agno 1.8.1__py3-none-any.whl → 2.0.0__py3-none-any.whl

agno/{embedder → knowledge/embedder}/jina.py

@@ -4,13 +4,18 @@ from typing import Any, Dict, List, Optional, Tuple
 
 from typing_extensions import Literal
 
-from agno.embedder.base import Embedder
+from agno.knowledge.embedder.base import Embedder
 from agno.utils.log import logger
 
 try:
     import requests
 except ImportError:
-    raise ImportError("requests not installed, use pip install requests")
+    raise ImportError("`requests` not installed, use pip install requests")
+
+try:
+    import aiohttp
+except ImportError:
+    raise ImportError("`aiohttp` not installed, use pip install aiohttp")
 
 
 @dataclass
@@ -71,3 +76,44 @@ class JinaEmbedder(Embedder):
         except Exception as e:
             logger.warning(f"Failed to get embedding and usage: {e}")
             return [], None
+
+    async def _async_response(self, text: str) -> Dict[str, Any]:
+        """Async version of _response using aiohttp."""
+        data = {
+            "model": self.id,
+            "late_chunking": self.late_chunking,
+            "dimensions": self.dimensions,
+            "embedding_type": self.embedding_type,
+            "input": [text],  # Jina API expects a list
+        }
+        if self.user is not None:
+            data["user"] = self.user
+        if self.request_params:
+            data.update(self.request_params)
+
+        timeout = aiohttp.ClientTimeout(total=self.timeout) if self.timeout else None
+
+        async with aiohttp.ClientSession(timeout=timeout) as session:
+            async with session.post(self.base_url, headers=self._get_headers(), json=data) as response:
+                response.raise_for_status()
+                return await response.json()
+
+    async def async_get_embedding(self, text: str) -> List[float]:
+        """Async version of get_embedding."""
+        try:
+            result = await self._async_response(text)
+            return result["data"][0]["embedding"]
+        except Exception as e:
+            logger.warning(f"Failed to get embedding: {e}")
+            return []
+
+    async def async_get_embedding_and_usage(self, text: str) -> Tuple[List[float], Optional[Dict]]:
+        """Async version of get_embedding_and_usage."""
+        try:
+            result = await self._async_response(text)
+            embedding = result["data"][0]["embedding"]
+            usage = result.get("usage")
+            return embedding, usage
+        except Exception as e:
+            logger.warning(f"Failed to get embedding and usage: {e}")
+            return [], None

agno/knowledge/embedder/langdb.py

@@ -0,0 +1,22 @@
+from dataclasses import dataclass
+from os import getenv
+from typing import Optional
+
+from agno.knowledge.embedder.openai import OpenAIEmbedder
+
+
+@dataclass
+class LangDBEmbedder(OpenAIEmbedder):
+    id: str = "text-embedding-ada-002"
+    dimensions: int = 1536
+    api_key: Optional[str] = getenv("LANGDB_API_KEY")
+    project_id: Optional[str] = getenv("LANGDB_PROJECT_ID")
+    base_url: Optional[str] = None
+
+    def __post_init__(self):
+        """Set the base_url based on project_id if not provided."""
+        if not self.project_id:
+            raise ValueError("LANGDB_PROJECT_ID not set in the environment")
+
+        if not self.base_url:
+            self.base_url = f"https://api.us-east-1.langdb.ai/{self.project_id}/v1"

