python-voiceio 0.3.0__tar.gz → 0.3.2__tar.gz

{python_voiceio-0.3.0/python_voiceio.egg-info → python_voiceio-0.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: python-voiceio
-Version: 0.3.0
+Version: 0.3.2
 Summary: Speak → text, locally, instantly.
 Author: Hugo Montenegro
 License-Expression: MIT
@@ -56,6 +56,7 @@ Dynamic: license-file
 [![PyPI](https://img.shields.io/pypi/v/python-voiceio)](https://pypi.org/project/python-voiceio/)
 [![Python](https://img.shields.io/pypi/pyversions/python-voiceio)](https://pypi.org/project/python-voiceio/)
 [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
+[![Downloads](https://img.shields.io/pepy/dt/python-voiceio)](https://pepy.tech/projects/python-voiceio)
 
 Speak → text, locally, instantly.
 
@@ -153,6 +154,10 @@ Press your hotkey to start recording (1s pre-buffer catches the first syllable).
 - **Works everywhere**: IBus input method for GUI apps, clipboard for terminals
 - **Wayland + X11**: evdev hotkeys work on both, no root required
 - **Pre-buffer**: never miss the first syllable
+- **Voice commands**: "new line", "comma", "scratch that", punctuation by name
+- **Autocorrect**: LLM-powered review of recurring Whisper mistakes (`voiceio correct`)
+- **Text-to-speech**: hear selected text spoken back (Piper, eSpeak, Edge TTS)
+- **Smart post-processing**: numbers ("twenty five" → "25"), punctuation, capitalization
 - **Auto-healing**: falls back to the next working backend if one fails
 - **Autostart**: optional systemd service, restarts on crash
 - **Self-diagnosing**: `voiceio doctor` checks everything, `--fix` repairs it
@@ -176,7 +181,10 @@ voiceio                  Start the daemon
 voiceio setup            Interactive setup wizard
 voiceio doctor           Health check (--fix to auto-repair)
 voiceio test             Test microphone + live transcription
+voiceio demo             Interactive guided tour of all features
 voiceio toggle           Toggle recording on a running daemon
+voiceio correct          Review and fix recurring transcription errors
+voiceio history          View transcription history
 voiceio update           Update to latest version
 voiceio service install  Autostart on login (systemd / Windows Startup)
 voiceio logs             View recent logs
@@ -250,9 +258,8 @@ Contributions welcome! See [CONTRIBUTING.md](CONTRIBUTING.md) and [open issues](
 - [ ] Multiple engine backends (whisper.cpp for Vulkan/AMD, VOSK for low-end hardware)
 - [ ] Echo cancellation (filter system audio for meeting use)
 - [ ] Wake word activation ("Hey voiceio")
-- [ ] Text-to-speech output (Piper/espeak-ng — completes the "io")
-
 **Done**
+- [x] Text-to-speech output (Piper/eSpeak/Edge TTS — completes the "io")
 - [x] LLM auto-audit dictionary (`voiceio correct --auto` — scan history with LLM, interactive correction)
 - [x] LLM post-processing via Ollama (grammar cleanup, spelling fixes on final pass)
 - [x] Corrections dictionary — auto-replace misheard words, "correct that" voice command

{python_voiceio-0.3.0 → python_voiceio-0.3.2}/README.md

@@ -4,6 +4,7 @@
 [![PyPI](https://img.shields.io/pypi/v/python-voiceio)](https://pypi.org/project/python-voiceio/)
 [![Python](https://img.shields.io/pypi/pyversions/python-voiceio)](https://pypi.org/project/python-voiceio/)
 [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
+[![Downloads](https://img.shields.io/pepy/dt/python-voiceio)](https://pepy.tech/projects/python-voiceio)
 
 Speak → text, locally, instantly.
 
