According to the manual, magics is just normal functions or member functions of derived class of IPython.core.magic.Magics and should be registered using
get_ipython().register_magics(MyMagics)
My questions:
where should I put those class/function definitions? Should I add their path to $PYTHONPATH?
where is the standard place to register them?
is there a folder to hold all user-defined magics and is automatically loaded when IPython is launched?
Please completly read the page you are linking to, in particular at the top, is written in red:
This documentation is for an old version of IPython. You can find docs for newer versions [here].
By following this link you can find the new documentation, in particular this page.
On this page you will find things like:
Below we show code you can place in a file that you load from your configuration, such as any file in the startup subdirectory of your default IPython profile.
And
Here is a full example of a magic package. You can distribute magics using setuptools, distutils, or any other distribution tools like flit for pure Python packages.
To summarize:
Magics can be defined anywhere as long as they can be imported.
Usually author defined magics in packages that define the load_ipython_extension function, making them loadable with %load_ext, and the load_ipython_extension register the magic.
You can load extension by default using the configuration file and the c.InteractiveShellApp.extensions option.
Related
I have stored a JavaScript file in a TYPO3 extension (under Resources/Public/Css/JavaScript/Backend/Iconpack.js), which is used as a JavaScript module in the backend.
However, when developing the extension, I packed the file full of console.log() to allow debugging, and also to better document certain things.
Of course I want that in the final extension as well as when downloading via composer, only a minimized version of this JavaScript is loaded, but not the original source file.
On the other hand, of course, I want the original source file to be stored somewhere in the git repository, so that you can actually continue working on it.
In the frontend, I would normally just store Iconpack.js, for example, in the same folder, create a minimized version from it, i.e. Iconpack.min.js, and use this minimized file.
But unfortunately that doesn't work in this case, because the file path for JavaScript modules is also the namespace.
What would be the best way to solve this?
To summarize:
The minimized version should be able to be used directly in the backend via Iconpack.js.
The namespace should be TYPO3/CMS/Iconpack/Backend/Iconpack and not TYPO3/CMS/Iconpack/Backend/Iconpack.min or similar.
The source file should not end up in TER or be loaded via composer.
But the source file should be accessible via Github.
I was able to find out that I can at least prevent a certain file from being loaded via composer with .gitattributes.
Unfortunately, that doesn't solve the question of where I should best place it so that other developers have it easier and it's clear that it's the source file.
Currently I only have the minimized version stored in git for this reason, but that makes little sense for an open source repository: https://github.com/quellenform/t3x-iconpack/tree/main/Resources/Public/JavaScript/Backend
Ideas?
Is there a way to make a user created class/package to be read only in Modelica preferably through annotations? Make Modelica class read-only in Dymola gives a Dymola option, I am using OpenModelica and is required to verify a package across its two different versions, since both these versions are editable I am unknowingly making modifications in the older version. Thanks in advance.
I tried to search OpenModelica documentation to see if any OM specific annotations are available. But I couldn't spot them. I am pretty much sure that I have missed it in them, probably used a bad keyword.
In OMEdit there are two ways to open a library: either use "File->System Libraries" or "File->Open Modelica/Library File(s)".
The system libraries only shows packages installed at $HOME/.openmodelica/libraries (on Linux; other path in Windows). These are installed by the package manager or installed there manually. When loading libraries through "File->System Libraries", they are always read-only.
If you load the same library by pointing out the package.mo file, it is opened writable.
You can mark a class read-only in the filesystem by not making the file writable (and if you use a hierarchical file structure instead of the whole library in one file, you can restrict editing to only certain parts in this way).
When loading an encrypted library, it is possible to prevent certain operations on the package using annotations, but editing is always restricted.
I have a few utility functions, snippets and scratches that I want to be able to use in every project.
Currently I have the following setup for Clojure projects and Intellij IDEA/Cursive:
I have a user profile defined in .lein, where I have source-paths pointing to
when I sync and run REPL for project in IDEA/Cursive, I check that :user profile is selected (it is the default)
Cursive shows both project files and common files (i.e. scratches folder) in project pane
I can edit and eval in REPL real project files and my local helers/scratchpads seamlessly.
Is there a way to achieve this behaviour in VSCode?
Notes:
My clojure setup in Cursive is for illustration purposes only. I would like to find a way to get similar feature in other scripting languages (i.e. python, groovy). So ideally I don't want to use leiningen for that, but rather find a generic way to add common source folder to VS code Explorer pane.
I know that with multi-root workspaces I can achieve what I want, but it must be done manually per project. It would be much better if my folder with common utils was added to every workspace automatically. Something like a default workspace template would solve this, but I couldn't find anything similar. Am I missing something?
I'm developing an extension in VS Code to add language support for OpenSCAD (Script-based 3D modeling program). Currently, I have been working on a way to open / preview a file in OpenSCAD from VS Code, which I have been able to do successfully using my own preview manager.
My issue is that I want to add configurable naming formats when exporting an OpenSCAD file that use environment variables similar to those used in the tasks.json file. More info can be found here: https://code.visualstudio.com/docs/editor/variables-reference. As an example, taking the file test.scad and the export configuration ${fileBasenameNoExtension}.stl would export to the file test.stl.
Additionally, I want to add a custom variable, ${#} that would evaluate after all other variables as a unique version number to avoid duplicate exported files. Using the example file: test.scad and the export configuration, ${fileBasenameNoExtension}_${#}.stl, the extension would export to the file test_1.stl for the first time. Then, seeing that test_1.stl exists, it would export to test_2.stl, and so on. I implement similar functionality in all of my exporting utilities, so it is important I can implement it here.
Now that the intro is done, on to the actual question: To anyone who knows more about the VS Code API than I do, in order to best get the functionality described above, should I implement environment variable evaluation into my custom preview manager or reimplement the preview manager I have using tasks? Because I have already implemented my own preview manager that I am happy with, I would prefer to do the former. However, I have been unable to find any functions in the VS Code API that will evaluate the environment variables in a string. Is there a typescript function to evaluate environment variables in the API that I have missed?
If re-implementing this functionality using tasks is a better way to achieve my goal, would I have to sacrifice the control I have in my preview manager, such as being able to selectively kill open previews and dispalying active exports?
Or, is there a compromise that could use all of the power from tasks without losing any functionality I've already developed?
Link to branch of my extension's repository: https://github.com/Antyos/vscode-openscad/tree/PreviewModel
Is there a generic way to prevent a given set of files from being parsed by IntelliSense modules while it still shows up elsewhere?
For instance, I have a third-party library that provides an AppModel class I can customise. Such class is defined in several places:
The library bundles a template to copy from.
The library includes a usage example.
The library implements the class in a helper tool.
The library defines the class for its test harness.
I implement the class in my app.
My implementation is the only one that gets loaded when the app runs and the only one I care about when I navigate my code or use Go to Symbol yet I need to pick among the 5 definitions every time.
For languages where there isn't a specific setting, the only way to go appears to be files.exclude. But that directive basically makes the file or directory completely disappear from the program. It won't even show up in the Explorer pane. That feature seems designed for source control metadata and similar stuff and I don't think it's entirely convenient for third-party libraries.
Any idea?