abstractvoice 0.5.2__py3-none-any.whl → 0.6.1__py3-none-any.whl

abstractvoice/artifacts.py

@@ -0,0 +1,173 @@
+from __future__ import annotations
+
+import hashlib
+import json
+import mimetypes
+import re
+from dataclasses import asdict, is_dataclass
+from datetime import datetime, timezone
+from pathlib import Path
+from typing import Any, Dict, Optional, Protocol, Union
+
+
+_ARTIFACT_ID_RE = re.compile(r"^[a-zA-Z0-9_-]{1,128}$")
+
+
+def _utc_now_iso() -> str:
+    return datetime.now(timezone.utc).isoformat()
+
+
+def sha256_hex(content: bytes) -> str:
+    return hashlib.sha256(content).hexdigest()
+
+
+def compute_artifact_id(content: bytes) -> str:
+    return sha256_hex(content)[:32]
+
+
+def is_artifact_ref(value: Any) -> bool:
+    return isinstance(value, dict) and isinstance(value.get("$artifact"), str) and bool(value.get("$artifact"))
+
+
+def get_artifact_id(ref: Dict[str, Any]) -> str:
+    return str(ref["$artifact"])
+
+
+def make_media_ref(
+    artifact_id: str,
+    *,
+    content_type: Optional[str] = None,
+    filename: Optional[str] = None,
+    sha256: Optional[str] = None,
+    size_bytes: Optional[int] = None,
+    metadata: Optional[Dict[str, Any]] = None,
+) -> Dict[str, Any]:
+    out: Dict[str, Any] = {"$artifact": str(artifact_id)}
+    if content_type:
+        out["content_type"] = str(content_type)
+    if filename:
+        out["filename"] = str(filename)
+    if sha256:
+        out["sha256"] = str(sha256)
+    if size_bytes is not None:
+        out["size_bytes"] = int(size_bytes)
+    if isinstance(metadata, dict) and metadata:
+        out["metadata"] = metadata
+    return out
+
+
+class MediaStore(Protocol):
+    def store_bytes(
+        self,
+        content: bytes,
+        *,
+        content_type: str,
+        filename: Optional[str] = None,
+        metadata: Optional[Dict[str, Any]] = None,
+        tags: Optional[Dict[str, str]] = None,
+        run_id: Optional[str] = None,
+        artifact_id: Optional[str] = None,
+    ) -> Dict[str, Any]: ...
+
+    def load_bytes(self, artifact_id: str) -> bytes: ...
+
+    def get_metadata(self, artifact_id: str) -> Optional[Dict[str, Any]]: ...
+
+
+class RuntimeArtifactStoreAdapter:
+    """Duck-typed adapter for AbstractRuntime's ArtifactStore (no hard dependency)."""
+
+    def __init__(self, artifact_store: Any):
+        self._store = artifact_store
+
+    def store_bytes(
+        self,
+        content: bytes,
+        *,
+        content_type: str,
+        filename: Optional[str] = None,
+        metadata: Optional[Dict[str, Any]] = None,
+        tags: Optional[Dict[str, str]] = None,
+        run_id: Optional[str] = None,
+        artifact_id: Optional[str] = None,
+    ) -> Dict[str, Any]:
+        store_fn = getattr(self._store, "store", None)
+        if not callable(store_fn):
+            raise TypeError("Provided artifact_store does not have a callable .store(...)")
+
+        content_b = bytes(content)
+        content_type = str(content_type or "application/octet-stream")
+        sha = sha256_hex(content_b)
+
+        merged_tags: Dict[str, str] = {}
+        if isinstance(tags, dict):
+            merged_tags.update({str(k): str(v) for k, v in tags.items()})
+        if filename and "filename" not in merged_tags:
+            merged_tags["filename"] = str(filename)
+        if sha and "sha256" not in merged_tags:
+            merged_tags["sha256"] = sha
+
+        try:
+            meta = store_fn(
+                content_b,
+                content_type=content_type,
+                run_id=str(run_id) if run_id else None,
+                tags=merged_tags or None,
+                artifact_id=str(artifact_id) if artifact_id else None,
+            )
+        except TypeError:
+            meta = store_fn(
+                content_b,
+                content_type=content_type,
+                run_id=str(run_id) if run_id else None,
+                tags=merged_tags or None,
+            )
+
+        artifact_id_out = None
+        if isinstance(meta, dict):
+            artifact_id_out = meta.get("artifact_id")
+        elif hasattr(meta, "artifact_id"):
+            artifact_id_out = getattr(meta, "artifact_id", None)
+        if not isinstance(artifact_id_out, str) or not artifact_id_out.strip():
+            raise TypeError("artifact_store.store(...) did not return a usable artifact_id")
+
+        return make_media_ref(
+            str(artifact_id_out),
+            content_type=content_type,
+            filename=str(filename) if filename else None,
+            sha256=sha,
+            size_bytes=len(content_b),
+            metadata=metadata if isinstance(metadata, dict) else None,
+        )
+
+    def load_bytes(self, artifact_id: str) -> bytes:
+        load_fn = getattr(self._store, "load", None)
+        if not callable(load_fn):
+            raise TypeError("Provided artifact_store does not have a callable .load(...)")
+        artifact = load_fn(str(artifact_id))
+        if artifact is None:
+            raise FileNotFoundError(f"Artifact not found: {artifact_id}")
+        if isinstance(artifact, (bytes, bytearray)):
+            return bytes(artifact)
+        if hasattr(artifact, "content"):
+            return bytes(getattr(artifact, "content"))
+        raise TypeError("artifact_store.load(...) returned an unsupported value")
+
+    def get_metadata(self, artifact_id: str) -> Optional[Dict[str, Any]]:
+        meta_fn = getattr(self._store, "get_metadata", None)
+        if not callable(meta_fn):
+            return None
+        meta = meta_fn(str(artifact_id))
+        if meta is None:
+            return None
+        if isinstance(meta, dict):
+            return meta
+        to_dict = getattr(meta, "to_dict", None)
+        if callable(to_dict):
+            out = to_dict()
+            return out if isinstance(out, dict) else None
+        if is_dataclass(meta):
+            out = asdict(meta)
+            return out if isinstance(out, dict) else None
+        return None
+

