Visual Studio Code Jupyter Python Extension line wrap in output cell - how to set notebook.output.wordWrap

My Visual Studio Code (Mac) Jupyter extension (plug-in) output was not line wrapped. It was a single unreadable output string. I had to either copy and paste into a markdown cell (where VSC line wrap config is followed) or I had to come up with Python output formatting solutions.

Then I saw a video in which the developer's cell block output was wrapping. So I went looking again. This turned out to be a lot harder than it was, say, 10 years ago. I spent at least 40 minutes, and probably more, digging through obscure corners of the internet

At last I got a hint. There is something called "notebook.output.wordWrap" that defaults to false but can be set true. If you can figure out how to set it.

I couldn't figure out how to set this value for the Jupyter Notebook extension on my Mac however. I got some hints from Microsoft's documentation on editing settings.json even though it's not correct about how to navigate to preferences. I used the ⌘, shortcut. From there I searched for "wordwrap"

You can see the result below. I needed to set Notebook > Output > Word Wrap to true. The copied JSON version of this is "notebook.output.wordWrap": true. You can see the setting in the screen shot below (click for full size).

I set this value for User and Workspace both, I assume Workspace wins but I don't know. Now my output wraps so I can read it.

There's also supposed be a metadata tag that causes a Jupyter notebook cell to scroll output but I couldn't get I too work and it felt too obscure to rely on.

Rendering ChatGPT output in readable form in a Juypter Notebook

Update: This post is still useful, but there's also a way to enable line wrap in Visual Studio Code's Jupyter extension. You can also use the Python Print function instead of the Display example in my original post. For example:

output = client.completions.create(

    model="gpt-3.5-turbo-instruct",

    prompt="List the days of the week: ",

    max_tokens=100,

    stop = "Saturday", #put this in for fun

)

print(output.choices[0].text)

Note these print parameters are specific to this particular object's structure. (I think JSON but I'm a newbie.)

Original below

--------

This was a bit of a revelation. I don't know Python but I've been working through a ChatGPT / LLM tutorial using Visual Studio Code and a Juypter Notebook on macOS. In a Jupyter cell the output renders below the cell and it looks like this:

Completion(id='cmpl-....', choices=[CompletionChoice(finish_reason='stop', index=0, logprobs=None, text='1...

All in one unreadable line with \n as a paragraph deliver and no line wrap.

I asked ChatGPT 4 to help. Over a series of interactions I tried different things and got various error messages I passed to ChatGPT 4. In turn it analyzed my error message and suggested fixes.

This is what I ended up with in about 15 minutes, here added to a cell that ran a simple prompt query

from IPython.display import display, HTML

output = client.completions.create(

    model="gpt-3.5-turbo-instruct",

    prompt="write me a poem",

    max_tokens=100,

    n=3

)

text_content = output.choices[0].text if output.choices else ""

html_output = text_content.replace('\n', '<br>')

display(HTML(html_output))

This is what the output looks like now (the poetry is greeting card quality and mildly painful):

A poem, a weave of words and rhyme

A tapestry of thoughts and time

A magic spell from the poet's pen

A story of love, of loss, of when

The stars above, they guide my hand

As I write of distant lands

Of fiery sunsets and ocean tides

Of moments we hold and let slip by ...   

[Adolescent poetry truncated] 

This screenshot shows it best ...

We are in a new world.

AppleScripts to speed Java compile and execute with BBEdit

Students are often asked not to use an IDE (I like Visual Studio Java) for their Java projects. Instead they need to use BBEdit without an IDE. BBEdit does not appear to natively support facilitated compile/execute but it does run AppleScrips that can speed things a bit.

The davalign.com site has some scripts written for TextWranger. If you rename them to BBEdit they work well.

Reproducing them here in case that site vanishes. I didn't see much like this. Note if you open and save in macOS Script Editor they will be compiled. I like that the first action is to save the document, it's easy to forget to save before a compile.

Compile java.scpt

tell application "BBEdit"

save text document 1

set the_file to file of text document 1

end tell

set AppleScript's text item delimiters to ":"

set source_file to the last text item of (the_file as string)

tell application "Finder"

set the_folder to container of file the_file

end tell

tell application "Terminal"

activate

set p to POSIX path of (the_folder as string)

set shell_script to "cd " & (quoted form of p) & ¬

"; javac " & source_file

if (count windows) is 0 then

do script shell_script

else

do script shell_script in the front window

end if

end tell

Run java.scpt

tell application "BBEdit"

set the_file to file of text document 1

end tell

set AppleScript's text item delimiters to ":"

set source_file to the last text item of (the_file as string)

set compiled_file to text 1 thru -6 of source_file

tell application "Finder"

set the_folder to container of file the_file

end tell

tell application "Terminal"

activate

set p to POSIX path of (the_folder as string)

set shell_script to "cd " & (quoted form of p) & ¬

"; java " & compiled_file

if (count windows) is 0 then

do script shell_script

else

do script shell_script in the front window

end if

end tell

Getting Outlook to export Exchange contacts as vCards (vcf) with proper email addresses for use in macOS

In 2009 I wrote about how it was getting harder to move contact information out of Outlook into something else (like macOS Contacts). I wrote about some options, but that's not what I do now.

