PyTree A Generic Tree Object Viewer

Up to now, this chapter has been command-line-oriented. To wrap up, I want to show you a program that merges the GUI technology we studied earlier in the book with some of the data structure ideas we've met in this chapter.

This program is called PyTree, a generic tree data structure viewer written in Python with the Tkinter GUI library. PyTree sketches out the nodes of a tree on screen as boxes connected by arrows. It also knows how to route mouseclicks on drawn tree nodes back to the tree, to trigger tree-specific actions. Because PyTree lets you visualize the structure of the tree generated by a set of parameters, it's a fun way explore tree-based algorithms.

PyTree supports arbitrary tree types by "wrapping" real trees in interface objects. The interface objects implement a standard protocol by communicating with the underlying tree object. For the purposes of this chapter, PyTree is instrumented to display binary search trees ; for the next chapter, it's also set up to render expression parse trees. New trees can be viewed by coding wrapper classes to interface to new tree types.

The GUI interfaces PyTree utilizes were covered in depth earlier in this book, so I won't go over this code in much detail here. See Part II, for background details, and be sure to run this program on your own computer to get a better feel for its operation. Because it is written with Python and Tkinter, it should be portable to Windows, Unix, and Macs.

17.10.1 Running PyTree

Before we get to the code, let's see what PyTree looks like. You can launch PyTree from the PyDemos launcher bar (see the top-level of the examples distribution source tree at http://examples.oreilly.com/python2), or by directly running the treeview.py file listed in Example 17-27. Figure 17-2 shows PyTree in action displaying the binary tree created by the "testl" button. Trees are sketched as labels embedded in a canvas, and connected by lines with arrows. The lines reflect parent-to-child relationships in the actual tree; PyTree attempts to layout the tree to produce a more or less uniform display like this one.

Figure 17-2. PyTree viewing a binary search tree (test1)

Figure 17-2. PyTree viewing a binary search tree (test1)

Python Object Oriented Canvas

PyTree's window consists of a canvas with vertical and horizontal scrolls, and a set of controls at the bottom -- radiobuttons for picking the type of tree you wish to display, a set of buttons that trigger canned tree drawing tests, and an input field for typing text to specify and generate a new tree. The set of test buttons changes if you pick the Parser radiobutton (you get one less test button); PyTree use widget pack_forget and pack methods to hide and show tree-specific buttons on the fly.

When you pick one of the canned test buttons, it displays in the input field the string you would type to generate the tree drawn. For binary trees, type a list of values separated by spaces and press the "input" button or the Enter key to generate a new tree; the new tree is the result of inserting the typed values from left to right. For parse trees, you input an expression string in the input field instead (more on this later). Figure 17-3 shows the result of typing a set of values into the input field and submitting; the resulting binary tree shows up in the canvas.

Figure 17-3. A binary tree typed manually with on-click pop-up

Figure 17-3. A binary tree typed manually with on-click pop-up

Input Field Tree

Notice the pop-up in this screen shot; left-clicking on a displayed tree node with your mouse runs whatever action a tree wrapper class defines, and displays its result in the pop-up. Binary trees have no action to run, so we get a default message in the pop-up, but parse tress use the mouseclick to evaluate the subtree rooted at the clicked node (again, more on parse trees later).

Just for fun, maximize this window and press the "test4" button -- it inserts 100 numbers from through 99 into a new binary tree at random, and displays the result. Figure 17-4 captures one portion of this tree; it's much too large to fit on one screen (or on one book page), but you can move around the tree with the canvas scrollbars.

0 0

Responses

  • stanislao lucciano
    How to draw a node after a capture button in tkinter python?
    5 years ago

Post a comment