Docs does not include master file or DLCs.xlsx

Currently it can be hard for new users to start because they don't have a master htc and associated DLCs.xlsx. Also not having version tracking on DLCs.xlsx is causing me a lot of grief. So I think it would be good to have a subfolder in docs where we store these. I'll do this, I'm just adding the issue for tracking purposes.

Absolutely right!

There is one complication though, I do not like to add a binary files to a git repo (unless with Large File Support, see further). So we should either have a script that creates that file, or host them somewhere else. Alternatively, we could use git LFS, which is actually supported on Gitlab, see here. But someone should first have to figure out how to use it properly.

Some more background on binary files and git: https://git-scm.com/book/en/v2/Customizing-Git-Git-Attributes#_binary_files

Ah. Good point. I think I've addressed this issue (see description below), but there are three possible issues with my solution right now:

• It uses module xlsxwriter. Is that okay that I've added a new dependency?
• It currently does not have a test function.
• It's currently just a function and an if __name__=="__main__" block, no classes or anything fancy. Not sure if we need those.

Description of my solution:

1. I saved each tab in a DLCs.xlsx file to a tab-delimited text file using VBA.
2. I wrote code that scrapes the information from each text file and saves the resulting data as an Excel file. I've attached an example output of the code.DLCs.xlsx

So, we can store the text files and the generating code in the repository, and people can submit issues if they see something wrong with the Excel files. Only issues to resolve are the extra module and possibly writing test functions.

I don't see any problems with an additional dependency on xlsxwriter (it is actually an optional pandas dependency as well).

I am not sure I can figure out where you placed the implementation of your solution. Is it in your fork?

Why not reading the csv file with pandas and just saving it as xlsx again with pandas? Something like:

import pandas as pd
df = pd.read_csv('DLCs.csv')
df.to_excel('DLCs.xlsx', sheet_name='main')

In this case we'll have to make sure pandas does not convert anything to something we don't like.

Theoretically, these files don't have to be of type xlsx, but I think it is otherwise nice to be able for users to safe the files with some formatting (at the end of the day a csv file is rather basic).

Since the master file contains model specific data I would argue that ideally they should belong together with the htc master file and maybe some example htc files in a model repository (like the one you have now for the V52). But maybe we'll figure out how to handle that another day.

Thanks for getting this thing fixed!! Thumbs up!

I kept the code local temporarily so I could check these issues with you before I pushed them.

I didn't know pandas had xlsxwriter built in -- I think I can do everything with pandas, which would be great. Let me check this real quick, and if it works I'll add it to the repo and create a merge request.

Yes, I've taken to naming my master Excel file with names that describe some of the model specifics (i.e., DLCs_onshore_WTI_TC3.xlsx for onshore, wind turbine class 1, and turbulence class 3). But yeah, we can find a more elegant approach to solve this later.

I think pd.read_excel relies on xlsxreader

it uses xlrd

mentioned in merge request !30 (merged)

mentioned in commit e9a1ff17

Turns out you can specify the Excel driver in pandas, and one of the options is xlsxwriter. Thus, the code now works using pandas.

Issue is resolved with merge request !30 (merged).

Note that I have also created a tutorials folder in docs for new users. I hope to add to it as time goes on.

closed