vision-agents-plugins-wizper 0.2.3__tar.gz

vision_agents_plugins_wizper-0.2.3/.gitignore

@@ -0,0 +1,90 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+*.so
+.cursor/*
+# Distribution / packaging
+.Python
+build/
+dist/
+downloads/
+develop-eggs/
+eggs/
+.eggs/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+pip-wheel-metadata/
+MANIFEST
+*.egg-info/
+*.egg
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+coverage.xml
+nosetests.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+
+# Type checker / lint caches
+.mypy_cache/
+.dmypy.json
+dmypy.json
+.pytype/
+.pyre/
+.ruff_cache/
+
+# Environments
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+.env
+.env.local
+.env.*.local
+.env.bak
+pyvenv.cfg
+.python-version
+
+# Editors / IDEs
+.vscode/
+.idea/
+
+# Jupyter Notebook
+.ipynb_checkpoints/
+
+# OS / Misc
+.DS_Store
+*.log
+
+# Tooling & repo-specific
+pyrightconfig.json
+shell.nix
+bin/*
+lib/*
+stream-py/
+
+# Artifacts / assets
+*.pt
+*.kef
+*.onnx
+profile.html
+
+/opencode.json

vision_agents_plugins_wizper-0.2.3/PKG-INFO

@@ -0,0 +1,15 @@
+Metadata-Version: 2.4
+Name: vision-agents-plugins-wizper
+Version: 0.2.3
+Summary: Wizper plugin for Vision Agents
+Project-URL: Documentation, https://visionagents.ai/
+Project-URL: Website, https://visionagents.ai/
+Project-URL: Source, https://github.com/GetStream/Vision-Agents
+License-Expression: MIT
+Keywords: AI,agents,fal,turn detection,voice agents
+Requires-Python: >=3.10
+Requires-Dist: fal-client>=0.7.0
+Requires-Dist: vision-agents
+Description-Content-Type: text/markdown
+
+Wizper for vision agents

vision_agents_plugins_wizper-0.2.3/README.md

@@ -0,0 +1 @@
+Wizper for vision agents

vision_agents_plugins_wizper-0.2.3/pyproject.toml

@@ -0,0 +1,41 @@
+[build-system]
+requires = ["hatchling", "hatch-vcs"]
+build-backend = "hatchling.build"
+
+[project]
+name = "vision-agents-plugins-wizper"
+dynamic = ["version"]
+description = "Wizper plugin for Vision Agents"
+readme = "README.md"
+keywords = ["fal", "turn detection", "AI", "voice agents", "agents"]
+requires-python = ">=3.10"
+license = "MIT"
+dependencies = [
+    "vision-agents",
+    "fal-client>=0.7.0",
+]
+
+[project.urls]
+Documentation = "https://visionagents.ai/"
+Website = "https://visionagents.ai/"
+Source = "https://github.com/GetStream/Vision-Agents"
+
+[tool.hatch.version]
+source = "vcs"
+raw-options = { root = "..", search_parent_directories = true, fallback_version = "0.0.0" }
+
+[tool.hatch.build.targets.wheel]
+packages = ["."]
+
+[tool.hatch.build.targets.sdist]
+include = ["/vision_agents"]
+
+[tool.uv.sources]
+vision-agents = { workspace = true }
+
+[dependency-groups]
+dev = [
+    "pytest>=8.4.1",
+    "pytest-asyncio>=1.0.0",
+    "numpy>=2.2.6,<2.3",
+]

vision_agents_plugins_wizper-0.2.3/vision_agents/plugins/wizper/__init__.py

@@ -0,0 +1,5 @@
+from .stt import STT
+
+# Re-export under the new namespace for convenience
+
+__all__ = ["STT"]