@@ -101,6 +102,10 @@ Press your hotkey to start recording (1s pre-buffer catches the first syllable).
 - **Works everywhere**: IBus input method for GUI apps, clipboard for terminals
 - **Wayland + X11**: evdev hotkeys work on both, no root required
 - **Pre-buffer**: never miss the first syllable
+- **Voice commands**: "new line", "comma", "scratch that", punctuation by name
+- **Autocorrect**: LLM-powered review of recurring Whisper mistakes (`voiceio correct`)
+- **Text-to-speech**: hear selected text spoken back (Piper, eSpeak, Edge TTS)
+- **Smart post-processing**: numbers ("twenty five" → "25"), punctuation, capitalization
 - **Auto-healing**: falls back to the next working backend if one fails
 - **Autostart**: optional systemd service, restarts on crash
 - **Self-diagnosing**: `voiceio doctor` checks everything, `--fix` repairs it
@@ -124,7 +129,10 @@ voiceio                  Start the daemon
 voiceio setup            Interactive setup wizard
 voiceio doctor           Health check (--fix to auto-repair)
 voiceio test             Test microphone + live transcription
+voiceio demo             Interactive guided tour of all features
 voiceio toggle           Toggle recording on a running daemon
+voiceio correct          Review and fix recurring transcription errors
+voiceio history          View transcription history
 voiceio update           Update to latest version
 voiceio service install  Autostart on login (systemd / Windows Startup)
 voiceio logs             View recent logs
@@ -198,9 +206,8 @@ Contributions welcome! See [CONTRIBUTING.md](CONTRIBUTING.md) and [open issues](
 - [ ] Multiple engine backends (whisper.cpp for Vulkan/AMD, VOSK for low-end hardware)
 - [ ] Echo cancellation (filter system audio for meeting use)
 - [ ] Wake word activation ("Hey voiceio")
-- [ ] Text-to-speech output (Piper/espeak-ng — completes the "io")
-
 **Done**
+- [x] Text-to-speech output (Piper/eSpeak/Edge TTS — completes the "io")
 - [x] LLM auto-audit dictionary (`voiceio correct --auto` — scan history with LLM, interactive correction)
 - [x] LLM post-processing via Ollama (grammar cleanup, spelling fixes on final pass)
 - [x] Corrections dictionary — auto-replace misheard words, "correct that" voice command

{python_voiceio-0.3.0 → python_voiceio-0.3.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "python-voiceio"
-version = "0.3.0"
+version = "0.3.2"
 description = "Speak → text, locally, instantly."
 readme = "README.md"
 license = "MIT"

{python_voiceio-0.3.0 → python_voiceio-0.3.2/python_voiceio.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: python-voiceio
-Version: 0.3.0
+Version: 0.3.2
 Summary: Speak → text, locally, instantly.
 Author: Hugo Montenegro
 License-Expression: MIT
@@ -56,6 +56,7 @@ Dynamic: license-file
 [![PyPI](https://img.shields.io/pypi/v/python-voiceio)](https://pypi.org/project/python-voiceio/)
 [![Python](https://img.shields.io/pypi/pyversions/python-voiceio)](https://pypi.org/project/python-voiceio/)
 [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
+[![Downloads](https://img.shields.io/pepy/dt/python-voiceio)](https://pepy.tech/projects/python-voiceio)
 
 Speak → text, locally, instantly.
 
@@ -153,6 +154,10 @@ Press your hotkey to start recording (1s pre-buffer catches the first syllable).
 - **Works everywhere**: IBus input method for GUI apps, clipboard for terminals
 - **Wayland + X11**: evdev hotkeys work on both, no root required
 - **Pre-buffer**: never miss the first syllable
+- **Voice commands**: "new line", "comma", "scratch that", punctuation by name
+- **Autocorrect**: LLM-powered review of recurring Whisper mistakes (`voiceio correct`)
+- **Text-to-speech**: hear selected text spoken back (Piper, eSpeak, Edge TTS)
+- **Smart post-processing**: numbers ("twenty five" → "25"), punctuation, capitalization
 - **Auto-healing**: falls back to the next working backend if one fails
 - **Autostart**: optional systemd service, restarts on crash
 - **Self-diagnosing**: `voiceio doctor` checks everything, `--fix` repairs it
@@ -176,7 +181,10 @@ voiceio                  Start the daemon
 voiceio setup            Interactive setup wizard
 voiceio doctor           Health check (--fix to auto-repair)
 voiceio test             Test microphone + live transcription
+voiceio demo             Interactive guided tour of all features
 voiceio toggle           Toggle recording on a running daemon
+voiceio correct          Review and fix recurring transcription errors
+voiceio history          View transcription history
 voiceio update           Update to latest version
 voiceio service install  Autostart on login (systemd / Windows Startup)
 voiceio logs             View recent logs
@@ -250,9 +258,8 @@ Contributions welcome! See [CONTRIBUTING.md](CONTRIBUTING.md) and [open issues](
 - [ ] Multiple engine backends (whisper.cpp for Vulkan/AMD, VOSK for low-end hardware)
 - [ ] Echo cancellation (filter system audio for meeting use)
 - [ ] Wake word activation ("Hey voiceio")
-- [ ] Text-to-speech output (Piper/espeak-ng — completes the "io")
-
 **Done**
+- [x] Text-to-speech output (Piper/eSpeak/Edge TTS — completes the "io")
 - [x] LLM auto-audit dictionary (`voiceio correct --auto` — scan history with LLM, interactive correction)
 - [x] LLM post-processing via Ollama (grammar cleanup, spelling fixes on final pass)
 - [x] Corrections dictionary — auto-replace misheard words, "correct that" voice command

