Parsing a directory structure in Java

Question

I have to parse a set of directories given in the following text file:

# Note: The root folder's parent is labelled as "?"
#       Assume all directory has different name
#
A,?
B,A
C,A
D,C
E,C
F,C
G,F

The above file, depicts the directory structure this way:

A
|
+ B
|
+ C
| |
| + D
| |
| + E
| |
| + F
| | |
| | + G

Assuming the lines starting with # are comments, I have the following code right now:

String line;
BufferedReader f = new BufferedReader(new FileReader(new File("directory.txt")));
while ((line = f.readLine()) != null)
{
    if (!line.substring(0, 1).equals("#"))
    {
        String directory, parent;
        directory = line.split(",")[0];
        parent = line.split(",")[1];
        if (parent.equals("?"))
            System.out.println("Directory " + directory + " is the root.");
        else
            System.out.println("Directory " + directory + " found inside " + parent + ".");
    }
}

So, this just does the work of displaying the directories, that too not in a hierarchical way to parse them. It just gives an output of textual representation, this way:

Directory A is the root.
Directory B found inside A.
Directory C found inside A.
Directory D found inside C.
Directory E found inside C.
Directory F found inside C.
Directory G found inside F.

If it was PHP, I can convert it to JSON nodes and can parse the parent or sibling, in a hierarchical way, but I am not sure how am I supposed to make this in Java. Any heads up will be great for me.

For now, I created a class for the tree structure this way:

public class Directory {
    private String name;
    private Directory parent;
}

But I am not sure how do I link the directories as kind of linked lists in the main Java program. Any help here would be appreciated. So, when I do some kind of tree structure here, I would like to achieve something like a directory traversing program.

Say, if I give the input as DirectoryParser C, then, it should output me something like:

C
|
+ D
|
+ E
|
+ F
| |
| + G

Is this possible with the current approach of mine? Can someone please guide me in how to achieve this? Thanks in advance.

Disclaimer: I went through Java tree data-structure?, but I am supposed to get something simple in a single file without using any external plugins. :(

Answer 1

OKay Guys, I approached the same old method for this. I used an Iterator to walk through the ArrayList and used it this way:

static void RecursiveFolderFinder(String Root)
{
    Iterator<Folder> Folder = ListOfFolders.iterator();
    while (Folder.hasNext())
    {
        Folder curFolder = Folder.next();
        if (curFolder.getParent().equals(Root))
        {
            OutputList += " " + curFolder.getName() + " ";
            OutputList += "{";
            RecursiveFolderFinder(curFolder.getName());
            if (OutputList.substring(OutputList.length()-1, OutputList.length()).equals("{"))
                OutputList = OutputList.substring(0, OutputList.length()-1);
            else if (!OutputList.substring(OutputList.length()-1, OutputList.length()).equals(" "))
                OutputList += " }";
            else
                OutputList += "}";
        }
    }
}

I know, this kind of string manipulation is crazy and weak, I used his for making sure, no one else uses such a funny one! Just sarcastic. Thanks for @mrhobo's answer. If anyone could suggest me a better output, awesome then!

Answer 2

Assuming all referenced parent directories are directories that have already been listed on previous lines you could create your directory tree in one pass. Put each directory you create in a map. Then once you create a new directory search for the parent directory in that map.

So feed each line to something like this:

public void parseLine(String line) {
    if (isComment.matcher(line).matches()) {
        return; // comment, skip line
    }

    Matcher m = dirPattern.matcher(line);
    if (!m.matches()) {
        throw new InvalidSyntaxException("line " + line + " does not follow dir syntax.");
    }

    String dirName = m.group(0);
    String parentDirName = m.group(1);
    if (parentDirName == "?") {
        directories.put(dirName, new Directory(dirName));
    } else {
        Directory parentDir = directories.get(parentDirName);
        if (parentDir == null) {
            throw new DirectoryNotFoundException("Directory " + parentDir + " not found");
        }
        directories.put(dirName, new Directory(dirName, parentDir));
    }
}

where Directory is:

public class Directory {
    private String name;
    private Directory parent;
    private Set<Directory> children = new HashSet<Directory>();

    public Directory(String n) {
        this(n, null);
    }

    public Directory(String n, Directory p) {
        name = n;
        parent = p;
        if (parent != null) {
            parent.addChild(this);
        }
    }

    // keep private to ensure no child is set without a proper parent
    private void addChild(Directory child) {
        children.add(child);
    }

    @Override
    public String toString() {
        String str = name;
        for (Directory child : children) {
            str += child.toString();
        }
        return str;
    }
            .... be sure to implement equals and hashCode ...
}

Notes:

• if aforementioned order is not there you may achieve the same goal in two passes. One pass to get all the directories and the second to set all the parents.
• this assumes a \w+,\w+ format and unique directory names
• you can use toString to generate your tree