templang

A simple annotation-based language to generate code from templates

Templates

Templates are just source files with annotations. For example:

# BEGIN
from typing import Literal

Rank = Literal[
  # LOOP RANK
  'RANK',
  # END
]

CLEARANCE_LEVEL = int # DELETE (just to make mypy happy)
# LOOP RANK CLEARANCE_LEVEL
class RANK:
  clearance_level: CLEARANCE_LEVEL

# END
# END

Template Parsing

from templang import parse
with open('ranks.py') as f:
  source = f.read()

code = parse(source, translations={
  'RANK': ['Captain', 'Lieutenant', 'Sergeant'],
  'CLEARANCE_LEVEL': ['1', '2', '3']
})

print(code)

# from typing import Literal
# 
# Rank = Literal[
#   'Captain',
#   'Lieutenant',
#   'Sergeant',
# ]
# 
# class Captain:
#   clearance_level: 1
# 
# class Lieutenant:
#   clearance_level: 2
# 
# class Sergeant:
#   clearance_level: 3

Rules

Now, that's most of what you need to now. But, to be precise, here are the rules:

• # BEGIN and # END delimit a template

• Lines starting with # UNCOMMENT are stripped to whatever proceeds the annotation

• Lines ending with # DELETE are removed

• translations: Mapping[str, str | Sequence[str]] is the dictionary of translations:

  • If value = translations[key] is a string, all instances of key are replaced by value
  • If value = translations[key] is a sequence of strings, key will be allowed in a # LOOP declaration
  • The order of translations is not guaranteed. So, keys substrings of other keys will probably yield inentended results

• # LOOP VAR1 ... VARN and # END delimit a loop:

  • translations[VARi] must be a sequence of strings
  • LEN = len(translations[VAR1]) == ... == len(translations[VARN])
  • The template inside the loop is substituted with translations[VARi][j] for each j in range(LEN)
  • Nested loops are allowed
  • If nested loops involve a same variable VAR multiple times, then translations[VAR] must be a nested sequence of strings (matching the loop depth)

    • E.g:

      # BEGIN
      from typing import Union, Literal
      
      Marriage = Union[
      # LOOP NAMES
        tuple [
        # LOOP NAMES
          Literal['NAMES'],
        # END
        ],
      # END
      ]
      # END
      
      
      parse(open(__file__).read(), {
        'NAMES': [['John', 'Jane'], ['Stuart', 'Alice']]
      })
      
      # from typing import Union, Literal
      #
      # Marriage = Union[
      #   tuple [
      #     Literal['John'],
      #     Literal['Jane'],
      #   ],
      #   tuple [
      #     Literal['Stuart'],
      #     Literal['Alice'],
      #   ],
      # ]
      

• Oh, and actually, # can be replaced by any custom string (e.g. //)