reminix-llamaindex 0.0.1__py3-none-any.whl

reminix_llamaindex/__init__.py

@@ -0,0 +1,3 @@
+from .adapter import LlamaIndexAdapter, wrap
+
+__all__ = ["LlamaIndexAdapter", "wrap"]

reminix_llamaindex/adapter.py

@@ -0,0 +1,179 @@
+"""LlamaIndex adapter for Reminix Runtime."""
+
+import json
+from collections.abc import AsyncIterator
+from typing import Any, Protocol, runtime_checkable
+
+from reminix_runtime import (
+    BaseAdapter,
+    ChatRequest,
+    ChatResponse,
+    InvokeRequest,
+    InvokeResponse,
+    Message,
+)
+
+
+@runtime_checkable
+class ChatEngine(Protocol):
+    """Protocol for LlamaIndex chat engines."""
+
+    async def achat(self, message: str) -> Any:
+        """Async chat method."""
+        ...
+
+    async def astream_chat(self, message: str) -> Any:
+        """Async streaming chat method."""
+        ...
+
+
+class LlamaIndexAdapter(BaseAdapter):
+    """Adapter for LlamaIndex chat engines."""
+
+    adapter_name = "llamaindex"
+
+    def __init__(self, engine: ChatEngine, name: str = "llamaindex-agent") -> None:
+        """Initialize the adapter.
+
+        Args:
+            engine: A LlamaIndex chat engine (e.g., SimpleChatEngine, ContextChatEngine).
+            name: Name for the agent.
+        """
+        self._engine = engine
+        self._name = name
+
+    @property
+    def name(self) -> str:
+        return self._name
+
+    def _get_last_user_message(self, messages: list[Message]) -> str:
+        """Get the last user message from the conversation."""
+        for message in reversed(messages):
+            if message.role == "user":
+                return message.content or ""
+        # Fallback to last message if no user message found
+        return messages[-1].content or "" if messages else ""
+
+    async def invoke(self, request: InvokeRequest) -> InvokeResponse:
+        """Handle an invoke request.
+
+        For task-oriented operations. Expects input with 'query' or 'prompt' key.
+
+        Args:
+            request: The invoke request with input data.
+
+        Returns:
+            The invoke response with the output.
+        """
+        # Extract query from input
+        if "query" in request.input:
+            query = request.input["query"]
+        elif "prompt" in request.input:
+            query = request.input["prompt"]
+        elif "message" in request.input:
+            query = request.input["message"]
+        else:
+            query = str(request.input)
+
+        # Call the chat engine
+        response = await self._engine.achat(query)
+
+        # Extract content from response
+        output = str(response.response) if hasattr(response, "response") else str(response)
+
+        return InvokeResponse(output=output)
+
+    async def chat(self, request: ChatRequest) -> ChatResponse:
+        """Handle a chat request.
+
+        For conversational interactions. Sends the last user message to the engine.
+
+        Args:
+            request: The chat request with messages.
+
+        Returns:
+            The chat response with output and messages.
+        """
+        # Get the last user message to send to the engine
+        message = self._get_last_user_message(request.messages)
+
+        # Call the chat engine
+        response = await self._engine.achat(message)
+
+        # Extract content from response
+        output = str(response.response) if hasattr(response, "response") else str(response)
+
+        # Build response messages (original + assistant response)
+        response_messages: list[dict[str, Any]] = [
+            {"role": m.role, "content": m.content} for m in request.messages
+        ]
+        response_messages.append({"role": "assistant", "content": output})
+
+        return ChatResponse(output=output, messages=response_messages)
+
+    async def invoke_stream(self, request: InvokeRequest) -> AsyncIterator[str]:
+        """Handle a streaming invoke request.
+
+        Args:
+            request: The invoke request with input data.
+
+        Yields:
+            JSON-encoded chunks from the stream.
+        """
+        # Extract query from input
+        if "query" in request.input:
+            query = request.input["query"]
+        elif "prompt" in request.input:
+            query = request.input["prompt"]
+        elif "message" in request.input:
+            query = request.input["message"]
+        else:
+            query = str(request.input)
+
+        # Stream from the chat engine
+        response = await self._engine.astream_chat(query)
+        async for token in response.async_response_gen():
+            yield json.dumps({"chunk": token})
+
+    async def chat_stream(self, request: ChatRequest) -> AsyncIterator[str]:
+        """Handle a streaming chat request.
+
+        Args:
+            request: The chat request with messages.
+
+        Yields:
+            JSON-encoded chunks from the stream.
+        """
+        # Get the last user message to send to the engine
+        message = self._get_last_user_message(request.messages)
+
+        # Stream from the chat engine
+        response = await self._engine.astream_chat(message)
+        async for token in response.async_response_gen():
+            yield json.dumps({"chunk": token})
+
+
+def wrap(engine: ChatEngine, name: str = "llamaindex-agent") -> LlamaIndexAdapter:
+    """Wrap a LlamaIndex chat engine for use with Reminix Runtime.
+
+    Args:
+        engine: A LlamaIndex chat engine (e.g., SimpleChatEngine, ContextChatEngine).
+        name: Name for the agent.
+
+    Returns:
+        A LlamaIndexAdapter instance.
+
+    Example:
+        ```python
+        from llama_index.core.chat_engine import SimpleChatEngine
+        from llama_index.llms.openai import OpenAI
+        from reminix_llamaindex import wrap
+        from reminix_runtime import serve
+
+        llm = OpenAI(model="gpt-4")
+        engine = SimpleChatEngine.from_defaults(llm=llm)
+        agent = wrap(engine, name="my-agent")
+        serve([agent], port=8080)
+        ```
+    """
+    return LlamaIndexAdapter(engine, name=name)