{python_voiceio-0.3.0 → python_voiceio-0.3.2}/tests/test_llm_api.py

@@ -6,7 +6,7 @@ import urllib.error
 from unittest.mock import MagicMock, patch
 
 from voiceio.config import AutocorrectConfig
-from voiceio.llm_api import chat, check_api_key, resolve_api_key
+from voiceio.llm_api import chat, check_api_key, detect_provider, resolve_api_key
 
 
 def _mock_response(data: dict) -> MagicMock:
@@ -116,3 +116,61 @@ def test_check_empty_key():
     cfg = _cfg(api_key="")
     with patch.dict("os.environ", {}, clear=True):
         assert check_api_key(cfg) is False
+
+
+# ── Anthropic native API ────────────────────────────────────────────────
+
+
+@patch("urllib.request.urlopen")
+def test_chat_anthropic_native(mock_urlopen):
+    mock_urlopen.return_value = _mock_response({
+        "content": [{"type": "text", "text": "Fixed text."}]
+    })
+    cfg = _cfg(base_url="https://api.anthropic.com/v1")
+    result = chat(cfg, "system prompt", "user message")
+    assert result == "Fixed text."
+
+    req = mock_urlopen.call_args[0][0]
+    assert req.get_header("X-api-key") == "test-key"
+    assert req.get_header("Anthropic-version") == "2023-06-01"
+    assert "Authorization" not in dict(req.header_items())
+    body = json.loads(req.data)
+    assert body["system"] == "system prompt"
+    assert body["messages"] == [{"role": "user", "content": "user message"}]
+    assert "/messages" in req.full_url
+
+
+@patch("urllib.request.urlopen")
+def test_check_api_key_anthropic(mock_urlopen):
+    mock_urlopen.return_value = _mock_response({
+        "content": [{"type": "text", "text": ""}]
+    })
+    cfg = _cfg(base_url="https://api.anthropic.com/v1")
+    assert check_api_key(cfg, "sk-ant-test") is True
+    req = mock_urlopen.call_args[0][0]
+    assert "/messages" in req.full_url
+
+
+# ── detect_provider ─────────────────────────────────────────────────────
+
+
+def test_detect_openrouter():
+    base_url, model = detect_provider("sk-or-abc123")
+    assert "openrouter" in base_url
+    assert "claude" in model
+
+
+def test_detect_anthropic():
+    base_url, model = detect_provider("sk-ant-abc123")
+    assert "anthropic.com" in base_url
+    assert "claude" in model
+
+
+def test_detect_openai():
+    base_url, model = detect_provider("sk-proj-abc123")
+    assert "openai.com" in base_url
+
+
+def test_detect_unknown_defaults_openrouter():
+    base_url, _ = detect_provider("unknown-key-format")
+    assert "openrouter" in base_url

{python_voiceio-0.3.0 → python_voiceio-0.3.2}/tests/test_tts.py

@@ -142,7 +142,7 @@ def test_player_empty_audio():
 
 def test_tts_config_defaults():
     cfg = TTSConfig()
-    assert cfg.enabled is False
+    assert cfg.enabled is True
     assert cfg.engine == "auto"
     assert cfg.hotkey == "ctrl+alt+s"
     assert cfg.voice == ""
@@ -155,4 +155,4 @@ def test_tts_config_in_main_config():
     cfg = Config()
     assert hasattr(cfg, "tts")
     assert isinstance(cfg.tts, TTSConfig)
-    assert cfg.tts.enabled is False
+    assert cfg.tts.enabled is True

python_voiceio-0.3.2/voiceio/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.3.2"

{python_voiceio-0.3.0 → python_voiceio-0.3.2}/voiceio/config.py

@@ -105,7 +105,7 @@ class AutocorrectConfig:
 
