sie-langchain 0.1.7__tar.gz

sie_langchain-0.1.7/.gitignore

@@ -0,0 +1,256 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[codz]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+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/
+cover/
+
+# 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
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .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
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+#poetry.toml
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#   pdm recommends including project-wide configuration in pdm.toml, but excluding .pdm-python.
+#   https://pdm-project.org/en/latest/usage/project/#working-with-version-control
+#pdm.lock
+#pdm.toml
+.pdm-python
+.pdm-build/
+
+# pixi
+#   Similar to Pipfile.lock, it is generally recommended to include pixi.lock in version control.
+#pixi.lock
+#   Pixi creates a virtual environment in the .pixi directory, just like venv module creates one
+#   in the .venv directory. It is recommended not to include this directory in version control.
+.pixi
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/
+
+# Abstra
+# Abstra is an AI-powered process automation framework.
+# Ignore directories containing user credentials, local state, and settings.
+# Learn more at https://abstra.io/docs
+.abstra/
+
+# Visual Studio Code
+#  Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore
+#  that can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
+#  and can be added to the global gitignore or merged into this file. However, if you prefer,
+#  you could uncomment the following to ignore the entire vscode folder
+# .vscode/
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Cursor
+#  Cursor is an AI-powered code editor. `.cursorignore` specifies files/directories to
+#  exclude from AI features like autocomplete and code analysis. Recommended for sensitive data
+#  refer to https://docs.cursor.com/context/ignore-files
+.cursorignore
+.cursorindexingignore
+
+# vcscode
+.vscode
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/
+
+# SIE specific
+# Model weights cache
+.cache/
+*.safetensors
+*.bin
+
+# Secrets (never commit)
+*.pem
+*.key
+credentials.json
+*-key.json
+
+# Terraform
+# Local .terraform directories (cached providers/modules)
+**/.terraform/*
+.terraform.lock*
+# State files (contain sensitive data)
+*.tfstate
+*.tfstate.*
+# Crash log files
+crash.log
+crash.*.log
+# Override files (local developer overrides)
+override.tf
+override.tf.json
+*_override.tf
+*_override.tf.json
+# tfvars files may contain secrets
+*.tfvars
+*.tfvars.json
+# Keep .terraform.lock.hcl for reproducibility (provider versions)
+
+# Node.js
+node_modules/
+
+# OS
+.DS_Store
+Thumbs.db
+
+# VIM
+*.swp
+
+# Worktree metadata
+.base-branch
+
+# Temporary files
+tmp/
+.tmp/
+.local/
+
+.requirements-modal.txt

sie_langchain-0.1.7/PKG-INFO

@@ -0,0 +1,12 @@
+Metadata-Version: 2.4
+Name: sie-langchain
+Version: 0.1.7
+Summary: SIE integration for LangChain
+Requires-Python: >=3.10
+Requires-Dist: langchain-core>=0.2.0
+Requires-Dist: sie-sdk>=0.1.0
+Provides-Extra: dev
+Requires-Dist: chromadb>=0.4; extra == 'dev'
+Requires-Dist: langchain-chroma>=0.1; extra == 'dev'
+Requires-Dist: pytest-asyncio>=0.23; extra == 'dev'
+Requires-Dist: pytest>=8.0; extra == 'dev'

sie_langchain-0.1.7/pyproject.toml

@@ -0,0 +1,25 @@
+[project]
+name = "sie-langchain"
+version = "0.1.7"
+description = "SIE integration for LangChain"
+requires-python = ">=3.10"
+dependencies = [
+    "sie-sdk>=0.1.0",
+    "langchain-core>=0.2.0",
+]
+
+[project.optional-dependencies]
+dev = [
+    "pytest>=8.0",
+    "pytest-asyncio>=0.23",
+    # For integration tests
+    "chromadb>=0.4",
+    "langchain-chroma>=0.1",
+]
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/sie_langchain"]