vision_agents_plugins_wizper-0.2.3/vision_agents/plugins/wizper/stt.py

@@ -0,0 +1,148 @@
+"""
+Fal Wizper STT Plugin for Stream
+
+Provides real-time audio transcription and translation using fal-ai/wizper (Whisper v3).
+This plugin integrates with Stream's audio processing pipeline to provide high-quality
+speech-to-text capabilities.
+"""
+
+import aiofiles
+import asyncio
+import logging
+import os
+import tempfile
+from pathlib import Path
+from typing import TYPE_CHECKING, Optional
+
+import fal_client
+from getstream.video.rtc.track_util import PcmData
+
+from vision_agents.core import stt
+from vision_agents.core.stt import TranscriptResponse
+
+if TYPE_CHECKING:
+    from vision_agents.core.edge.types import Participant
+
+logger = logging.getLogger(__name__)
+
+
+class STT(stt.STT):
+    """
+    Audio transcription and translation using fal-ai/wizper (Whisper v3).
+
+    This plugin provides real-time speech-to-text capabilities using the fal-ai/wizper
+    service, which is based on OpenAI's Whisper v3 model. It supports both transcription
+    and translation tasks.
+
+    Attributes:
+        task: The task type - either "transcribe" or "translate"
+        target_language: Target language code for translation (e.g., "pt" for Portuguese)
+    """
+
+    def __init__(
+        self,
+        task: str = "transcribe",
+        target_language: Optional[str] = None,
+        client: Optional[fal_client.AsyncClient] = None,
+    ):
+        """
+        Initialize Wizper STT.
+
+        Args:
+            task: "transcribe" or "translate"
+            target_language: Target language code (e.g., "pt" for Portuguese)
+            client: Optional fal_client.AsyncClient instance for testing
+        """
+        super().__init__(provider_name="wizper")
+        self.task = task
+        self.sample_rate = 48000
+        self.target_language = target_language
+        self._fal_client = client if client is not None else fal_client.AsyncClient()
+
+    async def process_audio(
+        self,
+        pcm_data: PcmData,
+        participant: Optional["Participant"] = None,
+    ):
+        """
+        Process audio through fal-ai/wizper for transcription.
+
+        Args:
+            pcm_data: The PCM audio data to process
+            participant: Optional participant metadata
+        """
+        if self.closed:
+            logger.warning("Wizper STT is closed, ignoring audio")
+            return
+
+        if pcm_data.samples.size == 0:
+            logger.debug("No audio data to process")
+            return
+
+        try:
+            logger.debug(
+                "Sending speech audio to fal-ai/wizper",
+                extra={"audio_bytes": pcm_data.samples.nbytes},
+            )
+            # Convert PCM to WAV format for upload using shared PcmData method
+            wav_data = pcm_data.to_wav_bytes()
+
+            # Create temporary file for upload (async to avoid blocking)
+            temp_file_path = os.path.join(
+                tempfile.gettempdir(), f"wizper_{os.getpid()}_{id(pcm_data)}.wav"
+            )
+            async with aiofiles.open(temp_file_path, "wb") as f:
+                await f.write(wav_data)
+
+            try:
+                input_params = {
+                    "task": "transcribe",  # TODO: make this dynamic, currently there's a bug in the fal-ai/wizper service where it only works with "transcribe"
+                    "chunk_level": "segment",
+                    "version": "3",
+                }
+                # Add language for translation
+                if self.target_language is not None:
+                    input_params["language"] = self.target_language
+
+                # Upload file and get URL
+                audio_url = await self._fal_client.upload_file(Path(temp_file_path))
+                input_params["audio_url"] = audio_url
+
+                # Use regular subscribe since streaming isn't supported
+                result = await self._fal_client.subscribe(
+                    "fal-ai/wizper", arguments=input_params
+                )
+                if "text" in result:
+                    text = result["text"].strip()
+                    if text:
+                        # Create a default participant if none provided
+                        if participant is None:
+                            from vision_agents.core.edge.types import Participant
+
+                            participant = Participant(
+                                original=None, user_id="test-user"
+                            )
+
+                        response_metadata = TranscriptResponse()
+                        self._emit_transcript_event(
+                            text, participant, response_metadata
+                        )
+            finally:
+                # Clean up temporary file (async to avoid blocking)
+                try:
+                    await asyncio.to_thread(os.unlink, temp_file_path)
+                except OSError:
+                    pass
+
+        except Exception as e:
+            logger.error(f"Wizper processing error: {str(e)}")
+            self._emit_error_event(e, "Wizper processing")
+
+    async def close(self):
+        """Close the Wizper STT service and release any resources."""
+        if self.closed:
+            logger.debug("Wizper STT service already closed")
+            return
+
+        logger.info("Closing Wizper STT service")
+        await super().close()