abstractvoice/audio/__init__.py

@@ -0,0 +1,7 @@
+"""Audio utilities (small, dependency-light)."""
+
+from .resample import linear_resample_mono
+from .recorder import record_wav
+
+__all__ = ["linear_resample_mono", "record_wav"]
+

abstractvoice/audio/recorder.py

@@ -0,0 +1,46 @@
+from __future__ import annotations
+
+import time
+from pathlib import Path
+
+import numpy as np
+import soundfile as sf
+
+
+def record_wav(
+    output_path: str | Path,
+    *,
+    seconds: float = 6.0,
+    sample_rate: int = 24000,
+    channels: int = 1,
+) -> str:
+    """Record microphone audio to a WAV file (best-effort).
+
+    This is intended for interactive REPL usage (e.g., `/clone-my-voice`).
+    """
+    try:
+        import sounddevice as sd
+    except Exception as e:
+        raise ImportError(
+            "Microphone recording requires sounddevice.\n"
+            "Install with: pip install abstractvoice\n"
+            f"Original error: {e}"
+        ) from e
+
+    out = Path(output_path)
+    out.parent.mkdir(parents=True, exist_ok=True)
+
+    frames = int(sample_rate * float(seconds))
+    if frames <= 0:
+        raise ValueError("seconds must be > 0")
+
+    audio = sd.rec(frames, samplerate=sample_rate, channels=channels, dtype="float32")
+    sd.wait()
+
+    # downmix to mono if needed
+    if channels > 1:
+        audio = np.mean(audio, axis=1, keepdims=True)
+
+    sf.write(str(out), audio, sample_rate, subtype="PCM_16")
+    return str(out)
+

abstractvoice/audio/resample.py

@@ -0,0 +1,25 @@
+from __future__ import annotations
+
+import numpy as np
+
+
+def linear_resample_mono(audio: np.ndarray, src_sr: int, dst_sr: int) -> np.ndarray:
+    """Lightweight mono resampler (linear interpolation).
+
+    Good enough for short prompts and avoids adding heavy DSP dependencies.
+    """
+    src_sr = int(src_sr)
+    dst_sr = int(dst_sr)
+    if src_sr <= 0 or dst_sr <= 0:
+        return audio
+    if src_sr == dst_sr:
+        return audio
+    if audio is None or len(audio) < 2:
+        return audio
+
+    ratio = float(dst_sr) / float(src_sr)
+    new_len = max(1, int(round(len(audio) * ratio)))
+    x_old = np.linspace(0.0, 1.0, num=len(audio), endpoint=True)
+    x_new = np.linspace(0.0, 1.0, num=new_len, endpoint=True)
+    return np.interp(x_new, x_old, audio).astype(np.float32)
+

abstractvoice/cloning/__init__.py

@@ -0,0 +1,7 @@
+"""Optional voice cloning support (behind `abstractvoice[cloning]`)."""
+
+from .store import VoiceCloneStore
+from .manager import VoiceCloner
+
+__all__ = ["VoiceCloneStore", "VoiceCloner"]
+