agno/knowledge/embedder/mistral.py

@@ -0,0 +1,139 @@
+from dataclasses import dataclass
+from os import getenv
+from typing import Any, Dict, List, Optional, Tuple
+
+from agno.knowledge.embedder.base import Embedder
+from agno.utils.log import logger
+
+try:
+    from mistralai import Mistral  # type: ignore
+    from mistralai.models.embeddingresponse import EmbeddingResponse  # type: ignore
+except ImportError:
+    logger.error("`mistralai` not installed")
+    raise
+
+
+@dataclass
+class MistralEmbedder(Embedder):
+    id: str = "mistral-embed"
+    dimensions: int = 1024
+    # -*- Request parameters
+    request_params: Optional[Dict[str, Any]] = None
+    # -*- Client parameters
+    api_key: Optional[str] = getenv("MISTRAL_API_KEY")
+    endpoint: Optional[str] = None
+    max_retries: Optional[int] = None
+    timeout: Optional[int] = None
+    client_params: Optional[Dict[str, Any]] = None
+    # -*- Provide the Mistral Client manually
+    mistral_client: Optional[Mistral] = None
+
+    @property
+    def client(self) -> Mistral:
+        if self.mistral_client:
+            return self.mistral_client
+
+        _client_params: Dict[str, Any] = {
+            "api_key": self.api_key,
+            "endpoint": self.endpoint,
+            "max_retries": self.max_retries,
+            "timeout": self.timeout,
+        }
+        _client_params = {k: v for k, v in _client_params.items() if v is not None}
+
+        if self.client_params:
+            _client_params.update(self.client_params)
+
+        self.mistral_client = Mistral(**_client_params)
+
+        return self.mistral_client
+
+    def _response(self, text: str) -> EmbeddingResponse:
+        _request_params: Dict[str, Any] = {
+            "inputs": text,
+            "model": self.id,
+        }
+        if self.request_params:
+            _request_params.update(self.request_params)
+        response = self.client.embeddings.create(**_request_params)
+        if response is None:
+            raise ValueError("Failed to get embedding response")
+        return response
+
+    def get_embedding(self, text: str) -> List[float]:
+        try:
+            response: EmbeddingResponse = self._response(text=text)
+            if response.data and response.data[0].embedding:
+                return response.data[0].embedding
+            return []
+        except Exception as e:
+            logger.warning(f"Error getting embedding: {e}")
+            return []
+
+    def get_embedding_and_usage(self, text: str) -> Tuple[List[float], Dict[str, Any]]:
+        try:
+            response: EmbeddingResponse = self._response(text=text)
+            embedding: List[float] = (
+                response.data[0].embedding if (response.data and response.data[0].embedding) else []
+            )
+            usage: Dict[str, Any] = response.usage.model_dump() if response.usage else {}
+            return embedding, usage
+        except Exception as e:
+            logger.warning(f"Error getting embedding and usage: {e}")
+            return [], {}
+
+    async def async_get_embedding(self, text: str) -> List[float]:
+        """Async version of get_embedding."""
+        try:
+            # Check if the client has an async version of embeddings.create
+            if hasattr(self.client.embeddings, "create_async"):
+                response: EmbeddingResponse = await self.client.embeddings.create_async(
+                    inputs=text, model=self.id, **self.request_params if self.request_params else {}
+                )
+            else:
+                # Fallback to running sync method in thread executor
+                import asyncio
+
+                loop = asyncio.get_running_loop()
+                response: EmbeddingResponse = await loop.run_in_executor(  # type: ignore
+                    None,
+                    lambda: self.client.embeddings.create(
+                        inputs=text, model=self.id, **self.request_params if self.request_params else {}
+                    ),
+                )
+
+            if response.data and response.data[0].embedding:
+                return response.data[0].embedding
+            return []
+        except Exception as e:
+            logger.warning(f"Error getting embedding: {e}")
+            return []
+
+    async def async_get_embedding_and_usage(self, text: str) -> Tuple[List[float], Dict[str, Any]]:
+        """Async version of get_embedding_and_usage."""
+        try:
+            # Check if the client has an async version of embeddings.create
+            if hasattr(self.client.embeddings, "create_async"):
+                response: EmbeddingResponse = await self.client.embeddings.create_async(
+                    inputs=text, model=self.id, **self.request_params if self.request_params else {}
+                )
+            else:
+                # Fallback to running sync method in thread executor
+                import asyncio
+
+                loop = asyncio.get_running_loop()
+                response: EmbeddingResponse = await loop.run_in_executor(  # type: ignore
+                    None,
+                    lambda: self.client.embeddings.create(
+                        inputs=text, model=self.id, **self.request_params if self.request_params else {}
+                    ),
+                )
+
+            embedding: List[float] = (
+                response.data[0].embedding if (response.data and response.data[0].embedding) else []
+            )
+            usage: Dict[str, Any] = response.usage.model_dump() if response.usage else {}
+            return embedding, usage
+        except Exception as e:
+            logger.warning(f"Error getting embedding and usage: {e}")
+            return [], {}

agno/{embedder → knowledge/embedder}/nebius.py

@@ -2,7 +2,7 @@ from dataclasses import dataclass
 from os import getenv
 from typing import Optional
 
