thaloco/aman
1
Fork 0
Code Pull requests Releases Packages Activity Actions

Remove unused vocabulary and domain mode options

Browse source
This commit is contained in:
Thales Maciel 2026-02-25 11:26:23 -03:00
parent a6e75f9c16
commit 7af8750258
5 changed files with 34 additions and 93 deletions
Download patch file Download diff file Expand all files Collapse all files

39
README.md
View file

@ -5,7 +5,7 @@ Python X11 STT daemon that records audio, runs Whisper, applies local AI cleanup
## Requirements
- X11 (Wayland support scaffolded but not available yet)
- X11
- `sounddevice` (PortAudio)
- `faster-whisper`
- `llama-cpp-python`
@ -16,7 +16,7 @@ Python X11 STT daemon that records audio, runs Whisper, applies local AI cleanup
System packages (example names): `portaudio`/`libportaudio2`.
<details>
<summary>Ubuntu (X11)</summary>
<summary>Ubuntu/Debian</summary>
```bash
sudo apt install -y portaudio19-dev libportaudio2 python3-gi gir1.2-gtk-3.0 libayatana-appindicator3-1
@ -25,16 +25,7 @@ sudo apt install -y portaudio19-dev libportaudio2 python3-gi gir1.2-gtk-3.0 liba
</details>
<details>
<summary>Debian (X11)</summary>
```bash
sudo apt install -y portaudio19-dev libportaudio2 python3-gi gir1.2-gtk-3.0 libayatana-appindicator3-1
```
</details>
<details>
<summary>Arch Linux (X11)</summary>
<summary>Arch Linux</summary>
```bash
sudo pacman -S --needed portaudio gtk3 libayatana-appindicator
@ -43,7 +34,7 @@ sudo pacman -S --needed portaudio gtk3 libayatana-appindicator
</details>
<details>
<summary>Fedora (X11)</summary>
<summary>Fedora</summary>
```bash
sudo dnf install -y portaudio portaudio-devel gtk3 libayatana-appindicator-gtk3
@ -52,7 +43,7 @@ sudo dnf install -y portaudio portaudio-devel gtk3 libayatana-appindicator-gtk3
</details>
<details>
<summary>openSUSE (X11)</summary>
<summary>openSUSE</summary>
```bash
sudo zypper install -y portaudio portaudio-devel gtk3 libayatana-appindicator3-1
@ -70,18 +61,6 @@ X11 (supported):
uv sync --extra x11
```
Wayland (scaffold only):
```bash
uv sync --extra wayland
```
Run:
```bash
uv run python3 src/aman.py --config ~/.config/aman/config.json
```
## Config
Create `~/.config/aman/config.json`:
@ -100,11 +79,9 @@ Create `~/.config/aman/config.json`:
{ "from": "Martha", "to": "Marta" },
{ "from": "docker", "to": "Docker" }
],
"terms": ["Systemd", "Kubernetes"],
"max_rules": 500,
"max_terms": 500
"terms": ["Systemd", "Kubernetes"]
},
"domain_inference": { "enabled": true, "mode": "auto" }
"domain_inference": { "enabled": true }
}
```
@ -124,11 +101,9 @@ Vocabulary correction:
- `vocabulary.terms` is a preferred spelling list used as hinting context.
- Wildcards are intentionally rejected (`*`, `?`, `[`, `]`, `{`, `}`) to avoid ambiguous rules.
- Rules are deduplicated case-insensitively; conflicting replacements are rejected.
- Limits are bounded by `max_rules` and `max_terms`.
Domain inference:
- `domain_inference.mode` currently supports `auto`.
- Domain context is advisory only and is used to improve cleanup prompts.
- When confidence is low, it falls back to `general` context.

