuals = nindividuals
    )
    print 'yep. my code is running'

if name == "main":
    main()
Jiadalee commented 5 years ago

Sure! Let me have a try

Jiadalee commented 5 years ago

Sorry. I got the same error. ...

Actually, I double checked the codes like swmm_mpc.py and run_ea.py file. I found it seems they are written in Python 3 since they use syntax such as print (' xxx ') but not print ' '. If so, I think running them in Python 3 is a correct way.

Can I run these codes in Windows system? If so, how should I do? Thanks

Jiada

Jiadalee commented 5 years ago

Ah....I think I have compiled it to my Local PC. But I got an ImportWarning.

image

Jiadalee commented 5 years ago

Ok, I fixed this local PC 'deap' problem of swmm_mpc. So here comes the same question just same as I run it in Linux system: I can run it but no output in 'results' folder.....

jsadler2 commented 5 years ago

Currently, running on Windows is not supported.

jsadler2 commented 5 years ago

I am trying your code on my computer to see if I can reproduce the behavior your seeing

jsadler2 commented 5 years ago

@Jiadalee, I ran your code and found a couple problems.

  1. In line 6 you spell "ORIFICE" wrong
  2. You didn't include control_str_ids in your function call
  3. make sure your if __name__ ... line is outside your main() function
  4. you need to include "Node" in front of St1 and St2 in the target_depth_dict (this was a problem in my example in the readme which has been fixed now.

After changing all those thing, I got it to work.

Jiadalee commented 5 years ago

Thanks. I will check it out.

Jiadalee commented 5 years ago

image

Jiadalee commented 5 years ago

I modified my codes. This time I got a log file in my 'results' folder. I passed this file to you via researchgate

Jiadalee commented 5 years ago

Oh. I fixed the last question

Here comes another one:

when I include the ifname...line inside the main() function, then, I run the codes. Finally, there is no error and I also got a newly-produced .inp file which is called 'simple_processed'. However, nothing in the results folder....

Jiadalee commented 5 years ago

I can get a new .rpt file now after running it in python3

Jiadalee commented 5 years ago

Hi:

Do you know how to solve the time_step problem?

image

jsadler2 commented 5 years ago

Awesome. So you are getting it to run. That is actually a known bug with a simple fix that I've neglected to push. I will push that today.

jsadler2 commented 5 years ago

I just pushed the commit that should fix that. After you run the code, it should update the control rules of the .inp file that you gave it. https://github.com/UVAdMIST/swmm_mpc/commit/3dd16c6e0ae4a9b6770089151e4016d41b584219

jsadler2 commented 5 years ago

I'm testing it now.

Jiadalee commented 5 years ago

Yes. It basically can run in the Linux system now.

jsadler2 commented 5 years ago

I just ran it with the latest commit and there was no error. Pls let me know if you still have an error.

Jiadalee commented 5 years ago

so you added a line '''python sim_start_dt = pd.to_datetime(sim_start_time) ''' before 'ct1_settings_df.....'

Jiadalee commented 5 years ago

Hi:

I still got this error. Actually, I have adjusted the code to make it run in Python3. So, could you let me know where you modify the code to fix the last error? so I can also do this in my side. Thanks

image

jsadler2 commented 5 years ago

Hi yes. If you just make the changes from this commit in your code, it should work: https://github.com/UVAdMIST/swmm_mpc/commit/3dd16c6e0ae4a9b6770089151e4016d41b584219 . Let me know if that makes sense.

jsadler2 commented 5 years ago

And that's cool you got it to run in Python3. What changes did you have to make?

Jiadalee commented 5 years ago

Hi Jeff:

Sorry for the late reply.

Basically, I modified the tabs and indentations. I will send you the updated scripts for running in Python3 Linux system soon.

