Closed vxe closed 4 years ago
The contents of the Untitled notebook created by New Notebook
is here:
https://gist.github.com/vxe/3923e5ae93e0d23b3b1f8e1adb97c941
Please let me know if I can provide any additional information to try and troubleshoot further
It looks like you're running two servers, 8888 and 8889, so all the 8888 output seems irrelevant.
It looks like 8889 is giving you Forbidden, so that's a showstopper. Try using the classic jupyter browser interface.
Here's what my EIN looks like:
Not for nothing, but clojupyter seems like a far less mature REPL to lein repl
in https://github.com/clojure-emacs/cider. I suppose getting clojupyter to work in EIN is a valid source of amusement.
Mainly just want the ui for tweaking notebooks. I'm using a custom kernel: https://github.com/clojupyter/clojupyter/blob/master/doc/usage-scenarios.md
so maybe that an issue, as a sanity check did you install like this : https://github.com/clojupyter/clojupyter/blob/master/doc/library.md or use conda to install clojupyter?
here are my full server and request logs, launching with ein:run, using the above steps
server: https://gist.github.com/vxe/0a7861b7263d75707dbbe3707773f107 request: https://gist.github.com/vxe/449c70a5c31c3c707ec3ce1fea185aad
(emacs "notebook instantiation failed" message is same as above)
additionally on subsequent open attempts I'm seeing the below:
My hard drive is at capacity, and conda is too big, not to mention redundant with pip.
I used https://github.com/clojupyter/clojupyter/blob/master/doc/docker.md and then ein:login
to 8888.
When things get so discombobulated due to two servers, you need to start over with one. Your latest diagnostics suggest you merely killed 8889 and are trying to make it work with the same 8888 and the same emacs. EIN is in an irrecoverable state of confusion -- you need to start over.
the docker version is useless as it doesn’t persist notebooks
No argument there. I only used it to confirm EIN's functional, if not airtight, integration with clojupyter.
If it works with this:
https://github.com/clojupyter/clojupyter/blob/master/doc/library.md
i’d agree with your assessment . No conda required
Sorry, I'm not going to do that. Too many words. Docker lets me get in and get out quick (I've already deleted the image).
If you can't make it happen, I reassert CIDER as a manifestly less frustrating approach for an emacs clojure repl.
Didn't ask for a recommendation, but don't worry about it. Words are hard
@vxe - interestingly enough I can get clojupyter to work if I do not have clojure-mode installed in emacs. Another workaround that is working is calling (add-to-list 'polymode-mode-name-aliases '(clojure . lisp))
before opening the notebook.
Something seems to be going wrong in either ein:notebook-install-kernel
or ein:notebook-from-json
. Code stepping is a bit tedious, but with some luck I might get to the bottom of this - eventually.
@vxe - Don't think this is an issue with ein. I think the issue is a conflict is with cider. Check to see if you have an entry in clojure-mode-hook
that makes a call to cider
.
I'm running to what appears to a similar issue as:
https://github.com/millejoh/emacs-ipython-notebook/issues/589
when setting up clojupyter
I'm seeing a kernel appear ein:notebooklist, enableing
ein:poly-mode
, andtoggle-debug-on-error
and then clicking New Notebook I gethere is ein:log-all: https://gist.github.com/vxe/f901343361e69a29bf16edd45f437db4 here is *ein server: https://gist.github.com/vxe/5077c6ad13e1b5af48ba44b973feadf1
However a notebook book is indeed created,
I do a M-x customize-variable and set Ein:Jupyter Default Kernel. I open the notebook buffer that is created, its in python mode and looks like this:
with Ein Server showing
I click the unknown kernel message and select my clojure kernel again:
I get the elisp debugger stating notebook instantiation failed
and what looks like a giving up attempt in Ein Server