mnemonic_port/rebuild/build_inline.py

#!/usr/bin/env python3
"""
Build an inlined variant of mnemonicode.py with the wordlist embedded as a
Python literal instead of loaded from wordlist.json at runtime.

Usage:
  python3 build_inline.py [source.py] [wordlist.json] [output.py]

Defaults:
  source      mnemonicode.py
  wordlist    wordlist.json
  output      mnemonicode_inline.py
"""

import json
import sys
import textwrap
from pathlib import Path

COLS = 79  # target line length for the word list

# Markers that delimit the section to replace in the source file.
IMPORT_JSON   = "import json\n"
IMPORT_PATH   = "from pathlib import Path\n"
WORDLIST_LINE = "_WORDLIST_JSON = Path(__file__).parent / \"wordlist.json\"\n"
LOAD_OLD = '''\
@lru_cache(maxsize=1)
def _load() -> tuple[list[str], dict[str, int], str]:
  data = json.loads(_WORDLIST_JSON.read_text())
  words: list[str] = data["words"]
  index: dict[str, int] = {w: i for i, w in enumerate(words)}
  return words, index, data["version"]
'''
LOAD_NEW_TMPL = '''\
_WORDLIST: tuple[list[str], str] = (
{words_literal},
  {version_literal},
)


@lru_cache(maxsize=1)
def _load() -> tuple[list[str], dict[str, int], str]:
  words, version = _WORDLIST
  index: dict[str, int] = {{w: i for i, w in enumerate(words)}}
  return words, index, version
'''


def format_words(words: list[str]) -> str:
  """Format a list of words as a Python list literal, wrapped at COLS."""
  items = [repr(w) for w in words]
  joined = ", ".join(items)
  # Fits on one line?
  one_line = f"  [{joined}]"
  if len(one_line) <= COLS:
    return one_line
  # Wrap into multiple lines indented by 4 spaces inside the outer 2-space indent.
  lines = []
  current = ""
  for item in items:
    candidate = f"{current}, {item}" if current else item
    if len(candidate) + 4 > COLS - 2:  # account for "  [" prefix on first line
      if current:
        lines.append(current)
      current = item
    else:
      current = candidate
  if current:
    lines.append(current)
  inner = ",\n    ".join(lines)
  return f"  [\n    {inner},\n  ]"


def build(source: Path, wordlist: Path, output: Path) -> None:
  src = source.read_text()
  data = json.loads(wordlist.read_text())
  words: list[str] = data["words"]
  version: str = data["version"]

  # Remove imports that are only needed for file-based loading.
  src = src.replace(IMPORT_JSON, "")
  src = src.replace(IMPORT_PATH, "")
  src = src.replace(WORDLIST_LINE + "\n", "")  # remove trailing blank line too

  # Replace _load().
  words_lit = format_words(words)
  version_lit = repr(version)
  load_new = LOAD_NEW_TMPL.format(
    words_literal=words_lit,
    version_literal=version_lit,
  )
  if LOAD_OLD not in src:
    sys.exit(f"error: could not find _load() definition in {source}")
  src = src.replace(LOAD_OLD, load_new)

  output.write_text(src)
  print(f"wrote {output}  ({output.stat().st_size:,} bytes)")


if __name__ == "__main__":
  args = sys.argv[1:]
  source   = Path(args[0]) if len(args) > 0 else Path("mnemonicode.py")
  wordlist = Path(args[1]) if len(args) > 1 else Path("wordlist.json")
  output   = Path(args[2]) if len(args) > 2 else Path("mnemonicode_inline.py")

  for p in (source, wordlist):
    if not p.exists():
      sys.exit(f"error: {p} not found")

  build(source, wordlist, output)