From: Matt Price <moptop99@gmail.com>
To: John Kitchin <jkitchin@andrew.cmu.edu>
Cc: Org Mode <emacs-orgmode@gnu.org>
Subject: Re: lisp: scoping vars in repetitive defuns
Date: Wed, 18 Sep 2019 16:52:11 -0400 [thread overview]
Message-ID: <CAN_Dec-x2UQW2apRJLhsTFQhfatKA76_g3WChQnXBoqRkTewEg@mail.gmail.com> (raw)
In-Reply-To: <m2woe5bwj2.fsf@andrew.cmu.edu>
[-- Attachment #1: Type: text/plain, Size: 11538 bytes --]
This is fun, thanks John. I really like the plist version put would also
like to loop through the variables in a let statement somehow.
I think what I'm missing is the equivalent of a javascript implicit
destructuring construct:
let { } = object;
which will define new variables prop1, prop2... forever enumerable property
of the object. Is thre away to do that kind of destructuring bind -- which
binds *everything* in the plist, without knowing the symbol names in
advance? that would be really great.
On Wed, Sep 18, 2019 at 8:48 AM John Kitchin <jkitchin@andrew.cmu.edu>
wrote:
> You don't really need a macro for this I think. I see it leads to pretty
> clean looking code, but as you noted at the expense of edebuggable
> functions. I don't think you need the lexical-let in your macro though.
> With empty arguments I am not sure it does anything.
>
> Here are some other approaches to do what you want. I wasn't
> able to test these, but I think they are close to correct and would
> provide edebuggable functions for you. The plist is more flexible and
> future proof, you can add anything you want to it, and it won't mess up
> existing functions. The first approach you can add what you want to the
> end of the list, but you can't change the order without (probably)
> messing existing functions up.
>
> #+BEGIN_SRC emacs-lisp
> (require 'f)
>
> (defun dh-variables ()
> "Return a list of variables for the problem at point."
> (let* ((gh (org-entry-get (point) "GITHUB"))
> (base (org-entry-get (point) "ORG_LMS_ASSIGNMENT_DIRECTORY"))
> (findFiles (list
> (format "Reflection/%s.md" gh)
> (format "students/%s.md" gh)))
> (browseFiles '("index.html"))
> (testOutput "TestResults/testresults.html")
> (testCommand "MARKING=instructor npm test"))
> (list gh base findFiles browseFiles testOutput testCommand)))
>
> ;; Here we assign all the values from the function above to variable names.
> (defun dh-view ()
> "Open viewable files in browser"
> (destructuring-bind
> (gh base findFiles browseFiles testOutput testCommand)
> (dh-variables)
> (loop for f in browseFiles
> do
> (browse-url-of-file (f-join base f)))))
>
>
> ;; alternative with a plist
> (defun dh-variables ()
> "Return a plist of variables for the problem at point."
> (let* ((gh (org-entry-get (point) "GITHUB"))
> (base (org-entry-get (point) "ORG_LMS_ASSIGNMENT_DIRECTORY"))
> (findFiles (list
> (format "Reflection/%s.md" gh)
> (format "students/%s.md" gh)))
> (browseFiles '("index.html"))
> (testOutput "TestResults/testresults.html")
> (testCommand "MARKING=instructor npm test"))
> (list :gh gh
> :base base
> :findFiles findFiles
> :browseFiles browseFiles
> :testOutput testOutput
> :testCommand testCommand)))
>
>
> (defun dh-view ()
> "Open viewable files in browser"
> (loop for f in (plist-get (dh-variables) :browseFiles)
> do
> (browse-url-of-file (f-join base f))))
>
>
> ;; Or with the plist as an argument
> (defun dh-view (var-plist)
> "Open viewable files in browser"
> (loop for f in (plist-get var-plist :browseFiles)
> do
> (browse-url-of-file (f-join base f))))
>
>
> ;; called like this (dh-view (dh-variables))
> #+END_SRC
>
> Matt Price <moptop99@gmail.com> writes:
>
> > On Tue, Sep 17, 2019 at 8:46 AM John Kitchin <jkitchin@andrew.cmu.edu>
> > wrote:
> >
> >> I don't totally understand what you are trying to do here.
> >>
> >
> > I think the explanation was a little unclear!
> >
> >
> >> If this were Python, it sounds like you want some kind of class that
> >> stores a variable and reuses it several different functions? Something
> kind
> >> of similar to that in elisp is a closure, which is like what you
> described.
> >> For example, here, we define a variable a, and then define two functions
> >> that use it persistently.
