fix(commands): keep line breaks from standalone voice commands (#958)
* fix(commands): keep line breaks from standalone voice commands
Two stacked defects ate the newline "new line" / "new paragraph"
produce when spoken alone:
* _normalize_segment_text stripped all leading whitespace, so the
command's leading newline never reached the injector (a clean
"new line" segment was dropped entirely).
* whisper.cpp transcripts arrive punctuated ("New line."), which
missed the exact-match ladder and left a stray "." after the
substituted break, injected as ". ".
Normalize now strips only spaces/tabs, and the generic path
substitutes on a punctuation-stripped core when the whole segment
is command phrases, so engine edge punctuation cannot leak while
real text keeps it.
Fixes #955
* fix(injection): deliver newline-only segments to the focused app
inject_text treated whitespace-only text as empty and returned early,
so the "\n" a standalone "new line" now correctly produces was
discarded before any backend ran ("Empty text provided, skipping
injection"). Only a truly empty segment is skipped; whitespace-only
input is typed/pasted verbatim (xdotool types "\n" as Return).
Space/tab-only segments never reach this point: segment normalization
strips them upstream.
Fixes #955
* fix(injection): keep skipping space/tab-only segments
The previous commit let every whitespace-only string through, which
also delivered bare spaces and tabs a caller never wants typed.
Narrow the skip to text with no content outside space/tab so "\n"
still reaches the backend while " " stays a no-op (the pre-existing
test_empty_text_injection contract).
Fixes #955
* fix(injection): deliver newline segments over IBus too
IBusTextInjector.inject_text had the same whitespace guard and dropped
"\n" before it reached the engine socket; narrow it the same way so
a newline-only segment is sent while space/tab-only stays a no-op.
Also let capitalize_sentences uppercase the first letter after leading
whitespace, so Vosk + auto_capitalize keeps working on "\n hello".
Covers Greptile's findings: real-socket test for the IBus path, the
capitalize regression via the main() callback seam, and type hints on
the new injector tests.
Fixes #955
* test(injection): add type hints to new IBus socket test
AGENTS.md requires hints on every signature; the Greptile review
flagged the new test for missing them.
---------
Co-authored-by: Jatin K Malik <jatinkrmalik@gmail.com>
This commit is contained in:
devin-ai-integration[bot]
co-authored by
Jatin K Malik
Jatin K Malik
parent
64032fb999
commit
0ddb127437
@@ -22,6 +22,11 @@ _ATTACH_LEFT_REPLACEMENTS = {".", ",", "?", "!", ";", ":"}
|
||||
# line-break command just inserted (``\s*`` would collapse "new line" + "period").
|
||||
_ATTACH_LEFT_WS = r"[^\S\n]*"
|
||||
|
||||
# Sentence-final punctuation engines may wrap around a standalone command
|
||||
# utterance ("New line."). Dropped only when the whole segment is command
|
||||
# phrases — the same marks stay untouched on real text.
|
||||
_EDGE_PUNCTUATION = ".,!?;:…"
|
||||
|
||||
# Base English text commands (always present).
|
||||
_BASE_TEXT_COMMANDS = {
|
||||
# Line commands
|
||||
@@ -252,22 +257,36 @@ class CommandProcessor:
|
||||
|
||||
# If no exact match found, fallback to generic processing
|
||||
else:
|
||||
processed_text = text.strip()
|
||||
# Engines add sentence-final punctuation to an utterance that is
|
||||
# only a command ("New line.", "Select all."). When the segment is
|
||||
# nothing but command phrases and whitespace once edge punctuation
|
||||
# is dropped, substitute on that stripped core so the stray mark
|
||||
# cannot leak into the output; real text keeps its punctuation.
|
||||
core = text.strip().strip(_EDGE_PUNCTUATION).strip()
|
||||
residual = self.format_cmd_regex.sub(
|
||||
"", self.action_cmd_regex.sub("", self.text_cmd_regex.sub("", core))
|
||||
)
|
||||
segment_text = core if core and not residual.strip() else text
|
||||
processed_text = segment_text.strip()
|
||||
|
||||
# Handle action commands
|
||||
for cmd, action in self.action_commands.items():
|
||||
cmd_pattern = r"\b" + re.escape(cmd) + r"\b"
|
||||
|
||||
if re.search(cmd_pattern, text, re.IGNORECASE):
|
||||
if re.search(cmd_pattern, segment_text, re.IGNORECASE):
|
||||
actions.append(action)
|
||||
|
||||
# Check if there's text after the command
|
||||
match = re.search(r"\b" + re.escape(cmd) + r"\s+(.*)", text, re.IGNORECASE)
|
||||
match = re.search(
|
||||
r"\b" + re.escape(cmd) + r"\s+(.*)", segment_text, re.IGNORECASE
|
||||
)
|
||||
if match:
|
||||
remaining_text = match.group(1).strip()
|
||||
# Only add space if there's text before the command
|
||||
cmd_match = re.search(
|
||||
r"^(.*?)\b" + re.escape(cmd) + r"\b", text, re.IGNORECASE
|
||||
r"^(.*?)\b" + re.escape(cmd) + r"\b",
|
||||
segment_text,
|
||||
re.IGNORECASE,
|
||||
)
|
||||
if cmd_match and cmd_match.group(1).strip():
|
||||
processed_text = " " + remaining_text
|
||||
@@ -302,9 +321,11 @@ class CommandProcessor:
|
||||
for cmd, format_type in self.format_commands.items():
|
||||
cmd_pattern = r"\b" + re.escape(cmd) + r"\b"
|
||||
|
||||
if re.search(cmd_pattern, text, re.IGNORECASE):
|
||||
if re.search(cmd_pattern, segment_text, re.IGNORECASE):
|
||||
# Handle format command that modifies next word
|
||||
match = re.search(r"\b" + re.escape(cmd) + r"\s+(\w+)", text, re.IGNORECASE)
|
||||
match = re.search(
|
||||
r"\b" + re.escape(cmd) + r"\s+(\w+)", segment_text, re.IGNORECASE
|
||||
)
|
||||
if match:
|
||||
word = match.group(1)
|
||||
if format_type == "capitalize_next":
|
||||
@@ -320,7 +341,7 @@ class CommandProcessor:
|
||||
processed_text = re.sub(
|
||||
r"\b" + re.escape(cmd) + r"\s+" + re.escape(word) + r"\b",
|
||||
replacement,
|
||||
text,
|
||||
segment_text,
|
||||
flags=re.IGNORECASE,
|
||||
)
|
||||
else:
|
||||
@@ -347,8 +368,14 @@ def capitalize_sentences(text: str) -> str:
|
||||
if not text:
|
||||
return text
|
||||
|
||||
# Capitalize first character
|
||||
result = text[0].upper() + text[1:]
|
||||
# Capitalize the first letter, skipping leading whitespace a voice
|
||||
# command can leave behind ("new line hello" -> "\n hello").
|
||||
first_letter = re.match(r"\s*([a-z])", text)
|
||||
if first_letter:
|
||||
idx = first_letter.start(1)
|
||||
result = text[:idx] + text[idx].upper() + text[idx + 1 :]
|
||||
else:
|
||||
result = text[0].upper() + text[1:]
|
||||
|
||||
# Capitalize after sentence-ending punctuation + whitespace
|
||||
result = re.sub(
|
||||
|
||||
Reference in New Issue
Block a user