feat: first version of vecstore, docloader and embeddings working

dev.Dockerfile

@@ -3,7 +3,7 @@ FROM python:3.10-slim
 WORKDIR /app
 
 # Install Poetry
-RUN apt-get update && apt-get install gcc curl -y
+RUN apt-get update && apt-get install gcc curl build-essential -y
 RUN curl -sSL https://install.python-poetry.org | python3 -
 # # Add Poetry to PATH
 ENV PATH="${PATH}:/root/.local/bin"

pyproject.toml

@@ -34,6 +34,7 @@ openai = "^0.27.2"
 types-pyyaml = "^6.0.12.8"
 dill = "^0.3.6"
 pandas = "^1.5.3"
+chromadb = "^0.3.21"
 
 [tool.poetry.group.dev.dependencies]
 black = "^23.1.0"

src/backend/langflow/config.yaml

@@ -12,6 +12,7 @@ agents:
   - JsonAgent
   - CSVAgent
   - initialize_agent
+  - VectorStoreAgent
 
 prompts:
   - PromptTemplate

src/backend/langflow/custom/customs.py

@@ -8,6 +8,7 @@ CUSTOM_NODES = {
         "JsonAgent": nodes.JsonAgentNode(),
         "CSVAgent": nodes.CSVAgentNode(),
         "initialize_agent": nodes.InitializeAgentNode(),
+        "VectorStoreAgent": nodes.VectorStoreAgentNode(),
     },
 }
 

src/backend/langflow/graph/base.py

@@ -185,6 +185,10 @@ class Node:
     def build(self, force: bool = False) -> Any:
         if not self._built or force:
             self._build()
+        
+        #! Deepcopy is breaking for vectorstores
+        if self.base_type == 'vectorstores':
+            return self._built_object
         return deepcopy(self._built_object)
 
     def add_edge(self, edge: "Edge") -> None:

src/backend/langflow/graph/nodes.py

@@ -32,6 +32,10 @@ class AgentNode(Node):
                 chain_node.build(tools=self.tools)
 
             self._build()
+
+        #! Cannot deepcopy VectorStore
+        if self.node_type == "VectorStoreAgent":
+            return self._built_object
         return deepcopy(self._built_object)
 
 

src/backend/langflow/interface/agents/custom.py

@@ -11,7 +11,12 @@ from langchain.llms.base import BaseLLM
 from langchain.memory.chat_memory import BaseChatMemory
 from langchain.schema import BaseLanguageModel
 from langchain.tools.python.tool import PythonAstREPLTool
-
+from langchain.agents.agent_toolkits import (
+    VectorStoreToolkit,
+    VectorStoreInfo,
+)
+from langchain.vectorstores.base import VectorStore
+from langchain.agents.agent_toolkits.vectorstore.prompt import PREFIX as VECTORSTORE_PREFIX, ROUTER_PREFIX as VECTORSTORE_ROUTER_PREFIX
 
 class JsonAgent(AgentExecutor):
     """Json agent"""
@@ -97,6 +102,54 @@ class CSVAgent(AgentExecutor):
         return super().run(*args, **kwargs)
 
 
+class VectorStoreAgent(AgentExecutor):
+    """Vector Store agent"""
+
+    @staticmethod
+    def function_name():
+        return "VectorStoreAgent"
+
+    @classmethod
+    def initialize(cls, *args, **kwargs):
+        return cls.from_toolkit_and_llm(*args, **kwargs)
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+
+    @classmethod
+    def from_toolkit_and_llm(
+        cls,
+        llm: BaseLanguageModel,
+        name: str,
+        description: str,
+        vectorstore: VectorStore,
+        **kwargs: Any
+    ):
+        """Construct a vectorstore agent from an LLM and tools."""
+
+        vectorstore_info = VectorStoreInfo(
+            name=name,
+            description=description,
+            vectorstore=vectorstore
+        )
+        toolkit = VectorStoreToolkit(vectorstore_info=vectorstore_info, llm=llm)
+
+
+        tools = toolkit.get_tools()
+        prompt = ZeroShotAgent.create_prompt(tools, prefix=VECTORSTORE_PREFIX)
+        llm_chain = LLMChain(
+            llm=llm,
+            prompt=prompt,
+            callback_manager=None,
+        )
+        tool_names = [tool.name for tool in tools]
+        agent = ZeroShotAgent(llm_chain=llm_chain, allowed_tools=tool_names, **kwargs)
+        return AgentExecutor.from_agent_and_tools(agent=agent, tools=tools, verbose=True)
+
+    def run(self, *args, **kwargs):
+        return super().run(*args, **kwargs)
+
+
 class InitializeAgent(AgentExecutor):
     """Implementation of initialize_agent function"""
 