7
config.example.json
View file

@ -34,12 +34,9 @@
"systemd",
"Kubernetes",
"PostgreSQL"
],
"max_rules": 500,
"max_terms": 500
]
},
"domain_inference": {
"enabled": true,
"mode": "auto"
"enabled": true
}
}

47
src/config.py
View file

@ -12,10 +12,7 @@ DEFAULT_HOTKEY = "Cmd+m"
DEFAULT_STT_MODEL = "base"
DEFAULT_STT_DEVICE = "cpu"
DEFAULT_INJECTION_BACKEND = "clipboard"
DEFAULT_VOCAB_LIMIT = 500
DEFAULT_DOMAIN_INFERENCE_MODE = "auto"
ALLOWED_INJECTION_BACKENDS = {"clipboard", "injection"}
ALLOWED_DOMAIN_INFERENCE_MODES = {"auto"}
WILDCARD_CHARS = set("*?[]{}")
@ -51,14 +48,11 @@ class VocabularyReplacement:
class VocabularyConfig:
replacements: list[VocabularyReplacement] = field(default_factory=list)
terms: list[str] = field(default_factory=list)
max_rules: int = DEFAULT_VOCAB_LIMIT
max_terms: int = DEFAULT_VOCAB_LIMIT
@dataclass
class DomainInferenceConfig:
enabled: bool = True
mode: str = DEFAULT_DOMAIN_INFERENCE_MODE
@dataclass
@ -113,28 +107,11 @@ def validate(cfg: Config) -> None:
if not isinstance(cfg.injection.remove_transcription_from_clipboard, bool):
raise ValueError("injection.remove_transcription_from_clipboard must be boolean")
cfg.vocabulary.max_rules = _validated_limit(cfg.vocabulary.max_rules, "vocabulary.max_rules")
cfg.vocabulary.max_terms = _validated_limit(cfg.vocabulary.max_terms, "vocabulary.max_terms")
if len(cfg.vocabulary.replacements) > cfg.vocabulary.max_rules:
raise ValueError(
f"vocabulary.replacements cannot exceed vocabulary.max_rules ({cfg.vocabulary.max_rules})"
)
if len(cfg.vocabulary.terms) > cfg.vocabulary.max_terms:
raise ValueError(
f"vocabulary.terms cannot exceed vocabulary.max_terms ({cfg.vocabulary.max_terms})"
)
cfg.vocabulary.replacements = _validate_replacements(cfg.vocabulary.replacements)
cfg.vocabulary.terms = _validate_terms(cfg.vocabulary.terms)
if not isinstance(cfg.domain_inference.enabled, bool):
raise ValueError("domain_inference.enabled must be boolean")
mode = cfg.domain_inference.mode.strip().lower()
if mode not in ALLOWED_DOMAIN_INFERENCE_MODES:
allowed = ", ".join(sorted(ALLOWED_DOMAIN_INFERENCE_MODES))
raise ValueError(f"domain_inference.mode must be one of: {allowed}")
cfg.domain_inference.mode = mode
def _from_dict(data: dict[str, Any], cfg: Config) -> Config:
@ -186,17 +163,15 @@ def _from_dict(data: dict[str, Any], cfg: Config) -> Config:
if "terms" in vocabulary:
cfg.vocabulary.terms = _as_terms(vocabulary["terms"])
if "max_rules" in vocabulary:
cfg.vocabulary.max_rules = _as_int(vocabulary["max_rules"], "vocabulary.max_rules")
raise ValueError("vocabulary.max_rules is no longer supported")
if "max_terms" in vocabulary:
cfg.vocabulary.max_terms = _as_int(vocabulary["max_terms"], "vocabulary.max_terms")
raise ValueError("vocabulary.max_terms is no longer supported")
if "enabled" in domain_inference:
cfg.domain_inference.enabled = _as_bool(
domain_inference["enabled"], "domain_inference.enabled"
)
if "mode" in domain_inference:
cfg.domain_inference.mode = _as_nonempty_str(
domain_inference["mode"], "domain_inference.mode"
)
raise ValueError("domain_inference.mode is no longer supported")
return cfg
if "hotkey" in data:
@ -234,12 +209,6 @@ def _as_bool(value: Any, field_name: str) -> bool:
return value
def _as_int(value: Any, field_name: str) -> int:
if isinstance(value, bool) or not isinstance(value, int):
raise ValueError(f"{field_name} must be an integer")
return value
def _as_recording_input(value: Any) -> str | int | None:
if value is None:
return None
@ -276,16 +245,6 @@ def _as_terms(value: Any) -> list[str]:
return terms
def _validated_limit(value: int, field_name: str) -> int:
if isinstance(value, bool) or not isinstance(value, int):
raise ValueError(f"{field_name} must be an integer")
if value <= 0:
raise ValueError(f"{field_name} must be positive")
if value > 5000:
raise ValueError(f"{field_name} cannot exceed 5000")
return value
def _validate_replacements(value: list[VocabularyReplacement]) -> list[VocabularyReplacement]:
deduped: list[VocabularyReplacement] = []
seen: dict[str, str] = {}

