Results reader fails in recent (post 2018 hotfix 2?) versions

[jgill]

The results reader get_variables() call is failing in the most recent VE version.

This can be easily replicated by opening any project with a results file and running the 'read_aps_energy.py' script, giving the output:

>>> Run start, Fri Sep 28 14:26:06 2018
Traceback (most recent call last):
File "C:\Program Files (x86)\IES\VE 2018\apps\Scripts\api_examples\resultsreader\read_aps_energy.py", line 118, in <module>
aps_variables = results_file_reader.get_variables()
AttributeError: type object 'Error' has no attribute 'values'
>>> Runtime: 2.83 seconds

Not sure if there's some undocumented (or documented somewhere I haven't seen yet) change related to accessing vista generated variables in the python environment.

Has anyone seen anything similar? any suggestions?

[LeightonSS55]

jgill,

Just going to say that the same is happening for me.

It seems the .get functions which access the C++ data do not match the IES-VE user documentation.

Boost.Python.ArgumentError: Python argument types in
ResultsReader.get_room_results(ResultsReader, str, str, str, int, int)
did not match C++ signature:
get_room_results(class ResultsReader {lvalue}, class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >, class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >, class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> >, char, int start_day=-1, int end_day=-1)

Can someone from IES clarify ASAP please?

Thanks,

[Rhind]

Hi,

This is what's needed for get_room_results:

Code: Select all

get_room_results(room_id, aps_var, vista_var, var_level, start_day, end_day)

start_day and end_day can be omitted, in which case the full date range is used.

There's also the related function get_all_room_results which may be of use:

Code: Select all

get_all_room_results(room_id, aps_var, var_level, start_day, end_day)

Hope that helps.

[LeightonSS55]

Hi Rhind,

Thanks for getting back and the code makes sense for use in VE2018.

I note that between the VE2017 version and VE2018 (Hotfix 2), the method for get_room_results changed?

In VE2017, when running the example 'overheating_test.py' script; the method for accessing results is stipulated as below:

Code: Select all

np_room_temp = results_reader_file.get_room_results(room_id, 'Comfort temperature', 'z', 1, 365)

In VE 2018, when running the example 'overheating_test.py' script; the method for accessing results is as below (and aligns with the VE Python Script User Guide)

Code: Select all

np_room_temp = results_reader_file.get_room_results(room_id, 'Comfort temperature', 'Dry resultant temperature', 'z', 1, 365)

This explains why i was getting the above error - both the aps_var & the vista_var were required, whereas in VE2017 it seems it wasn't required.
-----------------------------------------------------------------

Now, linked to jgill's original comment; the 'get_variables()' isn't working in VE2018.

The below script is what I have used to access the dictionary of .aps variables and print it out to excel. I did this so I can see name of both the vista_var AND the aps_var, because as far as I can tell there is no documentation of the aps_var, and hence needed a method to find this out so i can access the correct results in any future scripts.

Code: Select all

'''Importing modules'''
import iesve
import xlsxwriter
import os
from ies_file_picker import IesFilePicker

'''Accessing Results via Results Reader'''
aps_file_name = IesFilePicker.pick_aps_file()
results_reader_file = iesve.ResultsReader.open(aps_file_name)
variables = results_reader_file.get_variables()

# # # print(variables)
# here we can print the function 'variables' which reads into the .aps file variables and prints out a LIST of DICTIONARIES

# Here we can just print the first variable - which has an .aps variable name of Lights & SP elec, but a display name of Equip Elec. CE
# print(variables[0])

# Here we can find out the length of the list of variables
# print(len(variables))

# Create an Excel worksheet and name it
workbook = xlsxwriter.Workbook('Results_Variables.xlsx')
worksheet = workbook.add_worksheet('Sheet1')

# Specify the Row and Column index to start at 0 
row = 0
col = 0

# Loop through the dictionaries and write to columns using the index method
for dict in variables[1:718]:
    worksheet.write(row, col, dict['units_type'])
    worksheet.write(row, col + 1, dict['post_process_spec'])
    worksheet.write(row, col + 2, dict['aps_varname'])
    worksheet.write(row, col + 3, dict['source'])
    worksheet.write(row, col + 4, dict['subtype'])
    worksheet.write(row, col + 5, dict['model_level'])
    worksheet.write(row, col + 6, dict['custom_type'])
    worksheet.write(row, col + 7, dict['display_name'])
    row += 1
    
# Close the workbook and use OS to open the Excel file
workbook.close()
os.startfile('Results_Variables.xlsx')

I get the same error message as Jgill, from running the code above in VE2018, but it runs fine in VE2017.

Please can you review and explain what has changed and/or what is required in the code? Hopefully this will answer jgill's question also.

Many thanks for your help so far,

Leighton

[Rhind]

Nothing has changed with the usage of get_variables() and the example script you've posted is correct. I've run it in a copy of VE 2018 and it worked as expected. I used the small office starter model, are you able to run your script with this?

MIght be worth getting in touch with support, who can investigate further.

[bbrannon4]

Did anyone ever figure out the issue with get_variables? I have some code that actually worked fine, then I must have dome something which broke it. In both cases it's in VE 2018 FP1, and I've sent this into support.

Here is my code up until the error:

Code: Select all

import iesve
from ies_file_picker import IesFilePicker
import pandas as pd
import pickle
import json

#### Select results file to use ####
aps_filename = IesFilePicker.pick_aps_file()
results = iesve.ResultsReader.open(aps_filename)

#### Check available rooms ####
rooms = pd.DataFrame(results.get_room_list()) # (name,ID,Area,Volume)
rooms.to_csv(r'C:\Users\ben.brannon\Desktop\rooms_out.csv')    #save to file if needed

#### Check available variables ####
var_names = results.get_variables()

And the resulting error:

Code: Select all

Traceback (most recent call last):
  File "C:\git\IESVE_Scripts\RoomConditionsExtract.py", line 20, in <module>
    var_names = results.get_variables()
AttributeError: type object 'NullImporter' has no attribute 'values'

get_room_list works fine, so at least it's accessing the results file in some manner.

[Terence]

bbrannon4, sorry to see you are having trouble with this... Our developers looked at the example you posted above and confirmed they were able to get the exception in this instance and they have logged it for investigation.

If they find a way around it then they will post back here with the information.

[bbrannon4]

Thanks Terence. In the meantime, is it possible for you to provide a full list of all available variables (vista name and aps name, plus the model level and units type)? I don't see a full list in the documentation and I'm not sure another way to verify the names. Specifically I'm trying to get the plant profiles and heating/cooling set-points out of the results, but it would be great to get a full list, similar to what I would get if my code above worked.

[Rhind]

bbrannon4, support will be in touch with some more information while we address this issue.

[LeightonSS55]

Hi Rhind / bbrannon4,

Appreciate the support.

I still am having the same issue - when this is resolved please may you post the solution in this thread please?

I also second bbrannon4's request to formalise the vista names and corresponding aps names (outside the python dictionary), as accessing the data is not the problem here - it's the method to find out the naming convention which is not working - Like the internet with no search engine

Thanks,