reminix_llamaindex-0.0.1.dist-info/METADATA

@@ -0,0 +1,148 @@
+Metadata-Version: 2.4
+Name: reminix-llamaindex
+Version: 0.0.1
+Summary: Reminix adapter for LlamaIndex
+License-Expression: Apache-2.0
+Requires-Python: >=3.10
+Requires-Dist: llama-index>=0.14.0
+Requires-Dist: reminix-runtime~=0.0.1
+Provides-Extra: dev
+Requires-Dist: pytest-asyncio>=0.24.0; extra == 'dev'
+Requires-Dist: pytest>=8.0.0; extra == 'dev'
+Description-Content-Type: text/markdown
+
+# reminix-llamaindex
+
+Reminix Runtime adapter for [LlamaIndex](https://www.llamaindex.ai/). Deploy LlamaIndex chat engines as a REST API.
+
+## Installation
+
+```bash
+pip install reminix-llamaindex
+```
+
+This will also install `reminix-runtime` as a dependency.
+
+## Quick Start
+
+```python
+from llama_index.core.chat_engine import SimpleChatEngine
+from llama_index.llms.openai import OpenAI
+from reminix_llamaindex import wrap
+from reminix_runtime import serve
+
+# Create a LlamaIndex chat engine
+llm = OpenAI(model="gpt-4o")
+engine = SimpleChatEngine.from_defaults(llm=llm)
+
+# Wrap it with the Reminix adapter
+agent = wrap(engine, name="my-chatbot")
+
+# Serve it as a REST API
+serve([agent], port=8080)
+```
+
+Your agent is now available at:
+- `POST /agents/my-chatbot/invoke` - Stateless invocation
+- `POST /agents/my-chatbot/chat` - Conversational chat
+
+## API Reference
+
+### `wrap(engine, name)`
+
+Wrap a LlamaIndex chat engine for use with Reminix Runtime.
+
+| Parameter | Type | Default | Description |
+|-----------|------|---------|-------------|
+| `engine` | `BaseChatEngine` | required | A LlamaIndex chat engine |
+| `name` | `str` | `"llamaindex-agent"` | Name for the agent (used in URL path) |
+
+**Returns:** `LlamaIndexAdapter` - A Reminix adapter instance
+
+### Example with RAG
+
+```python
+from llama_index.core import VectorStoreIndex, SimpleDirectoryReader
+from llama_index.llms.openai import OpenAI
+from reminix_llamaindex import wrap
+from reminix_runtime import serve
+
+# Load documents and create index
+documents = SimpleDirectoryReader("./data").load_data()
+index = VectorStoreIndex.from_documents(documents)
+
+# Create a chat engine with the index
+engine = index.as_chat_engine(llm=OpenAI(model="gpt-4o"))
+
+# Wrap and serve
+agent = wrap(engine, name="rag-chatbot")
+serve([agent], port=8080)
+```
+
+## Endpoint Input/Output Formats
+
+### POST /agents/{name}/invoke
+
+Stateless invocation for task-oriented operations.
+
+**Request:**
+```json
+{
+  "input": {
+    "query": "What is the capital of France?"
+  }
+}
+```
+
+Or with prompt:
+```json
+{
+  "input": {
+    "prompt": "Summarize this text: ..."
+  }
+}
+```
+
+**Response:**
+```json
+{
+  "output": "The capital of France is Paris."
+}
+```
+
+### POST /agents/{name}/chat
+
+Conversational chat. The adapter extracts the last user message.
+
+**Request:**
+```json
+{
+  "messages": [
+    {"role": "user", "content": "What is the capital of France?"}
+  ]
+}
+```
+
+**Response:**
+```json
+{
+  "output": "The capital of France is Paris.",
+  "messages": [
+    {"role": "user", "content": "What is the capital of France?"},
+    {"role": "assistant", "content": "The capital of France is Paris."}
+  ]
+}
+```
+
+## Runtime Documentation
+
+For information about the server, endpoints, request/response formats, and more, see the [`reminix-runtime`](https://pypi.org/project/reminix-runtime/) package.
+
+## Links
+
+- [GitHub Repository](https://github.com/reminix-ai/runtime-python)
+- [LlamaIndex Documentation](https://docs.llamaindex.ai/)
+
+## License
+
+Apache-2.0

reminix_llamaindex-0.0.1.dist-info/RECORD

@@ -0,0 +1,6 @@
+reminix_llamaindex/__init__.py,sha256=F6k9DeJDD5-cwD0VlVNnu4aaIssuWQ2Uyp5eEGkEc84,86
+reminix_llamaindex/adapter.py,sha256=7dVMR6nw8EL9A7KeavAwAvhFsKRO6pLdUIQRD6IH4UM,5691
+reminix_llamaindex/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+reminix_llamaindex-0.0.1.dist-info/METADATA,sha256=lo4BbUew1TOMCtbBvtCepDEu56tmJu22sCOd2g3tGRM,3368
+reminix_llamaindex-0.0.1.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+reminix_llamaindex-0.0.1.dist-info/RECORD,,

reminix_llamaindex-0.0.1.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.28.0
+Root-Is-Purelib: true
+Tag: py3-none-any