-from agno.embedder.openai import OpenAIEmbedder
+from agno.knowledge.embedder.openai import OpenAIEmbedder
 
 
 @dataclass

agno/{embedder → knowledge/embedder}/ollama.py

@@ -1,12 +1,13 @@
 from dataclasses import dataclass
 from typing import Any, Dict, List, Optional, Tuple
 
-from agno.embedder.base import Embedder
-from agno.utils.log import logger
+from agno.knowledge.embedder.base import Embedder
+from agno.utils.log import log_error, logger
 
 try:
     import importlib.metadata as metadata
 
+    from ollama import AsyncClient as AsyncOllamaClient
     from ollama import Client as OllamaClient
     from packaging import version
 
@@ -30,7 +31,7 @@ except ImportError as e:
 
 except Exception as e:
     # Catch-all for unexpected errors
-    print(f"An unexpected error occurred: {e}")
+    log_error(f"An unexpected error occurred: {e}")
 
 
 @dataclass
@@ -42,6 +43,7 @@ class OllamaEmbedder(Embedder):
     options: Optional[Any] = None
     client_kwargs: Optional[Dict[str, Any]] = None
     ollama_client: Optional[OllamaClient] = None
+    async_client: Optional[AsyncOllamaClient] = None
 
     @property
     def client(self) -> OllamaClient:
@@ -58,6 +60,21 @@ class OllamaEmbedder(Embedder):
         self.ollama_client = OllamaClient(**_ollama_params)
         return self.ollama_client
 
+    @property
+    def aclient(self) -> AsyncOllamaClient:
+        if self.async_client:
+            return self.async_client
+
+        _ollama_params: Dict[str, Any] = {
+            "host": self.host,
+            "timeout": self.timeout,
+        }
+        _ollama_params = {k: v for k, v in _ollama_params.items() if v is not None}
+        if self.client_kwargs:
+            _ollama_params.update(self.client_kwargs)
+        self.async_client = AsyncOllamaClient(**_ollama_params)
+        return self.async_client
+
     def _response(self, text: str) -> Dict[str, Any]:
         kwargs: Dict[str, Any] = {}
         if self.options is not None:
@@ -88,3 +105,37 @@ class OllamaEmbedder(Embedder):
         embedding = self.get_embedding(text=text)
         usage = None
         return embedding, usage
+
+    async def _async_response(self, text: str) -> Dict[str, Any]:
+        """Async version of _response using AsyncOllamaClient."""
+        kwargs: Dict[str, Any] = {}
+        if self.options is not None:
+            kwargs["options"] = self.options
+
+        response = await self.aclient.embed(input=text, model=self.id, **kwargs)
+        if response and "embeddings" in response:
+            embeddings = response["embeddings"]
+            if isinstance(embeddings, list) and len(embeddings) > 0 and isinstance(embeddings[0], list):
+                return {"embeddings": embeddings[0]}  # Use the first element
+            elif isinstance(embeddings, list) and all(isinstance(x, (int, float)) for x in embeddings):
+                return {"embeddings": embeddings}  # Return as-is if already flat
+        return {"embeddings": []}  # Return an empty list if no valid embedding is found
+
+    async def async_get_embedding(self, text: str) -> List[float]:
+        """Async version of get_embedding."""
+        try:
+            response = await self._async_response(text=text)
+            embedding = response.get("embeddings", [])
+            if len(embedding) != self.dimensions:
+                logger.warning(f"Expected embedding dimension {self.dimensions}, but got {len(embedding)}")
+                return []
+            return embedding
+        except Exception as e:
+            logger.warning(f"Error getting embedding: {e}")
+            return []
+
+    async def async_get_embedding_and_usage(self, text: str) -> Tuple[List[float], Optional[Dict]]:
+        """Async version of get_embedding_and_usage."""
+        embedding = await self.async_get_embedding(text=text)
+        usage = None
+        return embedding, usage

