mirror of https://github.com/hwchase17/langchain
Implemented Milvus translator for self-querying (#10162)
- Implemented the MilvusTranslator for self-querying using Milvus vector store - Made unit tests to test its functionality - Documented the Milvus self-queryingpull/10087/head^2
parent
7403faa063
commit
de3322609e
@ -0,0 +1,375 @@
|
||||
{
|
||||
"cells": [
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"# Self-querying with Milvus\n",
|
||||
"\n",
|
||||
"In the walkthrough we'll demo the `SelfQueryRetriever` with a `Milvus` vector store."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Creating a Milvus vectorstore\n",
|
||||
"First we'll want to create a Milvus VectorStore and seed it with some data. We've created a small demo set of documents that contain summaries of movies.\n",
|
||||
"\n",
|
||||
"I have used the cloud version of Milvus, thus I need `uri` and `token` as well.\n",
|
||||
"\n",
|
||||
"NOTE: The self-query retriever requires you to have `lark` installed (`pip install lark`). We also need the `pymilvus` package."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"#!pip install lark"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"#!pip install pymilvus"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"We want to use `OpenAIEmbeddings` so we have to get the OpenAI API Key."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 1,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"import os\n",
|
||||
"\n",
|
||||
"OPENAI_API_KEY = \"Use your OpenAI key:)\"\n",
|
||||
"\n",
|
||||
"os.environ[\"OPENAI_API_KEY\"] = OPENAI_API_KEY"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 2,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"from langchain.schema import Document\n",
|
||||
"from langchain.embeddings.openai import OpenAIEmbeddings\n",
|
||||
"from langchain.vectorstores import Milvus\n",
|
||||
"\n",
|
||||
"embeddings = OpenAIEmbeddings()"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"docs = [\n",
|
||||
" Document(page_content=\"A bunch of scientists bring back dinosaurs and mayhem breaks loose\", metadata={\"year\": 1993, \"rating\": 7.7, \"genre\": \"action\"}),\n",
|
||||
" Document(page_content=\"Leo DiCaprio gets lost in a dream within a dream within a dream within a ...\", metadata={\"year\": 2010,\"genre\": \"thriller\", \"rating\": 8.2}),\n",
|
||||
" Document(page_content=\"A bunch of normal-sized women are supremely wholesome and some men pine after them\", metadata={\"year\": 2019, \"rating\": 8.3, \"genre\": \"drama\"}),\n",
|
||||
" Document(page_content=\"Three men walk into the Zone, three men walk out of the Zone\", metadata={\"year\": 1979, \"rating\": 9.9, \"genre\": \"science fiction\"}),\n",
|
||||
" Document(\n",
|
||||
" page_content='A psychologist / detective gets lost in a series of dreams within dreams within dreams and Inception reused the idea',\n",
|
||||
" metadata={\"year\": 2006, \"genre\": \"thriller\", 'rating': 9.0},\n",
|
||||
" ),\n",
|
||||
" Document(page_content=\"Toys come alive and have a blast doing so\", metadata={\"year\": 1995, \"genre\": \"animated\", \"rating\": 9.3 }),\n",
|
||||
"]\n",
|
||||
"\n",
|
||||
"vector_store = Milvus.from_documents(\n",
|
||||
" docs,\n",
|
||||
" embedding=embeddings,\n",
|
||||
" connection_args={\"uri\": 'Use your uri:)', \"token\":'Use your token:)'}\n",
|
||||
")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Creating our self-querying retriever\n",
|
||||
"Now we can instantiate our retriever. To do this we'll need to provide some information upfront about the metadata fields that our documents support and a short description of the document contents."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 5,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"from langchain.llms import OpenAI\n",
|
||||
"from langchain.retrievers.self_query.base import SelfQueryRetriever\n",
|
||||
"from langchain.chains.query_constructor.base import AttributeInfo\n",
|
||||
"\n",
|
||||
"metadata_field_info = [\n",
|
||||
" AttributeInfo(\n",
|
||||
" name=\"genre\",\n",
|
||||
" description=\"The genre of the movie\",\n",
|
||||
" type=\"string\",\n",
|
||||
" ),\n",
|
||||
" AttributeInfo(\n",
|
||||