sie_langchain-0.1.7/src/sie_langchain/__init__.py

@@ -0,0 +1,36 @@
+"""SIE integration for LangChain.
+
+Provides LangChain-compatible wrappers for SIE's encoding, reranking,
+and entity extraction capabilities.
+
+Example:
+    >>> from sie_langchain import SIEEmbeddings, SIEReranker, SIEExtractor
+    >>>
+    >>> # Create embeddings
+    >>> embeddings = SIEEmbeddings(base_url="http://localhost:8080", model="BAAI/bge-m3")
+    >>> vectors = embeddings.embed_documents(["Hello world"])
+    >>>
+    >>> # Create reranker
+    >>> reranker = SIEReranker(base_url="http://localhost:8080")
+    >>> reranked = reranker.compress_documents(documents, query)
+    >>>
+    >>> # Create extractor tool
+    >>> extractor = SIEExtractor(base_url="http://localhost:8080")
+    >>> entities = extractor.invoke("John Smith works at Acme Corp")
+
+Hybrid search example:
+    >>> from langchain_pinecone import PineconeHybridSearchRetriever
+    >>> from sie_langchain import SIEEmbeddings, SIESparseEncoder
+    >>>
+    >>> retriever = PineconeHybridSearchRetriever(
+    ...     embeddings=SIEEmbeddings(model="BAAI/bge-m3"),
+    ...     sparse_encoder=SIESparseEncoder(model="BAAI/bge-m3"),
+    ...     index=pinecone_index,
+    ... )
+"""
+
+from sie_langchain.embeddings import SIEEmbeddings, SIESparseEncoder
+from sie_langchain.extractors import SIEExtractor
+from sie_langchain.rerankers import SIEReranker
+
+__all__ = ["SIEEmbeddings", "SIEExtractor", "SIEReranker", "SIESparseEncoder"]