agno/knowledge/embedder/openai.py

@@ -0,0 +1,223 @@
+from dataclasses import dataclass
+from typing import Any, Dict, List, Optional, Tuple
+
+from typing_extensions import Literal
+
+from agno.knowledge.embedder.base import Embedder
+from agno.utils.log import logger
+
+try:
+    from openai import AsyncOpenAI
+    from openai import OpenAI as OpenAIClient
+    from openai.types.create_embedding_response import CreateEmbeddingResponse
+except ImportError:
+    raise ImportError("`openai` not installed")
+
+
+@dataclass
+class OpenAIEmbedder(Embedder):
+    id: str = "text-embedding-3-small"
+    dimensions: Optional[int] = None
+    encoding_format: Literal["float", "base64"] = "float"
+    user: Optional[str] = None
+    api_key: Optional[str] = None
+    organization: Optional[str] = None
+    base_url: Optional[str] = None
+    request_params: Optional[Dict[str, Any]] = None
+    client_params: Optional[Dict[str, Any]] = None
+    openai_client: Optional[OpenAIClient] = None
+    async_client: Optional[AsyncOpenAI] = None
+
+    def __post_init__(self):
+        if self.dimensions is None:
+            self.dimensions = 3072 if self.id == "text-embedding-3-large" else 1536
+
+    @property
+    def client(self) -> OpenAIClient:
+        if self.openai_client:
+            return self.openai_client
+
+        _client_params: Dict[str, Any] = {
+            "api_key": self.api_key,
+            "organization": self.organization,
+            "base_url": self.base_url,
+        }
+        _client_params = {k: v for k, v in _client_params.items() if v is not None}
+        if self.client_params:
+            _client_params.update(self.client_params)
+        self.openai_client = OpenAIClient(**_client_params)
+        return self.openai_client
+
+    @property
+    def aclient(self) -> AsyncOpenAI:
+        if self.async_client:
+            return self.async_client
+        params = {
+            "api_key": self.api_key,
+            "organization": self.organization,
+            "base_url": self.base_url,
+        }
+        filtered_params: Dict[str, Any] = {k: v for k, v in params.items() if v is not None}
+        if self.client_params:
+            filtered_params.update(self.client_params)
+        self.async_client = AsyncOpenAI(**filtered_params)
+        return self.async_client
+
+    def response(self, text: str) -> CreateEmbeddingResponse:
+        _request_params: Dict[str, Any] = {
+            "input": text,
+            "model": self.id,
+            "encoding_format": self.encoding_format,
+        }
+        if self.user is not None:
+            _request_params["user"] = self.user
+        if self.id.startswith("text-embedding-3"):
+            _request_params["dimensions"] = self.dimensions
+        if self.request_params:
+            _request_params.update(self.request_params)
+        return self.client.embeddings.create(**_request_params)
+
+    def get_embedding(self, text: str) -> List[float]:
+        response: CreateEmbeddingResponse = self.response(text=text)
+        try:
+            return response.data[0].embedding
+        except Exception as e:
+            logger.warning(e)
+            return []
+
+    def get_embedding_and_usage(self, text: str) -> Tuple[List[float], Optional[Dict]]:
+        response: CreateEmbeddingResponse = self.response(text=text)
+
+        embedding = response.data[0].embedding
+        usage = response.usage
+        if usage:
+            return embedding, usage.model_dump()
+        return embedding, None
+
+    async def async_get_embedding(self, text: str) -> List[float]:
+        req: Dict[str, Any] = {
+            "input": text,
+            "model": self.id,
+            "encoding_format": self.encoding_format,
+        }
+        if self.user is not None:
+            req["user"] = self.user
+        if self.id.startswith("text-embedding-3"):
+            req["dimensions"] = self.dimensions
+        if self.request_params:
+            req.update(self.request_params)
+
+        try:
+            response: CreateEmbeddingResponse = await self.aclient.embeddings.create(**req)
+            return response.data[0].embedding
+        except Exception as e:
+            logger.warning(e)
+            return []
+
+    async def async_get_embedding_and_usage(self, text: str):
