refactoring plans

[poldrack]

The initial release of the code (1.0a0) leaves a lot to be desired. In the next few weeks I am planning to refactor it, with a few goals in mind:

• define Publications class

  • move away from using DOIs as index - use a hash instead

• define Researcher class

• separate downloading from rendering -- ultimately would like to be able to render to generic markup (e.g. using jinja2) but probably won't get there for this next version.

• better command line interface

  • allow specification of:
    • base directory
    • template file

• need to rethink docker workflow. right now, the calls to docker use the version of the package that is installed within the container, rather than the current version on the host system, which could lead to version mismatches.

• Modularity (suggested by @chrisgorgo): try driving the refactor by making it easier for people to contribute new modules/components. For example layouts, styles, data sources etc.

• tests

  • what kinds of tests should we include (beyond an end-to-end smoke test)?

[thesamovar]

Still haven't had a chance to look at code, but is use of docker necessary? It makes it very tricky for people on Windows to use it, which is a shame.

[poldrack]

It’s not necessary but it avoids having to install latex...

On Mon, May 25, 2020 at 10:01 AM Dan Goodman notifications@github.com wrote:

Still haven't had a chance to look at code, but is use of docker necessary? It makes it very tricky for people on Windows to use it, which is a shame.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/poldrack/autoCV/issues/6#issuecomment-633653521, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGUVEHQPMG3PLLT2WFS3R3RTKP7VANCNFSM4NJVCEJQ .

[thesamovar]

It would probably not be too difficult to set it up so that it can be run either inside a docker with auto-installation of latex, or directly on host and you're responsible for setting up latex yourself?