sie_langchain-0.1.7/src/sie_langchain/embeddings.py

@@ -0,0 +1,328 @@
+"""SIE embeddings integration for LangChain.
+
+Provides drop-in replacement for OpenAI embeddings using SIE's inference server:
+- SIEEmbeddings: Dense embeddings for vector stores
+- SIESparseEncoder: Sparse encoder for hybrid search with PineconeHybridSearchRetriever
+"""
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+from langchain_core.embeddings import Embeddings
+
+if TYPE_CHECKING:
+    from sie_sdk import SIEAsyncClient, SIEClient
+
+
+class SIEEmbeddings(Embeddings):
+    """LangChain Embeddings implementation using SIE.
+
+    Wraps SIEClient.encode() to implement the Embeddings interface.
+
+    Example:
+        >>> # Basic usage
+        >>> embeddings = SIEEmbeddings(base_url="http://localhost:8080", model="BAAI/bge-m3")
+        >>> vectors = embeddings.embed_documents(["Hello world"])
+
+        >>> # With GPU routing for multi-GPU clusters
+        >>> embeddings = SIEEmbeddings(base_url="https://cluster.example.com", model="BAAI/bge-m3", gpu="a100-80gb")
+
+    Args:
+        base_url: URL of the SIE server.
+        model: Model name/ID to use for encoding.
+        client: Optional pre-configured SIEClient instance.
+        async_client: Optional pre-configured SIEAsyncClient instance.
+        instruction: Optional instruction prefix for embedding (model-dependent).
+        output_dtype: Output dtype: "float32" (default), "float16", "int8", "binary".
+        options: Runtime options dict passed to the model adapter. Available options
+            depend on the model - see model documentation for details.
+        gpu: Target GPU type for routing (e.g., "l4", "a100-80gb").
+        timeout_s: Request timeout in seconds.
+    """
+
+    def __init__(
+        self,
+        *,
+        base_url: str = "http://localhost:8080",
+        model: str = "BAAI/bge-m3",
+        client: SIEClient | None = None,
+        async_client: SIEAsyncClient | None = None,
+        instruction: str | None = None,
+        output_dtype: str | None = None,
+        options: dict[str, object] | None = None,
+        gpu: str | None = None,
+        timeout_s: float = 180.0,
+    ) -> None:
+        """Initialize SIE embeddings."""
+        self._base_url = base_url
+        self._model = model
+        self._instruction = instruction
+        self._output_dtype = output_dtype
+        self._options = options
+        self._gpu = gpu
+        self._timeout_s = timeout_s
+
+        # Store provided clients or create lazily
+        self._client = client
+        self._async_client = async_client
+
+    @property
+    def client(self) -> SIEClient:
+        """Get or create the sync SIEClient."""
+        if self._client is None:
+            from sie_sdk import SIEClient
+
+            self._client = SIEClient(
+                self._base_url,
+                timeout_s=self._timeout_s,
+                gpu=self._gpu,
+                options=self._options,
+            )
+        return self._client
+
+    @property
+    def async_client(self) -> SIEAsyncClient:
+        """Get or create the async SIEClient."""
+        if self._async_client is None:
+            from sie_sdk import SIEAsyncClient
+
+            self._async_client = SIEAsyncClient(
+                self._base_url,
+                timeout_s=self._timeout_s,
+                gpu=self._gpu,
+                options=self._options,
+            )
+        return self._async_client
+
+    def embed_documents(self, texts: list[str]) -> list[list[float]]:
+        """Embed a list of documents.
+
+        Args:
+            texts: List of document texts to embed.
+
+        Returns:
+            List of embedding vectors (as lists of floats).
+        """
+        if not texts:
+            return []
+
+        from sie_sdk.types import Item
+
+        items = [Item(text=text) for text in texts]
+        results = self.client.encode(
+            self._model,
+            items,
+            output_types=["dense"],
+            instruction=self._instruction,
+            output_dtype=self._output_dtype,
+        )
+
+        return [self._extract_dense(result) for result in results]
+
+    def embed_query(self, text: str) -> list[float]:
+        """Embed a single query text.
+
+        Args:
+            text: Query text to embed.
+
+        Returns:
+            Embedding vector as list of floats.
+        """
+        from sie_sdk.types import Item
+
+        result = self.client.encode(
+            self._model,
+            Item(text=text),
+            output_types=["dense"],
+            instruction=self._instruction,
+            output_dtype=self._output_dtype,
+            options={"is_query": True},
+        )
+
+        return self._extract_dense(result)
+
+    async def aembed_documents(self, texts: list[str]) -> list[list[float]]:
+        """Async embed a list of documents.
+
+        Args:
+            texts: List of document texts to embed.
+
+        Returns:
+            List of embedding vectors (as lists of floats).
+        """
+        if not texts:
+            return []
+
+        from sie_sdk.types import Item
+
+        items = [Item(text=text) for text in texts]
+        results = await self.async_client.encode(
+            self._model,
+            items,
+            output_types=["dense"],
+            instruction=self._instruction,
+            output_dtype=self._output_dtype,
