feat: nvidia triton embedding integration

docs/docs/api_reference/embeddings/nvidia_triton.md

@@ -0,0 +1,4 @@
+::: llama_index.embeddings.nvidia_triton
+    options:
+      members:
+        - NvidiaTritonEmbedding

docs/mkdocs.yml

@@ -940,6 +940,7 @@ nav:
           - ./api_reference/embeddings/netmind.md
           - ./api_reference/embeddings/nomic.md
           - ./api_reference/embeddings/nvidia.md
+          - ./api_reference/embeddings/nvidia_triton.md
           - ./api_reference/embeddings/oci_data_science.md
           - ./api_reference/embeddings/oci_genai.md
           - ./api_reference/embeddings/octoai.md
@@ -2488,6 +2489,7 @@ plugins:
             - ../llama-index-integrations/protocols/llama-index-protocols-ag-ui
             - ../llama-index-integrations/vector_stores/llama-index-vector-stores-openGauss
             - ../llama-index-integrations/tools/llama-index-tools-hive
+            - ../llama-index-integrations/embeddings/llama-index-embeddings-nvidia-triton
             - ../llama-index-integrations/vector_stores/llama-index-vector-stores-db2
             - ../llama-index-integrations/voice_agents/llama-index-voice-agents-openai
   - redirects:

llama-index-integrations/embeddings/llama-index-embeddings-nvidia-triton/.gitignore

@@ -0,0 +1,153 @@
+llama_index/_static
+.DS_Store
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+bin/
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+etc/
+include/
+lib/
+lib64/
+parts/
+sdist/
+share/
+var/
+wheels/
+pip-wheel-metadata/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+.ruff_cache
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+notebooks/
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+.python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+pyvenv.cfg
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# Jetbrains
+.idea
+modules/
+*.swp
+
+# VsCode
+.vscode
+
+# pipenv
+Pipfile
+Pipfile.lock
+
+# pyright
+pyrightconfig.json

llama-index-integrations/embeddings/llama-index-embeddings-nvidia-triton/LICENSE

@@ -0,0 +1,21 @@
+The MIT License
+
+Copyright (c) Jerry Liu
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+THE SOFTWARE.

llama-index-integrations/embeddings/llama-index-embeddings-nvidia-triton/Makefile

@@ -0,0 +1,17 @@
+GIT_ROOT ?= $(shell git rev-parse --show-toplevel)
+
+help:	## Show all Makefile targets.
+	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | awk 'BEGIN {FS = ":.*?## "}; {printf "\033[33m%-30s\033[0m %s\n", $$1, $$2}'
+
+format:	## Run code autoformatters (black).
+	pre-commit install
+	git ls-files | xargs pre-commit run black --files
+
+lint:	## Run linters: pre-commit (black, ruff, codespell) and mypy
+	pre-commit install && git ls-files | xargs pre-commit run --show-diff-on-failure --files
+
+test:	## Run tests via pytest.
+	pytest tests
+
+watch-docs:	## Build and watch documentation.
+	sphinx-autobuild docs/ docs/_build/html --open-browser --watch $(GIT_ROOT)/llama_index/

llama-index-integrations/embeddings/llama-index-embeddings-nvidia-triton/README.md

@@ -0,0 +1,25 @@
+# LlamaIndex Embeddings Integration: Nvidia Triton
+
+This integration allows LlamaIndex to use embedding models hosted on a [Triton Inference Server](https://github.com/triton-inference-server/server).
+
+## Usage:
+
+```python
+from llama_index.embeddings.nvidia_triton import NvidiaTritonEmbedding
+
+embedding = NvidiaTritonEmbedding(
+    model_name="text_embeddings",
+    server_url="localhost:8000",
+    client_kwargs={"ssl": False},
+)
+
+print(embedding.get_text_embedding("hello world"))
+```
+
+Parameters:
+
+- `model_name`: the name of the embedding model.
+- `server_url`: the URL to the Triton Inference Server, normally on the HTTP port.
+- `client_kwargs`: additional arguments to be passed to the `tritonclient.http.InferenceServerClient` instance, such us timeouts, ssl, etc.
+- `input_tensor_name`: the name of the tensor the embedding model expects the input to be. Default: `INPUT_TEXT`.
+- `output_tensor_name`: the name of the tensor the embedding model will serve the output embedding to. Default: `OUTPUT_EMBEDDINGS`.

llama-index-integrations/embeddings/llama-index-embeddings-nvidia-triton/llama_index/embeddings/nvidia_triton/__init__.py

@@ -0,0 +1,3 @@
+from llama_index.embeddings.nvidia_triton.base import NvidiaTritonEmbedding
+
+__all__ = ["NvidiaTritonEmbedding"]

llama-index-integrations/embeddings/llama-index-embeddings-nvidia-triton/llama_index/embeddings/nvidia_triton/base.py

