Customizing the Parser with Your Own Configuration — Nameparser 1.1.3 documentation (original) (raw)

Recognition of titles, prefixes, suffixes and conjunctions is handled by matching the lower case characters of a name piece with pre-defined sets of strings located in nameparser.config. You can adjust these predefined sets to help fine tune the parser for your dataset.

Changing the Parser Constants¶

There are a few ways to adjust the parser configuration depending on your needs. The config is available in two places.

The first is via from nameparser.config import CONSTANTS.

from nameparser.config import CONSTANTS CONSTANTS <Constants() instance>

The other is the C attribute of a HumanName instance, e.g.hn.C.

from nameparser import HumanName hn = HumanName("Dean Robert Johns") hn.C <Constants() instance>

Both places are usually a reference to the same shared module-levelCONSTANTS instance, depending on how you instantiate the HumanName class (see below).

Editable attributes of nameparser.config.CONSTANTS¶

• TITLES - Pieces that come before the name. Includes all first_name_titles. Cannot include things that may be first names.
• FIRST_NAME_TITLES - Titles that, when followed by a single name, that name is a first name, e.g. “King David”.
• SUFFIX_ACRONYMS - Pieces that come at the end of the name that may or may not have periods separating the letters, e.g. “m.d.”.
• SUFFIX_NOT_ACRONYMS - Pieces that come at the end of the name that never have periods separating the letters, e.g. “Jr.”.
• CONJUNCTIONS - Connectors like “and” that join the preceding piece to the following piece.
• PREFIXES - Connectors like “del” and “bin” that join to the following piece but not the preceding, similar to titles but can appear anywhere in the name.
• CAPITALIZATION_EXCEPTIONS - Dictionary of pieces that do not capitalize the first letter, e.g. “Ph.D”.
• REGEXES - Regular expressions used to find words, initials, nicknames, etc.

Each set of constants comes with add() and remove() methods for tuning the constants for your project. These methods automatically lower case and remove punctuation to normalize them for comparison.

Other editable attributes¶

• string_format - controls output from str()
• empty_attribute_default - value returned by empty attributes, defaults to empty string
• capitalize_name - If set, applies capitalize() to HumanName instance.
• force_mixed_case_capitalization - If set, forces the capitalization of mixed case strings when capitalize() is called.

Parser Customization Examples¶

Removing a Title¶

Take a look at the nameparser.config documentation to see what’s in the constants. Here’s a quick walk through of some examples where you might want to adjust them.

“Hon” is a common abbreviation for “Honorable”, a title used when addressing judges, and is included in the default tiles constants. This means it will never be considered a first name, because titles are the pieces before first names.

But “Hon” is also sometimes a first name. If your dataset contains more “Hon”s than “Honorable”s, you may wish to remove it from the titles constant so that “Hon” can be parsed as a first name.

from nameparser import HumanName hn = HumanName("Hon Solo") hn <HumanName : [ title: 'Hon' first: '' middle: '' last: 'Solo' suffix: '' nickname: '' ]> from nameparser.config import CONSTANTS CONSTANTS.titles.remove('hon') SetManager({'right', ..., 'tax'}) hn = HumanName("Hon Solo") hn <HumanName : [ title: '' first: 'Hon' middle: '' last: 'Solo' suffix: '' nickname: '' ]>

If you don’t want to detect any titles at all, you can remove all of them:

CONSTANTS.titles.remove(*CONSTANTS.titles)

Adding a Title¶

You can also pass a Constants instance to HumanName on instantiation.

“Dean” is a common first name so it is not included in the default titles constant. But in some contexts it is more common as a title. If you would like “Dean” to be parsed as a title, simply add it to the titles constant.

You can pass multiple strings to both the add()and remove()methods and each string will be added or removed. Both functions automatically normalize the strings for the parser’s comparison method by making them lower case and removing periods.

from nameparser import HumanName from nameparser.config import Constants constants = Constants() constants.titles.add('dean', 'Chemistry') SetManager({'right', ..., 'tax'}) hn = HumanName("Assoc Dean of Chemistry Robert Johns", constants=constants) hn <HumanName : [ title: 'Assoc Dean of Chemistry' first: 'Robert' middle: '' last: 'Johns' suffix: '' nickname: '' ]>