I would say this is "needed" quite often: I have personally been frustrated at the lack of this many times, and other people have too; someone so much so that they
Thanks. I did try these steps, but unfortunately no luck on either my M1 or older Intel Mac. I keep getting an error saying that homebrew could not compile a libgccjit example program.
I always rally against the % of words understood because in my experience, it doesn't align directly to the % of sentences understood. You can understand 80% of the words in a sentence but get effectively 0% of the sentence - or certainly miss the key parts.
1 small tweak to this which would make it beyond perfect... can you think of any way in which to have the items proceeded with an ordinal numbering list (ie. 1. Item1 2. Item 2 3. Item 3... etc etc...?). Handy for counts when filtering between types (esp booklist.).
Hmm, normally one sees the line number of point in the mode line, so one would have to place point at the last item and offset it by two (in my setup) to get the number of items. Not really convenient I guess.
Oh wow... this is great. This pretty much does exactly what I want (or at least uses the pattern I need0. Thank you!! One thing though, I modified the function a teeny bit but it means the sort keeps returning an error if I use booleans or wildcards on the `org-agenda-custom-command`.
Not sure... can't reproduce any of these issues. Works for me with "Type={.}", also with the custom sorting and timestamps. Try turn debugging on (M-x toggle-debug-on-error) and looking at the backtrace.
There’s an interesting conversation to be had here on vocabulary vs. semantics, one that I’m not qualified to weigh in on! The analysis seems to suggest that knowing 100 French words will let you understand around half of the average French book, but does being able to parse the most common, repetitive words such as pronouns and articles (un, une, le, la, il, elle, etc.) really get you that far if you don’t know any of the (less repeated) verbs and nouns they’re referring to? Interesting analysis but I’d love to see the next level up: if you trained an AI on only 100 common French words, what would its translations of a passage back to English look like?
Have you tried consult-org-heading? I think expanding on that with a custom completion category then adding flexible keyword metadata via a marginalia category would be a quick route. It uses org-map-entries under the hood (which is indeed getting quite fast in recent org versions).
Thanks for your suggestion! Yes, this is exactly what I tried in the beginning (I). The main problem is a marginalia annotator acts independent from and after the construction of completion candidates. So one would have to call org-map-entries once to get all the headings, and then one would have to again navigate to each heading within the annotator function to get the other entry data. This is too inefficient I'd say.
One other idea is to forgo marginalia and roll your own annotation function, packaging the keywords/etc. as text properties in midflight, then extracting them when the annotation gets requested.
I think I agree, mostly, but I suspect that many people (me included) actually spend almost all their time in org and don't just use it as a convenient format for producing html/pdf's. Hence, it would be desirable to have things looking nice in org too.
Hm, while git log has some arguments which can be used as a switch but also optionally take a value, e.g. --branches[=], I don't see any of these implemented in magit.
This is interesting, I will try it out. I already experimented a bit with async-shell-command but it doesn’t result in a nice experience when one wants the pdf output to pop up.
Some time back I faced a similar problem when I wanted to migrate all my existing Anki notes to Emacs anki-editor (and also sync back and forth between them). It ended up being quite the undertaking and was never published – until now:
Apologies for the delayed response. I did take a look for the code and it's not on my computer. It should be on a hard drive, i just haven't had a chance to check. I hope to look today and will share.
Thanks. I'm interested since I wrote some similar code a year or so ago, which however didn't work too well. Edit: I shared that code at top level now.
It seems to be possible to co-opt the mouse pointer to act as a cursor by using set-mouse-pixel-position, thus avoiding the slow re-rendering. This also seems to be the only way in Emacs to display something at an arbitrary pixel position...
I haven't seen the code yet (on mobile, saved the post to check it in a bit), but just wanted to let you know that pdf-tools allows keyboard navigation, albeit the keys used are different (I think N is next page, can't remember the others)
Sure, pdf-tools already provides ways to navigate between pages and to scroll around within pages. Plus occur, isearch, and isearch for links. This minor-mode extends this by introducing a cursor and allowing navigation by character, word, line, sentence, paragraph, etc.
A related query does a package exist to transcribe specific languages to IPA?
I am not aware of any. However, I have found this to be a good source for IPA data
I would say this is "needed" quite often: I have personally been frustrated at the lack of this many times, and other people have too; someone so much so that they
Thanks. I did try these steps, but unfortunately no luck on either my M1 or older Intel Mac. I keep getting an error saying that homebrew could not compile a libgccjit example program.
