python data structures

Given some string data comprising a series of headings
with a variable number of items under each heading.
As for example directory names and filename items.
What is a suitable Python structure for the Data?
It does not seem to map onto any of the four built-in
collection data types.

A dictionary of lists might be a possibility?

With C++ containers, I would have used a map < string, vector >
A map is synonymous with dictionary?

I looked at using list of lists and dictionary of lists as suggested by Steve, but would prefer something a bit more specifically tailored to tree-structure data.

There is a PIP installable package called Bigtree which looks hopeful.
It installs and can be imported.

Unfortunately it reports an error when it trying to draw a tree at the terminal because of missing Unicode characters.

If it is actually a file system you are concerned with, Python has a number of existing routines for handling such things. But they’re for querying an existing structure rather than for modelling one in memory.

Does your data structure need to handle a hierarchical structure, I.e. where directories can contain other directories?

Does your data structure need to handle a hierarchical structure, I.e. where directories can contain other directories?

The structure has only two layers. First layer directory names, second layer leaf names of runable programs

The purpose of the program is to list all the RISC OS executables.
It starts with Run$Path and enumerates each of the directories in it.
A data structure is not strictly necessary but it makes sense if I want to go beyond just printing out the results.

If you need to go beyond just storing filenames, I would suggest investigating DataClasses (I’m assuming that the RISC OS port supports them).

@ John

If i understand your question correctly then, in Python, the best approach you have should be to use nested dictionary:

john_data = {
    "Heading1": ["item1", "item2", "item3"],
    "Heading2": ["item4", "item5"],
    "Heading3": ["item6", "item7", "item8", "item9"]
}

What the above means:

• Each key (first column, aka before the “:”) represents your heading.
• The value corresponding to each key is a list containing the items under that heading.

HTH

@ Jean

A map is synonymous with dictionary?

In many ways, yes, a C++ std::map is similar to a Python dictionary, but there are also some differences. So, I would not really call it synonymous, but more like similar concept if I am making sense.

@ Steve

If you need to go beyond just storing filenames

Good idea, and I would say: I would invest in learning DataClasses anyway. It should work fine on Python for RO as long as a library is pure python and has dependecies that are pure python it should work straight out of the box (some lib may have some issues if they make certain assumptions like with syscalls and/or file paths with ENV variables etc., but DataClasses should not have any of this IIRC)

If you need to go beyond just storing filenames, I would suggest investigating DataClasses (I’m assuming that the RISC OS port supports them).

Thanks Steve and Paolo. I really did not did not want to hear that:-)

I usually start writing top down linear procedural code, and after 10 lines or so start thinking about defining a function or two.

For this particual program I had achieved 90 percent of what I needed in 20 lines code. Unfortunately it was at this point that I began to realize that further development would require a rethink about the approach I had taken. The data needed structure.
Data Classes came to mind, but as OOP is not something I am comfortable with I was reluctant to go back down the rabbithole and find myself in a maze of dark twisty passages.

But you have convinced me. I know it it the right way to go so will bite the bullet, study Data Classes, light a smoky torch and climb down the OOP hole.

Version 1.0 release may be some time.

Unfortunately it reports an error when it trying to draw a tree at the terminal because of missing Unicode characters.

Hi, creator of bigtree here. May I know how you are creating a tree and if this error can be replicated? Feel free to raise an issue here as well!

Hi, creator of bigtree here. May I know how you are creating a tree and if this error can be replicated?

Hello Kay It is a pleasant surprise to hear from the developer, but I am surprised you managed to find your way into this little RISC OS backwater:)

The problem is reproducible on a machine running RISC OS with python 3.8 installed.

from bigtree import Node
root = Node("a", age=90)
b = Node("b", age=65, parent=root)
c = Node("c", age=60, parent=root)
root.hshow()

Running the above gives error output:

Traceback (most recent call last):
File "...sandbox.Bigtree.eg0/py", line 7, in <module>
    root.hshow()
File "...bigtree.node.node/py", line 222, in hshow
    hprint_tree(self, **kwargs)
File "...bigtree.tree.export/py", line 476, in hprint_tree
    print("\n".join(result))
UnicodeEncodeError: 'latin-1' codec can't encode characters in position 5-6:
    ordinal not in range(256)

I don’t think the problem is in your code but rather in the RISC OS Python implementation. RISC OS Python is still a bit raw and RISC OS does not have comprehensive UniCode support.

RISC OS Python is still a bit raw and RISC OS does not have comprehensive UniCode support.

I was just googling around to see how I can support my own package haha I see.. I do not have RISC OS so I am unable to replicate the error on my machine, unfortunately.

I see that printing out the characters is the issue – there is another function that can extract the string – and from there you can perform any encoding as you wish. In the example below, I tried using utf-8 encoding but the results are not “pretty” though since they got encoded to byte format

from bigtree import hyield_tree
result = hyield_tree(root)
result_string = "\n".join(result)  # this was the line that caused error when you print it directly
print(result_string.encode("utf-8"))

Hope this can be a workaround :)