 @dataclass
 class TTSConfig:
-    enabled: bool = False
+    enabled: bool = True
     engine: str = "auto"         # "auto" | "piper" | "espeak" | "edge-tts"
     hotkey: str = "ctrl+alt+s"   # "s" for speak
     voice: str = ""              # empty = engine default

python_voiceio-0.3.2/voiceio/llm_api.py

@@ -0,0 +1,183 @@
+"""Multi-provider chat completions API client.
+
+Supports OpenRouter, OpenAI, Anthropic (native Messages API), Together, Groq,
+local Ollama (via /v1/chat/completions), etc. Zero dependencies beyond stdlib.
+"""
+from __future__ import annotations
+
+import json
+import logging
+import os
+import urllib.error
+import urllib.request
+
+from voiceio.config import AutocorrectConfig
+
+log = logging.getLogger(__name__)
+
+
+def _is_anthropic(base_url: str) -> bool:
+    """Check if the base URL points to Anthropic's native API."""
+    return "api.anthropic.com" in base_url
+
+
+def resolve_api_key(cfg: AutocorrectConfig) -> str:
+    """Resolve API key from config or environment variables."""
+    if cfg.api_key:
+        return cfg.api_key
+    # Check common env vars in priority order
+    for var in ("OPENROUTER_API_KEY", "OPENAI_API_KEY", "ANTHROPIC_API_KEY"):
+        val = os.environ.get(var, "")
+        if val:
+            return val
+    return ""
+
+
+def _anthropic_request(
+    base_url: str,
+    model: str,
+    system: str,
+    messages: list[dict],
+    api_key: str,
+    max_tokens: int,
+    timeout: float,
+) -> str | None:
+    """Send a request using Anthropic's native Messages API."""
+    url = f"{base_url}/messages"
+
+    body: dict = {
+        "model": model,
+        "max_tokens": max_tokens,
+        "messages": messages,
+    }
+    if system:
+        body["system"] = system
+
+    headers = {
+        "Content-Type": "application/json",
+        "x-api-key": api_key,
+        "anthropic-version": "2023-06-01",
+    }
+
+    req = urllib.request.Request(
+        url, data=json.dumps(body).encode(), headers=headers, method="POST",
+    )
+    with urllib.request.urlopen(req, timeout=timeout) as resp:
+        data = json.loads(resp.read())
+    # Anthropic returns content as a list of blocks
+    blocks = data.get("content", [])
+    text = "".join(b.get("text", "") for b in blocks if b.get("type") == "text")
+    return text.strip() or None
+
+
+def _openai_request(
+    base_url: str,
+    model: str,
+    system: str,
+    messages: list[dict],
+    api_key: str,
+    max_tokens: int,
+    timeout: float,
+) -> str | None:
+    """Send a request using the OpenAI chat completions format."""
+    url = f"{base_url}/chat/completions"
+
+    all_messages = []
+    if system:
+        all_messages.append({"role": "system", "content": system})
+    all_messages.extend(messages)
+
+    body = {
+        "model": model,
+        "max_tokens": max_tokens,
+        "messages": all_messages,
+    }
+
+    headers = {
+        "Content-Type": "application/json",
+        "Authorization": f"Bearer {api_key}",
+    }
+
+    req = urllib.request.Request(
+        url, data=json.dumps(body).encode(), headers=headers, method="POST",
+    )
+    with urllib.request.urlopen(req, timeout=timeout) as resp:
+        data = json.loads(resp.read())
+    return data["choices"][0]["message"]["content"].strip()
+
+
+def chat(
+    cfg: AutocorrectConfig,
+    system: str,
+    user_message: str,
+    *,
+    api_key: str = "",
+    max_tokens: int = 2048,
+) -> str | None:
+    """Send a chat completion request. Returns response text or None on failure.
+
+    Automatically detects Anthropic's native API vs OpenAI-compatible format
+    based on the configured base_url.
+    """
+    key = api_key or resolve_api_key(cfg)
+    if not key:
+        return None
+
+    base_url = cfg.base_url.rstrip("/")
+    messages = [{"role": "user", "content": user_message}]
+
+    try:
+        if _is_anthropic(base_url):
+            return _anthropic_request(
+                base_url, cfg.model, system, messages, key, max_tokens, cfg.timeout_secs,
+            )
+        return _openai_request(
+            base_url, cfg.model, system, messages, key, max_tokens, cfg.timeout_secs,
+        )
+    except urllib.error.HTTPError as e:
+        body_text = ""
+        try:
+            body_text = e.read().decode()[:200]
+        except Exception:
+            pass
+        log.warning("API request failed (HTTP %d): %s", e.code, body_text)
+        return None
+    except Exception as e:
+        log.warning("API request failed: %s", e)
+        return None
+
+
+def detect_provider(api_key: str) -> tuple[str, str]:
+    """Detect provider from API key prefix. Returns (base_url, model)."""
+    if api_key.startswith("sk-or-"):
+        return "https://openrouter.ai/api/v1", "anthropic/claude-sonnet-4"
+    if api_key.startswith("sk-ant-"):
+        return "https://api.anthropic.com/v1", "claude-sonnet-4-20250514"
+    if api_key.startswith(("sk-proj-", "sk-")):
+        return "https://api.openai.com/v1", "gpt-4o-mini"
+    # Default to OpenRouter (works with most keys)
+    return "https://openrouter.ai/api/v1", "anthropic/claude-sonnet-4"
+
+
+def check_api_key(cfg: AutocorrectConfig, api_key: str = "") -> bool:
+    """Validate an API key with a minimal request."""
+    key = api_key or resolve_api_key(cfg)
+    if not key:
+        return False
+
+    base_url = cfg.base_url.rstrip("/")
+    messages = [{"role": "user", "content": "hi"}]
+
+    try:
+        if _is_anthropic(base_url):
+            _anthropic_request(base_url, cfg.model, "", messages, key, 1, 10)
+        else:
+            _openai_request(base_url, cfg.model, "", messages, key, 1, 10)
+        return True
+    except urllib.error.HTTPError as e:
+        if e.code == 401:
+            return False
+        # Other errors (rate limit, etc.) mean the key itself is valid
+        return e.code != 403
+    except Exception:
+        return False