+        )
+
+        return [self._extract_dense(result) for result in results]
+
+    async def aembed_query(self, text: str) -> list[float]:
+        """Async embed a single query text.
+
+        Args:
+            text: Query text to embed.
+
+        Returns:
+            Embedding vector as list of floats.
+        """
+        from sie_sdk.types import Item
+
+        result = await self.async_client.encode(
+            self._model,
+            Item(text=text),
+            output_types=["dense"],
+            instruction=self._instruction,
+            output_dtype=self._output_dtype,
+            options={"is_query": True},
+        )
+
+        return self._extract_dense(result)
+
+    def _extract_dense(self, result: object) -> list[float]:
+        """Extract dense embedding from encode result.
+
+        Args:
+            result: EncodeResult from SIE client (dict with "dense" key containing numpy array).
+
+        Returns:
+            Dense embedding as list of floats.
+        """
+        # SDK returns {"dense": np.ndarray, ...}
+        dense = result.get("dense") if isinstance(result, dict) else getattr(result, "dense", None)
+        if dense is None:
+            msg = "Encode result missing dense embedding"
+            raise ValueError(msg)
+        # Convert numpy array to list
+        return dense.tolist() if hasattr(dense, "tolist") else list(dense)
+
+
+class SIESparseEncoder:
+    """Sparse encoder for LangChain hybrid search.
+
+    Compatible with PineconeHybridSearchRetriever's sparse_encoder interface.
+    Provides encode_queries() and encode_documents() methods.
+
+    Example:
+        >>> from langchain_pinecone import PineconeHybridSearchRetriever
+        >>> from sie_langchain import SIEEmbeddings, SIESparseEncoder
+        >>>
+        >>> retriever = PineconeHybridSearchRetriever(
+        ...     embeddings=SIEEmbeddings(model="BAAI/bge-m3"),
+        ...     sparse_encoder=SIESparseEncoder(model="BAAI/bge-m3"),
+        ...     index=pinecone_index,
+        ... )
+
+    Args:
+        base_url: URL of the SIE server.
+        model: Model name/ID to use for encoding. Must support sparse output.
+        gpu: Target GPU type for routing (e.g., "l4", "a100-80gb").
+        timeout_s: Request timeout in seconds.
+    """
+
+    def __init__(
+        self,
+        *,
+        base_url: str = "http://localhost:8080",
+        model: str = "BAAI/bge-m3",
+        gpu: str | None = None,
+        timeout_s: float = 180.0,
+    ) -> None:
+        """Initialize SIE sparse encoder."""
+        self._base_url = base_url
+        self._model = model
+        self._gpu = gpu
+        self._timeout_s = timeout_s
+        self._client: SIEClient | None = None
+
+    @property
+    def client(self) -> SIEClient:
+        """Get or create the sync SIEClient."""
+        if self._client is None:
+            from sie_sdk import SIEClient
+
+            self._client = SIEClient(
+                self._base_url,
+                timeout_s=self._timeout_s,
+                gpu=self._gpu,
+            )
+        return self._client
+
+    def encode_queries(self, texts: list[str]) -> list[dict[str, list]]:
+        """Encode query texts to sparse vectors.
+
+        Args:
+            texts: List of query texts to encode.
+
+        Returns:
+            List of dicts with "indices" and "values" keys.
+        """
+        if not texts:
+            return []
+
+        from sie_sdk.types import Item
+
+        items = [Item(text=text) for text in texts]
+        results = self.client.encode(
+            self._model,
+            items,
+            output_types=["sparse"],
+            options={"is_query": True},
+        )
+
+        return [self._extract_sparse(result) for result in results]
+
+    def encode_documents(self, texts: list[str]) -> list[dict[str, list]]:
+        """Encode document texts to sparse vectors.
+
+        Args:
+            texts: List of document texts to encode.
+
+        Returns:
+            List of dicts with "indices" and "values" keys.
+        """
+        if not texts:
+            return []
+
+        from sie_sdk.types import Item
+
+        items = [Item(text=text) for text in texts]
+        results = self.client.encode(
+            self._model,
+            items,
+            output_types=["sparse"],
+        )
+
+        return [self._extract_sparse(result) for result in results]
+
+    def _extract_sparse(self, result: object) -> dict[str, list]:
+        """Extract sparse embedding from encode result.
+
+        Args:
+            result: EncodeResult from SIE client with "sparse" key.
+
+        Returns:
+            Dict with "indices" and "values" lists.
+        """
+        # SDK returns {"sparse": {"indices": np.ndarray, "values": np.ndarray}, ...}
+        sparse = result.get("sparse") if isinstance(result, dict) else getattr(result, "sparse", None)
+        if sparse is None:
+            return {"indices": [], "values": []}
+
+        indices = sparse.get("indices") if isinstance(sparse, dict) else getattr(sparse, "indices", None)
+        values = sparse.get("values") if isinstance(sparse, dict) else getattr(sparse, "values", None)
+
+        return {
+            "indices": indices.tolist() if hasattr(indices, "tolist") else list(indices or []),
+            "values": values.tolist() if hasattr(values, "tolist") else list(values or []),
+        }