fullmetaphone 0.1.0

First CRAN release. The package grew out of cagmetaphone, written by Atharv Tyagi.

• double_metaphone() gives full-length primary and secondary Double Metaphone codes. It now takes whole vectors and returns a data frame.
• New metaphone() for the original single-code Metaphone algorithm.
• New sounds_like() to compare two vectors of names.
• New data set us_surnames, the 5000 most common surnames in the 2010 United States Census.
• New vignette with worked examples of deduplication, record linkage and name search.
• New max_length argument to set the code length. The default stays at 32 characters, as in the old version. Use Inf for no limit or 4 for the traditional short codes.
• New by_word argument to code each word of a name separately.
• Missing values now give NA instead of empty strings.
• Accented Latin letters are read as plain letters instead of being dropped.
• Spaces between words are kept while coding, as in the original algorithm, so rules such as “San Jacinto” and “Van” now work.
• Fixed a bug that coded a double L twice in names such as “Cabrillo” and “Gallegos”.