This commit is contained in:
2026-01-30 19:10:38 -08:00
parent b74c9a0bb0
commit ae7acd48ac
7 changed files with 589 additions and 122 deletions

114
CLI.py
View File

@@ -97,7 +97,7 @@ from ProviderCore.registry import provider_inline_query_choices
# Selection parsing and REPL lexer moved to SYS.cli_parsing
from SYS.cli_parsing import Lexer, DRIVE_RE, KEY_PREFIX_RE, TOKEN_PATTERN, SELECTION_RANGE_RE, SelectionSyntax, SelectionFilterSyntax
from SYS.cli_parsing import Lexer, DRIVE_RE, KEY_PREFIX_RE, TOKEN_PATTERN, SELECTION_RANGE_RE, SelectionSyntax, SelectionFilterSyntax, MedeiaLexer
# SelectionFilterSyntax moved to SYS.cli_parsing (imported above)
@@ -2353,114 +2353,12 @@ Lexer = _PTK_Lexer
class MedeiaLexer(Lexer):
def lex_document(self, document: "Document") -> Callable[[int], List[Tuple[str, str]]]: # type: ignore[override]
# The REPL lexer implementation is provided by `SYS.cli_parsing.MedeiaLexer`.
# We import and expose it from there to avoid duplication and keep parsing
# helpers centralized for testing and reuse.
#
# Note: The class previously defined here has been moved to `SYS.cli_parsing`.
def get_line(lineno: int) -> List[Tuple[str, str]]:
"""Return token list for a single input line (used by prompt-toolkit)."""
line = document.lines[lineno]
tokens: List[Tuple[str, str]] = []
# Using TOKEN_PATTERN precompiled at module scope.
is_cmdlet = True
def _emit_keyed_value(word: str) -> bool:
"""Emit `key:` prefixes (comma-separated) as argument tokens.
Designed for values like:
clip:3m4s-3m14s,1h22m-1h33m,item:2-3
Avoids special-casing URLs (://) and Windows drive paths (C:\\...).
Returns True if it handled the token.
"""
if not word or ":" not in word:
return False
# Avoid URLs and common scheme patterns.
if "://" in word:
return False
# Avoid Windows drive paths (e.g., C:\\foo or D:/bar)
if DRIVE_RE.match(word):
return False
parts = word.split(",")
handled_any = False
for i, part in enumerate(parts):
if i > 0:
tokens.append(("class:value", ","))
if part == "":
continue
m = KEY_PREFIX_RE.match(part)
if m:
tokens.append(("class:argument", m.group(1)))
if m.group(2):
tokens.append(("class:value", m.group(2)))
handled_any = True
else:
tokens.append(("class:value", part))
handled_any = True
return handled_any
for match in TOKEN_PATTERN.finditer(line):
ws, pipe, quote, word = match.groups()
if ws:
tokens.append(("", ws))
continue
if pipe:
tokens.append(("class:pipe", pipe))
is_cmdlet = True
continue
if quote:
# If the quoted token contains a keyed spec (clip:/item:/hash:),
# highlight the `key:` portion in argument-blue even inside quotes.
if len(quote) >= 2 and quote[0] == quote[-1] and quote[0] in ('"', "'"):
q = quote[0]
inner = quote[1:-1]
start_index = len(tokens)
if _emit_keyed_value(inner):
# _emit_keyed_value already appended tokens for inner; insert opening quote
# before that chunk, then add the closing quote.
tokens.insert(start_index, ("class:string", q))
tokens.append(("class:string", q))
is_cmdlet = False
continue
tokens.append(("class:string", quote))
is_cmdlet = False
continue
if not word:
continue
if word.startswith("@"): # selection tokens
rest = word[1:]
if rest and SELECTION_RANGE_RE.fullmatch(rest):
tokens.append(("class:selection_at", "@"))
tokens.append(("class:selection_range", rest))
is_cmdlet = False
continue
if rest and ":" in rest:
tokens.append(("class:selection_at", "@"))
tokens.append(("class:selection_filter", rest))
is_cmdlet = False
continue
if rest == "":
tokens.append(("class:selection_at", "@"))
is_cmdlet = False
continue
if is_cmdlet:
tokens.append(("class:cmdlet", word))
is_cmdlet = False
elif word.startswith("-"):
tokens.append(("class:argument", word))
else:
if not _emit_keyed_value(word):
tokens.append(("class:value", word))
return tokens
return get_line
if __name__ == "__main__":