Back to manual main page

The program uses hierarchical keywords in order to find hierarchical tag names for the defined faces. The implementation is inspired by the hierarchical keyword management available in Lightroom and should be compatible with it.

Hierarchical keywords are organized as a keyword tree. The elements of a keyword tree are categories, keywords and synonyms. The keyword tree can be loaded from and saved to an ordinary text file.

Keywords in the keyword tree are just strings. Keywords may have subtrees. The text representation for keywords in the text file is just the keyword string itself.

Categories are just the same like keywords and are used to create the structure of the tree without being used for tagging. They are ignored when calculating the keyword path and the keyword cloud for a keyword. The text representation for a category in the text file is the category name enclosed in square brackets.

Synonyms are additional keywords for a keyword. They cannot have subelements (children) and they are used to compute the keyword cloud for a given keyword. The text representation for a category in the text file is the synonym name enclosed in curly brackets.

The tree structure is defined by identation in text file. Even though the text file reader is able to parse space idented text, it's the best to use TAB characters for identation.

Let's have some examples:

{keyword test}
People
	{Picasa faces}
	Relatives
		[Families]
			Family One
				{The Kelly family}
				Michael Maroon
					{brother}
					{brown eyes}
				Paula Purple
					{sister}
			Others
				Betty Blue
	Friends
		{my friends}
		Bob Green
			{best friend}
	Collegues
		Joe Black

The keyword path for a given keyword is just walking up the tree and collecting all parent node names. Category names are skipped and therefore not used.

Let's see what the hierarchical keyword paths are:

The keyword cloud for a given keyword consist of all keywords in the keyword path and their synonyms (including the synonyms of the root element, which are used for all keywords).

For example, the keyword cloud for “Michael Maroon” is computed as follows:

• “Michael Maroon” (the keyword itself)
• “brother”, “brown eyes” (the synonyms for “Michael Maroon”
• “Family One” (parent of “Michael Maroon”)
• “The Kelly family” (synonym for “Family One”)
• “Relatives” (parent of “Family One”)
• “People” (parent of “Relatives”)
• “Picasa faces” (synonym for “People”)
• “keyword test” (synonym for the root node, they apply to all keywords)

As a result, the keyword cloud for “Michael Maroon” is “keyword test”, “Picasa faces”, “People”, “Relatives”, “The Kelly family”, “Family One”, “brother”, “brown eyes”, “Michael Maroon”.

On the left side you can see the tree view showing the hierarchical keyword tree. Categories have a “C” icon, keywords have a “K” icon and synonyms have a “S” icon. The root node has a “F” icon and it's label shows the current filename associated with the current keyword tree.

There is a right button context menu. With this menu you can easily change the type of a tree element, create a new element, edit an existing element or delete an element.

You can move tree nodes using drag&drop. When the SHIFT key is pressed while dragging, the nodes are moved, otherwise the nodes are copied. It's possible to delete selected notes by pressing the DELETE key.

On the right side there is a listbox containing the Picasa face names (taken from the Picasa contacts file) which are not yet part of the keyword tree. You can add them by drag&drop.

The “Full path” textbox shows the full path of the current selected keyword (including the categories in the path).

The “Hierarchical path” textbox shows the hierarchical path which will be used for tagging.

Above the “Full path” textbox on the right side there is a checkbox labeled “Show keyword cloud”. When this box is checked, a small window opens and shows the keyword cloud for the selected keyword. This keyword cloud is used for tagging (if it's set as an tagging option).

On the bottom there are several buttons for loading and saving keyword trees from a text file.

With the “New” button, the current keyword tree is just deleted and you can start from scratch to create a new keyword tree. When pressing the “Open…” button a dialog box opens where a text file can be chosen which is then loaded when you press OK. When using the “Add…” button, the selected text file will be read and appended to the current keyword tree. When pressing the “Save” button, the current keyword tree will be saved to the associated file (shown as the root node in the keyword tree). With the “Save as…” you can save the current keyword tree to a file you specify. In this case, the filename associated with the current tree will be set to this new filename.