@@ -128,4 +181,5 @@ CUSTOM_AGENTS = {
     "JsonAgent": JsonAgent,
     "CSVAgent": CSVAgent,
     "initialize_agent": InitializeAgent,
+    "VectorStoreAgent": VectorStoreAgent,
 }

src/backend/langflow/interface/tools/constants.py

@@ -7,7 +7,7 @@ from langchain.agents.load_tools import (
 )
 from langchain.tools.json.tool import JsonSpec
 
-from langflow.interface.custom.types import PythonFunction
+from langflow.interface.tools.custom import PythonFunction
 
 FILE_TOOLS = {"JsonSpec": JsonSpec}
 CUSTOM_TOOLS = {"Tool": Tool, "PythonFunction": PythonFunction}

src/backend/langflow/interface/tools/custom.py

@@ -1,10 +1,10 @@
-from typing import Callable, Optional
-
-from pydantic import BaseModel, validator
-
 from langflow.utils import validate
 
 
+from typing import Callable, Optional
+from pydantic import BaseModel, validator
+
+
 class Function(BaseModel):
     code: str
     function: Optional[Callable] = None
@@ -30,8 +30,8 @@ class Function(BaseModel):
 
         return validate.create_function(self.code, function_name)
 
-
 class PythonFunction(Function):
     """Python function"""
 
     code: str
+

src/backend/langflow/template/nodes.py

@@ -248,6 +248,48 @@ class CSVAgentNode(FrontendNode):
         return super().to_dict()
 
 
+class VectorStoreAgentNode(FrontendNode):
+    name: str = "VectorStoreAgent"
+    template: Template = Template(
+        type_name="vectorstore_agent",
+        fields=[
+            TemplateField(
+                field_type="str",
+                required=True,
+                show=True,
+                name="name",
+                value="",
+            ),
+            TemplateField(
+                field_type="str",
+                required=True,
+                show=True,
+                name="description",
+                value="",
+            ),
+            TemplateField(
+                field_type="VectorStore",
+                required=True,
+                show=True,
+                name="vectorstore",
+                display_name="Vector Store",
+            ),
+            TemplateField(
+                field_type="BaseLanguageModel",
+                required=True,
+                show=True,
+                name="llm",
+                display_name="LLM",
+            ),
+        ],
+    )
+    description: str = """Construct a json agent from a CSV and tools."""
+    base_classes: list[str] = ["AgentExecutor"]
+
+    def to_dict(self):
+        return super().to_dict()
+
+
 class PromptFrontendNode(FrontendNode):
     @staticmethod
     def format_field(field: TemplateField, name: Optional[str] = None) -> None:

src/frontend/src/utils.ts

@@ -9,6 +9,7 @@ import {
   ComputerDesktopIcon,
   Bars3CenterLeftIcon,
   GiftIcon,
+  PaperClipIcon,
   QuestionMarkCircleIcon,
 } from "@heroicons/react/24/outline";
 import { Connection, Edge, Node, ReactFlowInstance } from "reactflow";

tests/test_custom_types.py

@@ -1,6 +1,6 @@
 # Test this:
 import pytest
-from langflow.interface.custom.types import PythonFunction
+from langflow.interface.tools.custom import PythonFunction
 from langflow.utils import constants