" name=\"year\",\n",
|
||||
" description=\"The year the movie was released\",\n",
|
||||
" type=\"integer\",\n",
|
||||
" ),\n",
|
||||
" AttributeInfo(\n",
|
||||
" name=\"rating\", description=\"A 1-10 rating for the movie\", type=\"float\"\n",
|
||||
" ),\n",
|
||||
"]\n",
|
||||
"document_content_description = \"Brief summary of a movie\"\n",
|
||||
"llm = OpenAI(temperature=0)\n",
|
||||
"retriever = SelfQueryRetriever.from_llm(\n",
|
||||
" llm, vector_store, document_content_description, metadata_field_info, verbose=True\n",
|
||||
")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Testing it out\n",
|
||||
"And now we can try actually using our retriever!"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 6,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"query='dinosaur' filter=None limit=None\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"[Document(page_content='A bunch of scientists bring back dinosaurs and mayhem breaks loose', metadata={'year': 1993, 'rating': 7.7, 'genre': 'action'}),\n",
|
||||
" Document(page_content='Toys come alive and have a blast doing so', metadata={'year': 1995, 'rating': 9.3, 'genre': 'animated'}),\n",
|
||||
" Document(page_content='Three men walk into the Zone, three men walk out of the Zone', metadata={'year': 1979, 'rating': 9.9, 'genre': 'science fiction'}),\n",
|
||||
" Document(page_content='A psychologist / detective gets lost in a series of dreams within dreams within dreams and Inception reused the idea', metadata={'year': 2006, 'rating': 9.0, 'genre': 'thriller'})]"
|
||||
]
|
||||
},
|
||||
"execution_count": 6,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"# This example only specifies a relevant query\n",
|
||||
"retriever.get_relevant_documents(\"What are some movies about dinosaurs\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 8,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"query=' ' filter=Comparison(comparator=<Comparator.GT: 'gt'>, attribute='rating', value=9) limit=None\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"[Document(page_content='Toys come alive and have a blast doing so', metadata={'year': 1995, 'rating': 9.3, 'genre': 'animated'}),\n",
|
||||
" Document(page_content='Three men walk into the Zone, three men walk out of the Zone', metadata={'year': 1979, 'rating': 9.9, 'genre': 'science fiction'})]"
|
||||
]
|
||||
},
|
||||
"execution_count": 8,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"# This example specifies a filter\n",
|
||||
"retriever.get_relevant_documents(\"What are some highly rated movies (above 9)?\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 7,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"query='toys' filter=Comparison(comparator=<Comparator.GT: 'gt'>, attribute='rating', value=9) limit=None\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"[Document(page_content='Toys come alive and have a blast doing so', metadata={'year': 1995, 'rating': 9.3, 'genre': 'animated'}),\n",
|
||||
" Document(page_content='Three men walk into the Zone, three men walk out of the Zone', metadata={'year': 1979, 'rating': 9.9, 'genre': 'science fiction'})]"
|
||||
]
|
||||
},
|
||||
"execution_count": 7,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"# This example only specifies a query and a filter\n",
|
||||
"retriever.get_relevant_documents(\"I want to watch a movie about toys rated higher than 9\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 10,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"query=' ' filter=Operation(operator=<Operator.AND: 'and'>, arguments=[Comparison(comparator=<Comparator.EQ: 'eq'>, attribute='genre', value='thriller'), Comparison(comparator=<Comparator.GTE: 'gte'>, attribute='rating', value=9)]) limit=None\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"[Document(page_content='A psychologist / detective gets lost in a series of dreams within dreams within dreams and Inception reused the idea', metadata={'year': 2006, 'rating': 9.0, 'genre': 'thriller'})]"
|
||||
]
|
||||
},
|
||||
"execution_count": 10,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"# This example specifies a composite filter\n",
|
||||