I used to get that same error, and
I always rally against the % of words understood because in my experience, it doesn't align directly to the % of sentences understood. You can understand 80% of the words in a sentence but get effectively 0% of the sentence - or certainly miss the key parts.
I think this is an interesting discussion to have. It is a point also raised
Why are you looking at ngrams?
Because I think they are a useful resource for language learning. Some suggested uses are listed in the Github repository:
1 small tweak to this which would make it beyond perfect... can you think of any way in which to have the items proceeded with an ordinal numbering list (ie. 1. Item1 2. Item 2 3. Item 3... etc etc...?). Handy for counts when filtering between types (esp booklist.).
Hmm, normally one sees the line number of point in the mode line, so one would have to place point at the last item and offset it by two (in my setup) to get the number of items. Not really convenient I guess.
Oh wow... this is great. This pretty much does exactly what I want (or at least uses the pattern I need0. Thank you!! One thing though, I modified the function a teeny bit but it means the sort keeps returning an error if I use booleans or wildcards on the `org-agenda-custom-command`.
Not sure... can't reproduce any of these issues. Works for me with "Type={.}", also with the custom sorting and timestamps. Try turn debugging on (M-x toggle-debug-on-error) and looking at the backtrace.
☹️ It fails in the same way:
I used to have the same error and resolved it like
https://forvo.com
Thanks. Anki-forvo-dl, unfortunately, doesn't have American English, which is weird, and can only grab the top of the findings.
For anki-forvo-dl: Once you apply the changes made in
There’s an interesting conversation to be had here on vocabulary vs. semantics, one that I’m not qualified to weigh in on! The analysis seems to suggest that knowing 100 French words will let you understand around half of the average French book, but does being able to parse the most common, repetitive words such as pronouns and articles (un, une, le, la, il, elle, etc.) really get you that far if you don’t know any of the (less repeated) verbs and nouns they’re referring to? Interesting analysis but I’d love to see the next level up: if you trained an AI on only 100 common French words, what would its translations of a passage back to English look like?
See my response
Do they mean words or characters? Several thousand Chinese characters can get you to a lot of places this I can confirm.
Words
Have you tried consult-org-heading? I think expanding on that with a custom completion category then adding flexible keyword metadata via a marginalia category would be a quick route. It uses org-map-entries under the hood (which is indeed getting quite fast in recent org versions).
Thanks for your suggestion! Yes, this is exactly what I tried in the beginning (I). The main problem is a marginalia annotator acts independent from and after the construction of completion candidates. So one would have to call org-map-entries once to get all the headings, and then one would have to again navigate to each heading within the annotator function to get the other entry data. This is too inefficient I'd say.
One other idea is to forgo marginalia and roll your own annotation function, packaging the keywords/etc. as text properties in midflight, then extracting them when the annotation gets requested.
Thanks, I'll try it out.
org-cite-csl-activate
Do note that activate processor previews the bibliography entries in a pop-up when one hovers over the citation reference.
I think I agree, mostly, but I suspect that many people (me included) actually spend almost all their time in org and don't just use it as a convenient format for producing html/pdf's. Hence, it would be desirable to have things looking nice in org too.
Hm, while git log has some arguments which can be used as a switch but also optionally take a value, e.g. --branches[=], I don't see any of these implemented in magit.
Nice to see more command-line integrations. I did a small
This is interesting, I will try it out. I already experimented a bit with async-shell-command but it doesn’t result in a nice experience when one wants the pdf output to pop up.
Some time back I faced a similar problem when I wanted to migrate all my existing Anki notes to Emacs anki-editor (and also sync back and forth between them). It ended up being quite the undertaking and was never published – until now:
Apologies for the delayed response. I did take a look for the code and it's not on my computer. It should be on a hard drive, i just haven't had a chance to check. I hope to look today and will share.
Thanks. I'm interested since I wrote some similar code a year or so ago, which however didn't work too well. Edit: I shared that code at top level now.
This is great - thanks for sharing your code!
I updated the repo with your suggestions
Great, that is very helpful
It seems to be possible to co-opt the mouse pointer to act as a cursor by using set-mouse-pixel-position, thus avoiding the slow re-rendering. This also seems to be the only way in Emacs to display something at an arbitrary pixel position...
I haven't seen the code yet (on mobile, saved the post to check it in a bit), but just wanted to let you know that pdf-tools allows keyboard navigation, albeit the keys used are different (I think N is next page, can't remember the others)
Sure, pdf-tools already provides ways to navigate between pages and to scroll around within pages. Plus occur, isearch, and isearch for links. This minor-mode extends this by introducing a cursor and allowing navigation by character, word, line, sentence, paragraph, etc.