> >>
> >>
> >> I think you can wrap this in a macro to make new functions, e.g.
> >>
> >> #+BEGIN_SRC emacs-lisp
> >> (defmacro f-maker (a)
> >> `(lexical-let ((a ,a))
> >> (defun f1 (x)
> >> (* a x))
> >>
> >> (defun f2 (x)
> >> (+ a x))))
> >>
> >> (f-maker 3)
> >>
> >> (list (f1 2) (f2 2))
> >> #+END_SRC
> >>
> >> #+RESULTS:
> >> | 6 | 5 |
> >>
> >> This is basically what I want, except it turned out to be easier to just
> > wrap the body forms in a let *within*the function. THis is what I came
> up
> > with:
> >
> > #+BEGIN_SRC emacs-lisp
> >
> > (defmacro dh-factory (name body &optional docstring)
> > "A helper macro that sets up the environment to simplify defining
> multiple
> > functions with the same environment variables. NAME will bcome the
> functin
> > name, BODY is a list containing the lisp forms specific to the function,
> > and DOCSTRING is an optional ... docstring. NAME wil lbe wrapped in a
> > `let` statement setting all the remelvant variables."
> > `(lexical-let (())
> > (defun ,name ,()
> > ,docstring
> > (interactive)
> > (let* ((gh (org-entry-get (point) "GITHUB"))
> > (base (org-entry-get (point)
> "ORG_LMS_ASSIGNMENT_DIRECTORY"))
> > (findFiles `( ,(concat "Reflection/" gh ".md") ,(concat
> > "students/" gh ".json")))
> > (browseFiles `( "index.html" ))
> > (testOutput "TestResults/testresults.html")
> > (testCommand "MARKING=instructor npm test"))
> > ,@body))))
> >
> > (dh-factory dh-find-files
> > ((dolist (f findFiles)
> > (message "%s" (concat base "/" f))
> > (if (file-exists-p (concat base "/" f))
> > (find-file-other-window (concat base "/" f) )
> > (message "File %s does not exist, not opening." f))) )
> > "Open gradable files in Emacs windows")
> >
> > (dh-factory dh-view
> > ((loop for f in browseFiles
> > do
> > (browse-url-of-file (concat base "/" f ))))
> > "open viewable files in browser")
> >
> > ;; this one should really pass a variable to allow different branches! oh
> > well.
> > (dh-factory dh-status
> > ((magit-status base)
> > (let ((currentBranch (shell-command-to-string
> > (format "cd %s && git rev-parse
> > --abbrev-ref HEAD" base) ))
> > (currentCommit (shell-command-to-string
> > (format "cd %s && git rev-parse HEAD"
> > base))))
> > (magit-stash-worktree
> > (format "Stashed random files from %s after commit
> %s."
> > currentBranch currentCommit)))
> > (magit-checkout (concat gh "-master")))
> > "Open a magit-status buffer and check out this student's
> branch
> > in the repo")
> >
> >
> > (dh-factory dh-tests
> > ((let ((output))
> > (with-temp-buffer
> > (message (concat "cd " base " ; npm test"))
> > (setq output (shell-command-to-string (concat "cd " base
> " ;
> > npm test"))))
> > (message "WELL, sorta made it through: %s" output)
> > (browse-url-of-file (concat base testOutput))))
> > "Run tests directly from macs and view output")
> >
> >
> > #+END_SRC
> >
> > I quite like it, though it's a bit unwieldy not to have the the source
> code
> > available for edebug etc. It feels a little like having a namespace; I'm
> > not afraid of polluting the global namespace, but I can use these symbol
> > names throughout the code I'm processing via the factory.
> >
> > I'm not sure it's *all* that much more efficient though...
> >
> >>
> >> There is also a class system you can probably use like this called
> eieio,
> >> Here is one approach to doing this.
> >>
> >> #+BEGIN_SRC emacs-lisp
> >> (require 'eieio)
> >>
> >> (defclass Grade ()
> >> ((:a :initarg :a)))
> >>
> >> (cl-defmethod gf1 ((g Grade) x)
> >> (* (oref g :a) x))
> >>
> >> (cl-defmethod gf2 ((g Grade) x)
> >> (+ (oref g :a) x))
> >>
> >> (let ((G (Grade :a 3)))
> >> (list (gf1 G 2) (gf2 G 2)))
> >> #+END_SRC
> >>
> >> #+RESULTS:
> >> | 6 | 5 |
> >>
> >> I would love to learn to use eieio but it feels like a bit of a jump.
> >
> >>
> >> Finally, maybe the simplest thing to do is pass this information in as
> an
> >> argument, e.g. a plist or alist, or get them from a function that you
> only
> >> change in one place?