"retriever.get_relevant_documents(\"What's a highly rated (above or equal 9) thriller film?\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 11,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"query='dinosaur' filter=Operation(operator=<Operator.AND: 'and'>, arguments=[Comparison(comparator=<Comparator.GT: 'gt'>, attribute='year', value=1990), Comparison(comparator=<Comparator.LT: 'lt'>, attribute='year', value=2005), Comparison(comparator=<Comparator.EQ: 'eq'>, attribute='genre', value='action')]) limit=None\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"[Document(page_content='A bunch of scientists bring back dinosaurs and mayhem breaks loose', metadata={'year': 1993, 'rating': 7.7, 'genre': 'action'})]"
|
||||
]
|
||||
},
|
||||
"execution_count": 11,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"# This example specifies a query and composite filter\n",
|
||||
"retriever.get_relevant_documents(\n",
|
||||
" \"What's a movie after 1990 but before 2005 that's all about dinosaurs, \\\n",
|
||||
" and preferably has a lot of action\"\n",
|
||||
")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Filter k\n",
|
||||
"\n",
|
||||
"We can also use the self query retriever to specify `k`: the number of documents to fetch.\n",
|
||||
"\n",
|
||||
"We can do this by passing `enable_limit=True` to the constructor."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 12,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"retriever = SelfQueryRetriever.from_llm(\n",
|
||||
" llm, \n",
|
||||
" vector_store, \n",
|
||||
" document_content_description, \n",
|
||||
" metadata_field_info, \n",
|
||||
" verbose=True,\n",
|
||||
" enable_limit=True\n",
|
||||
")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 13,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"query='dinosaur' filter=None limit=2\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"[Document(page_content='A bunch of scientists bring back dinosaurs and mayhem breaks loose', metadata={'year': 1993, 'rating': 7.7, 'genre': 'action'}),\n",
|
||||
" Document(page_content='Toys come alive and have a blast doing so', metadata={'year': 1995, 'rating': 9.3, 'genre': 'animated'})]"
|
||||
]
|
||||
},
|
||||
"execution_count": 13,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"# This example only specifies a relevant query\n",
|
||||
"retriever.get_relevant_documents(\"What are two movies about dinosaurs?\")"
|
||||
]
|
||||
}
|
||||
],
|
||||
"metadata": {
|
||||
"kernelspec": {
|
||||
"display_name": "Python 3",
|
||||
"language": "python",
|
||||
"name": "python3"
|
||||
},
|
||||
"language_info": {
|
||||
"codemirror_mode": {
|
||||
"name": "ipython",
|
||||
"version": 3
|
||||
},
|
||||
"file_extension": ".py",
|
||||
"mimetype": "text/x-python",
|
||||
"name": "python",
|
||||
"nbconvert_exporter": "python",
|
||||
"pygments_lexer": "ipython3",
|
||||
"version": "3.11.4"
|
||||
},
|
||||
"orig_nbformat": 4
|
||||
},
|
||||
"nbformat": 4,
|
||||
"nbformat_minor": 2
|
||||
}
|
@ -0,0 +1,83 @@
|
||||
"""Logic for converting internal query language to a valid Milvus query."""
|
||||
from typing import Tuple, Union
|
||||
|
||||
from langchain.chains.query_constructor.ir import (
|
||||
Comparator,
|
||||
Comparison,
|
||||
Operation,
|
||||
Operator,
|
||||
StructuredQuery,
|
||||
Visitor,
|
||||
)
|
||||
|
||||
COMPARATOR_TO_BER = {
|
||||
Comparator.EQ: "==",
|
||||
Comparator.GT: ">",
|
||||
Comparator.GTE: ">=",
|
||||
Comparator.LT: "<",
|
||||
Comparator.LTE: "<=",
|
||||
}
|
||||
|
||||
UNARY_OPERATORS = [Operator.NOT]
|
||||
|
||||
|
||||
def process_value(value: Union[int, float, str]) -> str:
|
||||
# required for comparators involving strings
|
||||
if isinstance(value, str):
|
||||
# If the value is already a string, add double quotes
|
||||
return f'"{value}"'
|
||||
else:
|
||||
# If the valueis not a string, convert it to a string without double quotes
|
||||
return str(value)
|
||||
|
||||
|
||||
class MilvusTranslator(Visitor):
|
||||
"""Translate Milvus internal query language elements to valid filters."""
|
||||
|
||||
"""Subset of allowed logical operators."""
|
||||
allowed_operators = [Operator.AND, Operator.NOT, Operator.OR]
|
||||
|
||||
"""Subset of allowed logical comparators."""
