Rather than the simplistic: [Word] [URL Pattern] One would have [word] [type] [summary field and previously defined word dropdown] [delim] [regex]

Here [type] is a dropdown with three values temporary, summary, and meta.

summary field is any field that is normally extracted by Yioop: for example, URL, TITLE, DESCRIPTION, etc.

previously defined word is any previously defined word string

This dropdown should support selecting multiple items.

delim is a delimeter which should be used when concatenating multiple fields.

regex is any perl compatible regex.

To determine the word string corresponding to this rule, one starts with the text that results from concatenating the selected fields. Then one does a match using the regex and the matching text is the word string.

Words whose type is temp become neither new summary fields nor meta words but can be used in later words

Those whose name is summary become new summary field for crawl items

Those whose name begins with meta become new meta words to be used in indexing.

This kind of functionality was achieved in the Page Rules portion of Page Options. The result is actually more flexible and described in Version 0.94 and higher of the Yioop documentation.

Implemented by 798b78de