Since package names can be arbitrary and may begin with numberals (7zip) or contain characters that are not great as classnames (c++, C#), a function is needed to remap those names to more "safe" alternatives, in a way that maintains the following properties:
- readability; the new name should be readable, as much as possible
- reversibility & mapping: it should reflect well the original name
- standardization: it should follow established practice, if any
- extendibility: the logic should be extendable for future needs
- nameclash-free: class names are prefixed with "EB_", by default
The function encode_class_name returns the encoded version of a class name
and is in turned based on an EB_ prefix and the encode_string function.
For the programmer, this means that if, for example, you are extending
something called C++, your class name likely is EB_C_plus__plus_.
Samewise:
| package name | class name |
|---|---|
| 7zip | EB_7zip |
| C# | EB_C_hash_ |
| Charm++ | EB_Charm_plus__plus_ |
| R | EB_R |
| r | EB_r |
| map\reduce | EB_map_backslash_reduce |
The encoding function can handle even more complicated package names, like:
- name:
0_foo+0x0x#-$__ - becomes:
0_underscore_foo_plus_0x0x_hash__minus__dollar__underscore__underscore_
It has been inspired by the concepts seen at: (but in lowercase style)
- http://fossies.org/dox/netcdf-4.2.1.1/escapes_8c_source.html
- http://celldesigner.org/help/CDH_Species_01.html
- http://research.cs.berkeley.edu/project/sbp/darcsrepo-no-longer-updated/src/edu/berkeley/sbp/misc/ReflectiveWalker.java and can be extended freely as per ISO/IEC 10646:2012 / Unicode 6.1 names:
- http://www.unicode.org/versions/Unicode6.1.0/
For readability of >2 words, it is suggested to use _CamelCase_ style.
So, yes,
_GreekSmallLetterEtaWithPsiliAndOxia_could indeed be a fully valid package name; package "electron" in the original spelling anyone? ;-)
>>> from easybuild.tools.filetools import encode_class_name
>>> encode_class_name('GAMESS-US')
'EB_GAMESS_minus_US'