32
tests/test_config.py
View file

@ -28,10 +28,7 @@ class ConfigTests(unittest.TestCase):
self.assertFalse(cfg.injection.remove_transcription_from_clipboard)
self.assertEqual(cfg.vocabulary.replacements, [])
self.assertEqual(cfg.vocabulary.terms, [])
self.assertEqual(cfg.vocabulary.max_rules, 500)
self.assertEqual(cfg.vocabulary.max_terms, 500)
self.assertTrue(cfg.domain_inference.enabled)
self.assertEqual(cfg.domain_inference.mode, "auto")
def test_loads_nested_config(self):
payload = {
@ -48,10 +45,8 @@ class ConfigTests(unittest.TestCase):
{"from": "docker", "to": "Docker"},
],
"terms": ["Systemd", "Kubernetes"],
"max_rules": 100,
"max_terms": 200,
},
"domain_inference": {"enabled": True, "mode": "auto"},
"domain_inference": {"enabled": True},
}
with tempfile.TemporaryDirectory() as td:
path = Path(td) / "config.json"
@ -65,14 +60,11 @@ class ConfigTests(unittest.TestCase):
self.assertEqual(cfg.stt.device, "cuda")
self.assertEqual(cfg.injection.backend, "injection")
self.assertTrue(cfg.injection.remove_transcription_from_clipboard)
self.assertEqual(cfg.vocabulary.max_rules, 100)
self.assertEqual(cfg.vocabulary.max_terms, 200)
self.assertEqual(len(cfg.vocabulary.replacements), 2)
self.assertEqual(cfg.vocabulary.replacements[0].source, "Martha")
self.assertEqual(cfg.vocabulary.replacements[0].target, "Marta")
self.assertEqual(cfg.vocabulary.terms, ["Systemd", "Kubernetes"])
self.assertTrue(cfg.domain_inference.enabled)
self.assertEqual(cfg.domain_inference.mode, "auto")
def test_loads_legacy_keys(self):
payload = {
@ -200,13 +192,31 @@ class ConfigTests(unittest.TestCase):
with self.assertRaisesRegex(ValueError, "wildcard"):
load(str(path))
def test_invalid_domain_mode_raises(self):
def test_removed_domain_mode_raises(self):
payload = {"domain_inference": {"mode": "heuristic"}}
with tempfile.TemporaryDirectory() as td:
path = Path(td) / "config.json"
path.write_text(json.dumps(payload), encoding="utf-8")
with self.assertRaisesRegex(ValueError, "domain_inference.mode"):
with self.assertRaisesRegex(ValueError, "domain_inference.mode is no longer supported"):
load(str(path))
def test_removed_vocabulary_max_rules_raises(self):
payload = {"vocabulary": {"max_rules": 100}}
with tempfile.TemporaryDirectory() as td:
path = Path(td) / "config.json"
path.write_text(json.dumps(payload), encoding="utf-8")
with self.assertRaisesRegex(ValueError, "vocabulary.max_rules is no longer supported"):
load(str(path))
def test_removed_vocabulary_max_terms_raises(self):
payload = {"vocabulary": {"max_terms": 100}}
with tempfile.TemporaryDirectory() as td:
path = Path(td) / "config.json"
path.write_text(json.dumps(payload), encoding="utf-8")
with self.assertRaisesRegex(ValueError, "vocabulary.max_terms is no longer supported"):
load(str(path))

2
tests/test_vocabulary.py
View file

@ -17,7 +17,7 @@ class VocabularyEngineTests(unittest.TestCase):
replacements=replacements or [],
terms=terms or [],
)
domain = DomainInferenceConfig(enabled=domain_enabled, mode="auto")
domain = DomainInferenceConfig(enabled=domain_enabled)
return VocabularyEngine(vocab, domain)
def test_boundary_aware_replacement(self):