{python_voiceio-0.3.0 → python_voiceio-0.3.2}/voiceio/numbers.py

@@ -142,6 +142,7 @@ def convert_numbers(text: str, language: str = "en") -> str:
         # Collect consecutive number words
         if _is_number_word(low) and low != "a" and low != "and":
             num_words = []
+            last_category = None  # "ones", "tens", "scale"
             j = i
             while j < len(words):
                 w = words[j].lower().rstrip(".,;:?!")
@@ -153,6 +154,7 @@ def convert_numbers(text: str, language: str = "en") -> str:
                         # "a" at start: only if followed by scale word
                         if j + 1 < len(words) and words[j + 1].lower().rstrip(".,;:?!") in _SCALES:
                             num_words.append(w)
+                            last_category = "ones"
                             j += 1
                             continue
                         break
@@ -163,7 +165,14 @@ def convert_numbers(text: str, language: str = "en") -> str:
                             j += 1
                             continue
                         break
+                    # Two consecutive ones-words = separate numbers
+                    # e.g. "one two three" should NOT become 6
+                    # But "twenty three", "one hundred", "thirteen thousand" are valid
+                    cat = "scale" if w in _SCALES else ("tens" if w in _TENS else "ones")
+                    if cat == "ones" and last_category == "ones":
+                        break
                     num_words.append(w)
+                    last_category = cat
                     j += 1
                 else:
                     break

{python_voiceio-0.3.0 → python_voiceio-0.3.2}/voiceio/service.py

@@ -57,6 +57,7 @@ Type=simple
 ExecStart={bin_path}
 Restart=on-failure
 RestartSec=3
+PassEnvironment=DISPLAY WAYLAND_DISPLAY XDG_SESSION_TYPE XDG_RUNTIME_DIR
 
 [Install]
 WantedBy=default.target

{python_voiceio-0.3.0 → python_voiceio-0.3.2}/voiceio/tts/edge_engine.py

@@ -19,12 +19,25 @@ class EdgeEngine:
     def probe(self) -> ProbeResult:
         try:
             import edge_tts  # noqa: F401
-            return ProbeResult(ok=True)
         except ImportError:
             return ProbeResult(
                 ok=False, reason="edge-tts not installed",
                 fix_hint="pip install edge-tts",
             )
+        try:
+            import soundfile  # noqa: F401
+            return ProbeResult(ok=True)
+        except ImportError:
+            pass
+        try:
+            import pydub  # noqa: F401
+            return ProbeResult(ok=True)
+        except ImportError:
+            return ProbeResult(
+                ok=False,
+                reason="edge-tts needs soundfile or pydub to decode audio",
+                fix_hint="pip install soundfile",
+            )
 
     def synthesize(self, text: str, voice: str, speed: float) -> tuple[np.ndarray, int]:
         import asyncio

{python_voiceio-0.3.0 → python_voiceio-0.3.2}/voiceio/tts/piper_engine.py

@@ -22,10 +22,11 @@ class PiperEngine:
     def probe(self) -> ProbeResult:
         try:
             import piper  # noqa: F401
+            from piper.download import ensure_voice_exists, get_voices  # noqa: F401
             return ProbeResult(ok=True)
-        except ImportError:
+        except ImportError as e:
             return ProbeResult(
-                ok=False, reason="piper-tts not installed",
+                ok=False, reason=f"piper-tts not fully installed: {e}",
                 fix_hint="pip install piper-tts",
             )