Here's what I do (tested in Mojave, which I'm still using because Aperture):

1. Create a simple list Contacts view. I usually only want people so I sort by last name. In a few cases last name of people is blank so I fix that.
2. Now create an empty email. Drag Contacts from Outlook's view into the email body. It has to be to the email, dragging to desktop creates a .msg file. It might fail if you do too many so I distribute 300-400 contacts across 4 separate emails. Outlook creates a vCard file as an attachment. It resolves the email too, so instead of an Outlook x400 (?) you get a proper email address.
3. Send the email to your Mac
4. On the Mac download all attachments. They show as VCF files and macOS renders them quite well.  If they have photos the photos show within the card icon. Spotlight indexes them all. You don't even need to bother with dropping them into Contacts (though that's easy to do, you can drop them into your Contacts Groups (folders)).

It's pretty easy if you know the trick. I've not seen it described anywhere else but I'm sure others know it.

Python macOS environments for learners in 2020

My daughter is auditing Coursera's Intro to Python class. It's pretty standard stuff, but I was surprised by the development environment. For macOS there's a non-trivial Python install that requires some unix knowledge, use of the Homebrew package manager, dealing with admin vs non-admin user issues, consideration of pyenv, editing the path, and finally installing Python.

That's a long way from the ease of, say, TurboPascal circa 1983.

I figured there had to be a better way, but Google only found me some pro-level IDEs. It fell to Twitter to clue me in to the modern scene. The 4 good modern options turn out to be:

• Google Colab: absolute easiest and least painful. I believe the Python code executes in the browser, so it's substantially slower than execution directly in macOS.
• Microsoft Visual Studio Code for macOS: this does require the traditional Python install with Homebrew, but it's a very beginner friendly environment. The Python plugin provides Jupyter support.
• Homebrew Jupyter: similar to Colab but like Visual Studio is part of the Homebrew/Python path.
• Azure does Jupyter Notebooks (via @jhovland) at notebooks.azure.com.

Years ago I ran into iPython as a novice environment; turns out it morphed into Juypter.

It's a sign of the times that Google search didn't turn up a blog post with these options. (It won't find this one either, I'm way off Google's radar now.) Once I'd identified the above options however I could do a Google search to find an educational resource that did mention then:

There are many ways to write and execute Python code:

Python tutor (online, visual debugger)

Python interpreter (command line)

Visual Studio Code (editor, good debugger)

Jupyter (notebook)

Google Colab (online, collaborative)

 

During this lab we see all of them and familiarize with the exercises format. For now ignore the exercises zip and proceed reading.

That site is the University of Trento's data science lab course, updated 2019/2020.  The U of Trento was founded in 1962. Reading the wikipedia page it seems to have started out focusing on sociology (and, given the era, was likely a wee bit Left) but now seems to be very tech.

The course material is presumably translated from Italian. It's quite readable though it would benefit from a native speaker updating the GitHub content. Judging by my little test it may be one of the best resources of its kind.

See also:

1. Learning Python for middle school - from 2013

Update 2/19/2024

I came back to Python for course on working with the OpenAI ChatGPT LLM. This time around I used Visual Studio Code with the Jupyter support. I use the default Python PIP package manager but I think Microsoft favors Conda. As of 2024 CoPilot is an option but it is not free.

Micro-Frameworks for web app development

A developer colleague (M.A) sent me a brief list of micro-frameworks organized by language. His list is in the same vein as Microjs: Fantastic Micro-Frameworks and Micro-Libraries for Fun and Profit but quite a bit shorter.

For my own future reference, here's his list organized by server-side language

• Java – Spark or perhaps something old fashion like Tomcat or Spring MVC in Tomcat
• Groovy – Grails or Ratpack
• Javascript – node.js or Meteor
• Ruby – Sinatra
• PHP – PHP
• Python – Django or Bottle

For my own amusement (and perhaps my 14yo) I'd be inclined towards either Django (Python and packaged on DreamHost, my longtime hosting service) or Meteor (he likes).

PS. Clearly the world needs an AppleScript micro-framework. (ok, sick joke)