Reports coverage given a set of unicode values
Smörgåsbord tests coverage over unicode character sets.
The Smorgasbord class inherits from UnicodeSet and supports the same features.
Supports Python 2.6 – 3.x
$ pip install smorgasbord
from smorgasbord import Smorgasbord # Provide a path to a file or folder of character sets Smorgasbord.paths.prepend("/my/path/to/language/en.txt") >>> bord = Smorgasbord([97, "b", "c", u"ü", u"\u0660"]) Smorgasbord([u"a", u"c", u"b", u"\xfc", u"\u0660"]) # Reports are accessed though the "reports" dict using the language code >>> en = bord.reports["en"] # Basic information about the report's language is accessible >>> en.language.code "en" >>> en.language.name "English" >>> en.language.characters FrozenUnicodeSet([u"a", u"b", u"c", ...]) # Amount of coverage is availbe as float and string representations >>> en.coverage 0.057 >>> en.coverage.percentage u"5.7%" # Sets of glyphs can be accessed >>> en.covered FrozenUnicodeSet([u"a", u"b", u"c"]) >>> en.uncovered FrozenUnicodeSet([u"d", u"e", u"f", ...]) # Reports can also return a boolean value for completeness: >>> en.complete False >>> en.incomplete True
Character Set Files
Each character set is defined in a text file. Characters are separated by spaces, and lines starting with a # are ignored as comments. Line breaks can and should be used to wrap lines to 80 characters maximum.
For example, an en.txt definition for an English coverage character set:
# Language: English a b c d e f g h i j k l m n o p q r s t u v w x y z A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
The first line is a special comment that will be parsed as the language name. Other special comments may be added in the future, but for now only Language is supported.
Supplying Character Sets
Character sets are made available by supplying a path to a folder, or directly to a language file.
Character set files are searched for in each succesive folder, using the first matching file.
Alternatively, the paths array can be replaced entirely:
Smorgasbord.paths = ["/my/path/to/language/files/dir"]
This is a quick list of features that will need to be added in the near future (and will probably comprise a 1.0 release).
- Lazily evaluate reports. Currently the library loads all language files when a Smorgasbord is initialized, which will get slow, fast. This should happen at the latest possible moment.
- Unicode ranges in language files. Adding support for unicode ranges will probably be necesary for languages with large character sets.