Pull Request for #13: Migrate demo scripts to `/examples/` and docs to wiki

gentzkow commented 3 months ago

@ShiqiYang2022 @Xingtong-Jiang In this update, I have made two main changes to the template:

I moved all of the example scripts out of the main directories to /examples/. This way someone who downloads the template doesn't need to worry about having Python / R / Stata installed to be able to run run_all.py successfully. It also means there aren't so many example files floating around -- while my intention was that people would delete these, I've observed that they sometimes get left in repositories cluttering things up.
I migrated the documentation for using the template to the GentzkowLabTemplate wiki and updated the README to be a template for a repository readme.

I'd appreciate it if you could both try running the template "out of the box", look over the documentation, and then try copying over and running the Python / Stata / R / Latex example scripts following the instructions here.

ShiqiYang2022 commented 3 months ago

Thanks @gentzkow for setting this up! Confirming we have received this and will start working on testing soon.

gentzkow commented 3 months ago

@Xingtong-Jiang Thanks for your comments here. I updated the wiki to reflect them.

I made the change to make it a template repository.

@ShiqiYang2022 @Xingtong-Jiang I went through your proposed commits and accepted some, modified, some, and rejected some. Feel free to go through the comments above and resolve them if you agree.

gentzkow commented 3 months ago

@ShiqiYang2022 Also, I would propose we get rid of setup.sh. My idea is that check_setup.sh can perform the same function. If you think it's nicer to have it at the top level to make it more prominent, I'd be fine moving check_setup.sh back out of lib/shell/ to the root of the repo.

ShiqiYang2022 commented 3 months ago

@gentzkow Thanks! Re your https://github.com/gentzkow/GentzkowLabTemplate/pull/14#issuecomment-2272122797: I agree that we don't need setup.sh. I've also noticed that when I reviewed the repo (but I forgot to flag that). I went ahead and removed in 1888b97.

gentzkow commented 3 months ago

@ShiqiYang2022 @Xingtong-Jiang

I made a few more tweaks here

I added more extensions to .gitignore so we're ignoring all the auxiliary files I know of produced by LaTex builds
I set up the default .gitattributes for Git LFS and added code to check_setup.sh to verify the Git LFS is installed.
I revised the make.sh scripts so they are all identical; I want to preserve this property so we don't have to keep updating them separately every time there's a change
I tweaked the Stata scripts so they don't produce graphs with blue backgrounds
I reorganized inputs_for_examples again to simplify; I think we want to have just the .csv file plus the other intermediate outputs in case someone wants to run the paper or slides without setting up analysis. We don't need the intermediate data files because if someone is going to set up R / Stata / Python they can set up both data and analysis.
I revised the README.md for the LaTex example directory. I plan to revise the ones in the other directories similarly but I haven't had a chance. I think we should wait for that until this PR is merged then I can do it in another issue.

I would appreciate it if you guys could do one more quick round of testing to make sure I haven't broken anything.

Next steps for me/us after this issue is closed

Revise those readmes
Work more on how to robustly require Git LFS and also create an easy option for people who want to use the repo without Git LFS
Consider whether there's a good way to get rid of the redundancy of having multiple copies of make.sh. One would be to restore setup.sh and have it populate the various directories with copies of a single make.sh script. We could consider adding to this code to set up the LaTex, Stata, etc. examples and prompting the user to input which of these they would like.

ShiqiYang2022 commented 3 months ago

@gentzkow

I initialized a test branch 13-testing-0812, did an another round of test on that branch and pushed the outputs there. Everything looks good to me, I made some small tweaks in ef9b324. Feel free to revert changes if you disagree.

I agree with your roadmap per https://github.com/gentzkow/GentzkowLabTemplate/pull/14#issuecomment-2285161889, and I and @Xingtong-Jiang are willing to contribute. Shall we open the issues and allocate among ourselves after we closed this?

Thanks!

gentzkow commented 3 months ago

@ShiqiYang2022 Looks good, modulo one tiny comment I made.

gentzkow / GentzkowLabTemplate

Pull Request for #13: Migrate demo scripts to `/examples/` and docs to wiki #14