Vitessce & Easy_vitessce: Basis-Specific Embedding Functions

Hey guys! Today, let's dive into a fascinating discussion about basis-specific embedding functions, particularly in the context of tools like Vitessce and easy_vitessce. We're going to explore a streamlined approach to handling dimensionality reduction techniques like UMAP, PCA, and t-SNE. The core idea revolves around centralizing the plotting logic within a main sc.pl.embedding function and then creating handy wrapper functions for each specific technique. This not only makes our code cleaner and more maintainable but also provides a consistent interface for users. So, buckle up, and let's get started!

The Power of Centralized Embedding Logic

At the heart of our discussion is the concept of a centralized embedding function. By consolidating the core logic for plotting UMAP, PCA, and t-SNE into a single function, sc.pl.embedding, we achieve several key benefits. First and foremost, it promotes code reusability. Imagine having to duplicate the same plotting code across multiple functions – that's a recipe for maintenance nightmares! With a centralized approach, we write the code once and reuse it across different embedding techniques. This dramatically reduces the risk of errors and inconsistencies.

Secondly, a centralized function makes it easier to maintain and update our plotting logic. If we need to tweak the appearance of our plots, add new features, or fix bugs, we only need to modify the sc.pl.embedding function. This change will automatically propagate to all the embedding techniques that use it. This reduces the workload and time consumption that comes with changing something consistently throughout an application. This is a huge win for maintainability and scalability.

Furthermore, centralizing the logic allows us to enforce a consistent style and behavior across different embedding plots. This is crucial for providing a seamless user experience. Users shouldn't have to learn different plotting conventions for UMAP, PCA, and t-SNE. By using a single underlying function, we can ensure that all plots adhere to the same standards. This consistency extends to things like color schemes, axis labels, and plot titles, making it easier for users to interpret and compare results.

Wrapper Functions: A User-Friendly Interface

Now, let's talk about wrapper functions. While a centralized sc.pl.embedding function is great for internal code organization, it's not the most user-friendly interface. Users typically want to call specific functions like sc.pl.umap, sc.pl.tsne, and sc.pl.pca directly, rather than having to specify the embedding basis every time. This is where wrapper functions come in.

The beauty of wrapper functions lies in their simplicity. They act as convenient shortcuts for calling the sc.pl.embedding function with the appropriate basis argument. For example, the umap wrapper function would simply call sc.pl.embedding with `basis=