+        req: Dict[str, Any] = {
+            "input": text,
+            "model": self.id,
+            "encoding_format": self.encoding_format,
+        }
+        if self.user is not None:
+            req["user"] = self.user
+        if self.id.startswith("text-embedding-3"):
+            req["dimensions"] = self.dimensions
+        if self.request_params:
+            req.update(self.request_params)
+
+        response = await self.aclient.embeddings.create(**req)
+        embedding = response.data[0].embedding
+        usage = response.usage
+        return embedding, usage.model_dump() if usage else None
+
+    def get_embeddings_batch(self, texts: List[str], batch_size: int = 100) -> List[List[float]]:
+        """
+        Get embeddings for multiple texts in batches.
+
+        Args:
+            texts: List of text strings to embed
+            batch_size: Number of texts to process in each API call (max ~2048)
+
+        Returns:
+            List of embedding vectors
+        """
+        all_embeddings = []
+
+        for i in range(0, len(texts), batch_size):
+            batch_texts = texts[i : i + batch_size]
+
+            req: Dict[str, Any] = {
+                "input": batch_texts,
+                "model": self.id,
+                "encoding_format": self.encoding_format,
+            }
+            if self.user is not None:
+                req["user"] = self.user
+            if self.id.startswith("text-embedding-3"):
+                req["dimensions"] = self.dimensions
+            if self.request_params:
+                req.update(self.request_params)
+
+            try:
+                response: CreateEmbeddingResponse = self.client.embeddings.create(**req)
+                batch_embeddings = [data.embedding for data in response.data]
+                all_embeddings.extend(batch_embeddings)
+            except Exception as e:
+                logger.warning(f"Error in batch embedding: {e}")
+                # Fallback to individual calls for this batch
+                for text in batch_texts:
+                    try:
+                        embedding = self.get_embedding(text)
+                        all_embeddings.append(embedding)
+                    except Exception as e2:
+                        logger.warning(f"Error in individual embedding fallback: {e2}")
+                        all_embeddings.append([])
+
+        return all_embeddings
+
+    async def async_get_embeddings_batch(self, texts: List[str], batch_size: int = 100) -> List[List[float]]:
+        """
+        Get embeddings for multiple texts in batches (async version).
+
+        Args:
+            texts: List of text strings to embed
+            batch_size: Number of texts to process in each API call (max ~2048)
+
+        Returns:
+            List of embedding vectors
+        """
+        all_embeddings = []
+
+        for i in range(0, len(texts), batch_size):
+            batch_texts = texts[i : i + batch_size]
+
+            req: Dict[str, Any] = {
+                "input": batch_texts,
+                "model": self.id,
+                "encoding_format": self.encoding_format,
+            }
+            if self.user is not None:
+                req["user"] = self.user
+            if self.id.startswith("text-embedding-3"):
+                req["dimensions"] = self.dimensions
+            if self.request_params:
+                req.update(self.request_params)
+
+            try:
+                response: CreateEmbeddingResponse = await self.aclient.embeddings.create(**req)
+                batch_embeddings = [data.embedding for data in response.data]
+                all_embeddings.extend(batch_embeddings)
+            except Exception as e:
+                logger.warning(f"Error in async batch embedding: {e}")
+                # Fallback to individual async calls for this batch
+                for text in batch_texts:
+                    try:
+                        embedding = await self.async_get_embedding(text)
+                        all_embeddings.append(embedding)
+                    except Exception as e2:
+                        logger.warning(f"Error in individual async embedding fallback: {e2}")
+                        all_embeddings.append([])
+
+        return all_embeddings