Jiadalee commented 5 years ago

Hi: I test these modified codes in python 3 again. I used the simple.inp model and only got 15-min simulation results. Can this code only run 15 mins?

image

DangMH-TLU commented 3 years ago

Dear Authors, Could you possibly help me to fix the following error?

minhhai@MInhHai:/mnt/c/swmmmpc/models/runs$ python hgv11.py {'optimization_params': {u'num_cores': 20, u'nindividuals': 20, u'ngen': 5}, 'n_ctl_steps': 2, 'ctl_horizon': 0.5, 'work_dir': u'/mnt/c/swmmmpc/models/runs/work', 'inp_file_dir': u'/mnt/c/swmmmpc/models/runs', 'flood_weight': 1, 'inp_file_path': u'/mnt/c/swmmmpc/models/runs/hague_v11_template.inp', 'results_dir': u'/mnt/c/swmmmpc/models/runs/rtc_output', 'opt_method': u'genetic_algorithm', 'run_suffix': u'hgv11_base_00_2yr12hr_gate_ga', 'ctl_str_ids': [u'PUMP P0', u'WEIR W0', u'ORIFICE R2'], 'target_depth_dict': None, 'ctl_time_step': 900, 'dev_weight': 0, 'log_file': u'/mnt/c/swmmmpc/models/runs/rtc_output/log_hgv11_base_00_2yr12hr_gate_ga', 'inp_process_file_path': u'/mnt/c/swmmmpc/models/runs/work/hague_v11_template_process.inp', 'node_flood_weight_dict': None} Simulation start: 2021.02.09.16.04 2016-10-08 00:15:01

Traceback (most recent call last): File "hgv11.py", line 46, in main() File "hgv11.py", line 43, in main run_swmm_mpc(config_file) File "/home/minhhai/.local/lib/python2.7/site-packages/swmm_mpc/swmm_mpc.py", line 150, in run_swmm_mpc run.optimization_params) File "/home/minhhai/.local/lib/python2.7/site-packages/swmm_mpc/run_ea.py", line 54, in run_ea halloffame=hof, verbose=True) File "/home/minhhai/.local/lib/python2.7/site-packages/deap/algorithms.py", line 150, in eaSimple fitnesses = toolbox.map(toolbox.evaluate, invalid_ind) File "/usr/lib/python2.7/multiprocessing/pool.py", line 253, in map return self.map_async(func, iterable, chunksize).get() File "/usr/lib/python2.7/multiprocessing/pool.py", line 572, in get raise self._value IOError: [Errno 2] No such file or directory: '/mnt/c/swmmmpc/models/runs/work/hague_v11_template_process_tmp_fpsnd2f8w.rpt'

Thank you very much in advance. Regards, Dang Minh Hai

Jiadalee commented 3 years ago

@DangMH-TLU It seems the working directory is not correct. Can you introduce which programming environment are you using and how did you navigate to your working dir?

DangMH-TLU commented 3 years ago

Hi Jiadalee, I run Swmm_mpc on python 2.7 on Ubuntu 18.04. I navigate to the working directory as follows:cd /mnt/c/swmmmpc/models/runs/ Thank you very much. Regards, Dang Minh Hai

jsadler2 commented 3 years ago

Hi @DangMH-TLU - sorry for the slow reply. I think your error means that you haven't installed swmm5 correctly or it's not in your path. Did you follow this step in the readme? https://github.com/UVAdMIST/swmm_mpc#2-install-epaswmm5

@bdb3m - didn't you have a similar problem? Were you able to fix that?

bdb3m commented 3 years ago

@jsadler2 I'm fuzzy on the details since it's been a while but that sounds right. I'll see if I can find anything in my notes.

@DangMH-TLU I hope jsadler2's response can help you get it working!

DangMH-TLU commented 3 years ago

@jsadler2 : Thank you very much. I fixed sucessfully the error which originated from the wrong working directory in the .bashrc file. Again, thank you very much!

DangMH-TLU commented 3 years ago

@bdb3m Thank you very much!

jsadler2 commented 3 years ago

Glad you got it working!