@@ -0,0 +1,132 @@
+from typing import Any, Dict, List, Optional
+
+from llama_index.core.base.embeddings.base import BaseEmbedding
+from llama_index.core.constants import DEFAULT_EMBED_BATCH_SIZE
+from llama_index.core.callbacks.base import CallbackManager
+from llama_index.core.bridge.pydantic import PrivateAttr
+
+import random
+import numpy as np
+from tritonclient.http import (
+    InferenceServerClient,
+    InferInput,
+    InferRequestedOutput,
+)
+
+DEFAULT_INPUT_TENSOR_NAME = "INPUT_TEXT"
+DEFAULT_OUTPUT_TENSOR_NAME = "OUTPUT_EMBEDDINGS"
+
+
+class NvidiaTritonEmbedding(BaseEmbedding):
+    """
+    Nvidia Triton Embedding.
+
+    This connector allows for llama_index to interact with embedding models hosted on a Triton
+    inference server over HTTP.
+
+    [Triton Inference Server Github](https://github.com/triton-inference-server/server)
+
+    Examples:
+        `pip install llama-index-embeddings-nvidia-triton`
+
+        ```python
+        from llama_index.embeddings.nvidia_triton import NvidiaTritonEmbedding
+
+        # Ensure a Triton server instance is running and provide the correct HTTP URL for your Triton server instance
+        triton_url = "localhost:8000"
+
+        # Instantiate the NvidiaTritonEmbedding class
+        emb_client = NvidiaTritonEmbedding(
+            server_url=triton_url,
+            model_name="text_embeddings",
+        )
+
+        # Get a text embedding
+        embedding = emb_client.get_text_embedding("hello world")
+        print(f"Embedding for 'hello world': {embedding}")
+        print(f"Embedding length: {len(embedding)}")
+        ```
+
+    """
+
+    _client: InferenceServerClient = PrivateAttr()
+    _input_tensor_name: str = PrivateAttr()
+    _output_tensor_name: str = PrivateAttr()
+
+    def __init__(
+        self,
+        model_name: str,
+        server_url: str = "localhost:8000",
+        embed_batch_size: int = DEFAULT_EMBED_BATCH_SIZE,
+        input_tensor_name: str = DEFAULT_INPUT_TENSOR_NAME,
+        output_tensor_name: str = DEFAULT_OUTPUT_TENSOR_NAME,
+        callback_manager: Optional[CallbackManager] = None,
+        client_kwargs: Optional[Dict[str, Any]] = None,
+        **kwargs: Any,
+    ) -> None:
+        super().__init__(
+            model_name=model_name,
+            embed_batch_size=embed_batch_size,
+            callback_manager=callback_manager,  # type: ignore
+            **kwargs,
+        )
+
+        self._client = InferenceServerClient(
+            url=server_url,
+            **client_kwargs or {},
+        )
+        self._input_tensor_name = input_tensor_name
+        self._output_tensor_name = output_tensor_name
+
+    @classmethod
+    def class_name(cls) -> str:
+        return "NvidiaTritonEmbedding"
+
+    def _get_query_embedding(self, query: str) -> List[float]:
+        """Get query embedding."""
+        return self.get_general_text_embeddings([query])[0]
+
+    async def _aget_query_embedding(self, query: str) -> List[float]:
+        """The asynchronous version of _get_query_embedding."""
+        embs = await self.aget_general_text_embeddings([query])
+        return embs[0]
+
+    def _get_text_embedding(self, text: str) -> List[float]:
+        """Get text embedding."""
+        return self.get_general_text_embeddings([text])[0]
+
+    async def _aget_text_embedding(self, text: str) -> List[float]:
+        """Asynchronously get text embedding."""
+        embs = await self.aget_general_text_embeddings([text])
+        return embs[0]
+
+    def _get_text_embeddings(self, texts: List[str]) -> List[List[float]]:
+        """Get text embeddings."""
+        return self.get_general_text_embeddings(texts)
+
+    async def _aget_text_embeddings(self, texts: List[str]) -> List[List[float]]:
+        """Asynchronously get text embeddings."""
+        return await self.aget_general_text_embeddings(texts)
+
+    def get_general_text_embeddings(self, texts: List[str]) -> List[List[float]]:
+        """Get Triton embedding."""
+        input_data = InferInput(self._input_tensor_name, [len(texts)], "BYTES")
+        input_data.set_data_from_numpy(np.array(texts, dtype=np.object_))
+        output_data = InferRequestedOutput(self._output_tensor_name)
+        request_id = str(random.randint(1, 9999999))  # nosec
+
+        response = self._client.infer(
+            model_name=self.model_name,
+            inputs=[input_data],
+            outputs=[output_data],
+            request_id=request_id,
+        )
+
+        embeddings = response.as_numpy(self._output_tensor_name)
+        if embeddings is None:
+            raise ValueError("No embeddings returned from Triton server.")
+        return [e.tolist() for e in embeddings]
+
+    async def aget_general_text_embeddings(self, texts: List[str]) -> List[List[float]]:
+        """Asynchronously get Triton embedding."""
+        return self.get_general_text_embeddings(texts)