agno/{embedder → knowledge/embedder}/sentence_transformer.py

@@ -1,7 +1,7 @@
 from dataclasses import dataclass
 from typing import Dict, List, Optional, Tuple, Union
 
-from agno.embedder.base import Embedder
+from agno.knowledge.embedder.base import Embedder
 from agno.utils.log import logger
 
 try:
@@ -42,3 +42,18 @@ class SentenceTransformerEmbedder(Embedder):
 
     def get_embedding_and_usage(self, text: str) -> Tuple[List[float], Optional[Dict]]:
         return self.get_embedding(text=text), None
+
+    async def async_get_embedding(self, text: Union[str, List[str]]) -> List[float]:
+        """Async version using thread executor for CPU-bound operations."""
+        import asyncio
+
+        loop = asyncio.get_event_loop()
+        # Run the CPU-bound operation in a thread executor
+        return await loop.run_in_executor(None, self.get_embedding, text)
+
+    async def async_get_embedding_and_usage(self, text: str) -> Tuple[List[float], Optional[Dict]]:
+        """Async version using thread executor for CPU-bound operations."""
+        import asyncio
+
+        loop = asyncio.get_event_loop()
+        return await loop.run_in_executor(None, self.get_embedding_and_usage, text)

agno/{embedder → knowledge/embedder}/together.py

@@ -2,7 +2,7 @@ from dataclasses import dataclass
 from os import getenv
 from typing import Optional
 
-from agno.embedder.openai import OpenAIEmbedder
+from agno.knowledge.embedder.openai import OpenAIEmbedder
 
 
 @dataclass

agno/{embedder → knowledge/embedder}/voyageai.py

@@ -1,10 +1,11 @@
 from dataclasses import dataclass
 from typing import Any, Dict, List, Optional, Tuple
 
-from agno.embedder.base import Embedder
+from agno.knowledge.embedder.base import Embedder
 from agno.utils.log import logger
 
 try:
+    from voyageai import AsyncClient as AsyncVoyageClient
     from voyageai import Client as VoyageClient
     from voyageai.object import EmbeddingsObject
 except ImportError:
@@ -22,6 +23,7 @@ class VoyageAIEmbedder(Embedder):
     timeout: Optional[float] = None
     client_params: Optional[Dict[str, Any]] = None
     voyage_client: Optional[VoyageClient] = None
+    async_client: Optional[AsyncVoyageClient] = None
 
     @property
     def client(self) -> VoyageClient:
@@ -39,6 +41,22 @@ class VoyageAIEmbedder(Embedder):
         self.voyage_client = VoyageClient(**_client_params)
         return self.voyage_client
 
+    @property
+    def aclient(self) -> AsyncVoyageClient:
+        if self.async_client:
+            return self.async_client
+
+        _client_params = {
+            "api_key": self.api_key,
+            "max_retries": self.max_retries,
+            "timeout": self.timeout,
+        }
+        _client_params = {k: v for k, v in _client_params.items() if v is not None}
+        if self.client_params:
+            _client_params.update(self.client_params)
+        self.async_client = AsyncVoyageClient(**_client_params)
+        return self.async_client
+
     def _response(self, text: str) -> EmbeddingsObject:
         _request_params: Dict[str, Any] = {
             "texts": [text],
@@ -62,3 +80,33 @@ class VoyageAIEmbedder(Embedder):
         embedding = response.embeddings[0]
         usage = {"total_tokens": response.total_tokens}
         return embedding, usage
+
+    async def _async_response(self, text: str) -> EmbeddingsObject:
+        """Async version of _response using AsyncVoyageClient."""
+        _request_params: Dict[str, Any] = {
+            "texts": [text],
+            "model": self.id,
+        }
+        if self.request_params:
+            _request_params.update(self.request_params)
+        return await self.aclient.embed(**_request_params)
+
+    async def async_get_embedding(self, text: str) -> List[float]:
+        """Async version of get_embedding."""
+        try:
+            response: EmbeddingsObject = await self._async_response(text=text)
+            return response.embeddings[0]
+        except Exception as e:
+            logger.warning(f"Error getting embedding: {e}")
+            return []
+
+    async def async_get_embedding_and_usage(self, text: str) -> Tuple[List[float], Optional[Dict]]:
+        """Async version of get_embedding_and_usage."""
+        try:
+            response: EmbeddingsObject = await self._async_response(text=text)
+            embedding = response.embeddings[0]
+            usage = {"total_tokens": response.total_tokens}
+            return embedding, usage
+        except Exception as e:
+            logger.warning(f"Error getting embedding and usage: {e}")
+            return [], None