> >>
> >
> > That might have been more sensible but this was pretty fun!
> >
> > Thanks everyone for the help!
> >
> >
> >
> >> On Tue, Sep 17, 2019 at 7:31 AM Matt Price <moptop99@gmail.com> wrote:
> >>
> >>> I have a number of convenience functions define to help me with grading
> >>> assignments. As I go through the semester, i update all of these
> functions
> >>> modestly so that they'rehelpful for grading the current assignment.
> >>>
> >>> I big chunk of these simple functions is taken up just declaring
> >>> variables with (let (())) forms. Each function uses some ofhte same
> >>> variables, e.g:
> >>>
> >>> (defun dh-find-files ()
> >>> (interactive)
> >>> (let* ((base (org-entry-get (point) "ORG_LMS_ASSIGNMENT_DIRECTORY"))
> >>> (gh (org-entry-get (point) "GITHUB"))
> >>> (f2o `( ,(concat "Reflection/" gh ".md") ,(concat "students/"
> gh
> >>> ".json")))) ;;;; "01/index.html" "02/index.html" "03/style.css"
> >>> "04/style.css"
> >>> (message "%s" f2o)
> >>> ;; make more flexible for resubmits
> >>> (shell-command (concat "cd " base " && git checkout " gh
> "-master"))
> >>> (dolist (x f2o)
> >>> (if (file-exists-p (concat base "/" x))
> >>> (find-file-other-window (concat base "/" x) )
> >>> (message "File %s does not exist, not opening." x)))))
> >>>
> >>> (defun dh-tests ()
> >>> (interactive)
> >>> (let* ((base (org-entry-get (point) "ORG_LMS_ASSIGNMENT_DIRECTORY" ))
> >>> (gh (org-entry-get (point) "GITHUB")))
> >>> (with-temp-buffer (shell-command (concat "cd " base " && npm test")
> >>> t)) ;; the "t" lets us suppress buffer
> >>> (browse-url-of-file (concat base "/TestResults/testresults.html"))
> >>> ;; (dh-mocha-run)
> >>>
> >>> ))
> >>>
> >>> ----------
> >>>
> >>> This semester I changed some elements of my workflow and I had to
> update
> >>> all the (org-entry-get) calls to new values. It makes me think the
> code is
> >>> less maintainable than it could be. I would like to do something like
> this:
> >>>
> >>> (lexical-let ((base `(org-entry-get (point)
> >>> "ORG_LMS_ASSIGNMENT_DIRECTORY")
> >>> (gh `(org-entry-get (point) "GITHUB")) )
> >>> (defun dh-find-files ()
> >>> (with-temp-buffer (shell-command (concat "cd " base " && npm test") t))
> >>> ;; the "t" lets us suppress buffer
> >>> (browse-url-of-file (concat base
> "/TestResults/testresults.html")))))
> >>>
> >>>
> >>> Obviously it doesn't work this way. But is there any way to set macros
> >>> like this to be expanded later inside a function definition? I feel
> certain
> >>> there must be...
> >>>
> >>> Thanks,
> >>>
> >>> Matt
> >>>
> >>
>
>
> --
> Professor John Kitchin
> Doherty Hall A207F
> Department of Chemical Engineering
> Carnegie Mellon University
> Pittsburgh, PA 15213
> 412-268-7803
> @johnkitchin
> http://kitchingroup.cheme.cmu.edu
>
[-- Attachment #2: Type: text/html, Size: 15694 bytes --]
next prev parent reply other threads:[~2019-09-18 20:52 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-09-17 11:30 lisp: scoping vars in repetitive defuns Matt Price
2019-09-17 12:46 ` John Kitchin
2019-09-18 1:12 ` Matt Price
2019-09-18 12:48 ` John Kitchin
2019-09-18 20:52 ` Matt Price [this message]
2019-09-18 21:26 ` Joost Kremers
2019-09-18 21:44 ` Adam Porter
2019-09-18 23:17 ` John Kitchin
2019-09-19 0:10 ` Adam Porter
2019-09-18 21:42 ` Adam Porter
2019-09-17 13:42 ` Nick Dokos
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: https://www.orgmode.org/
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAN_Dec-x2UQW2apRJLhsTFQhfatKA76_g3WChQnXBoqRkTewEg@mail.gmail.com \
--to=moptop99@gmail.com \
--cc=emacs-orgmode@gnu.org \
--cc=jkitchin@andrew.cmu.edu \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://git.savannah.gnu.org/cgit/emacs/org-mode.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).