|
||||
allowed_comparators = [
|
||||
Comparator.EQ,
|
||||
Comparator.GT,
|
||||
Comparator.GTE,
|
||||
Comparator.LT,
|
||||
Comparator.LTE,
|
||||
]
|
||||
|
||||
def _format_func(self, func: Union[Operator, Comparator]) -> str:
|
||||
self._validate_func(func)
|
||||
value = func.value
|
||||
if isinstance(func, Comparator):
|
||||
value = COMPARATOR_TO_BER[func]
|
||||
return f"{value}"
|
||||
|
||||
def visit_operation(self, operation: Operation) -> str:
|
||||
if operation.operator in UNARY_OPERATORS and len(operation.arguments) == 1:
|
||||
operator = self._format_func(operation.operator)
|
||||
return operator + "(" + operation.arguments[0].accept(self) + ")"
|
||||
elif operation.operator in UNARY_OPERATORS:
|
||||
raise ValueError(
|
||||
f'"{operation.operator.value}" can have only one argument in Milvus'
|
||||
)
|
||||
else:
|
||||
args = [arg.accept(self) for arg in operation.arguments]
|
||||
operator = self._format_func(operation.operator)
|
||||
return "(" + (" " + operator + " ").join(args) + ")"
|
||||
|
||||
def visit_comparison(self, comparison: Comparison) -> str:
|
||||
comparator = self._format_func(comparison.comparator)
|
||||
processed_value = process_value(comparison.value)
|
||||
attribute = comparison.attribute
|
||||
|
||||
return "( " + attribute + " " + comparator + " " + processed_value + " )"
|
||||
|
||||
def visit_structured_query(
|
||||
self, structured_query: StructuredQuery
|
||||
) -> Tuple[str, dict]:
|
||||
if structured_query.filter is None:
|
||||
kwargs = {}
|
||||
else:
|
||||
kwargs = {"expr": structured_query.filter.accept(self)}
|
||||
return structured_query.query, kwargs
|
@ -0,0 +1,116 @@
|
||||
from typing import Dict, Tuple
|
||||
|
||||
from langchain.chains.query_constructor.ir import (
|
||||
Comparator,
|
||||
Comparison,
|
||||
Operation,
|
||||
Operator,
|
||||
StructuredQuery,
|
||||
)
|
||||
from langchain.retrievers.self_query.milvus import MilvusTranslator
|
||||
|
||||
DEFAULT_TRANSLATOR = MilvusTranslator()
|
||||
|
||||
|
||||
def test_visit_comparison() -> None:
|
||||
comp = Comparison(comparator=Comparator.LT, attribute="foo", value=4)
|
||||
expected = "( foo < 4 )"
|
||||
actual = DEFAULT_TRANSLATOR.visit_comparison(comp)
|
||||
|
||||
assert expected == actual
|
||||
|
||||
|
||||
def test_visit_operation() -> None:
|
||||
# Non-Unary operator
|
||||
|
||||
op = Operation(
|
||||
operator=Operator.AND,
|
||||
arguments=[
|
||||
Comparison(comparator=Comparator.LT, attribute="foo", value=2),
|
||||
Comparison(comparator=Comparator.EQ, attribute="bar", value="baz"),
|
||||
Comparison(comparator=Comparator.LT, attribute="abc", value="4"),
|
||||
],
|
||||
)
|
||||
|
||||
expected = '(( foo < 2 ) and ( bar == "baz" ) ' 'and ( abc < "4" ))'
|
||||
actual = DEFAULT_TRANSLATOR.visit_operation(op)
|
||||
|
||||
assert expected == actual
|
||||
|
||||
# Unary operator: normal execution
|
||||
op = Operation(
|
||||
operator=Operator.NOT,
|
||||
arguments=[
|
||||
Comparison(comparator=Comparator.LT, attribute="foo", value=2),
|
||||
],
|
||||
)
|
||||
|
||||
expected = "not(( foo < 2 ))"
|
||||
actual = DEFAULT_TRANSLATOR.visit_operation(op)
|
||||
|
||||
assert expected == actual
|
||||
|
||||
# Unary operator: error
|
||||
op = Operation(
|
||||
operator=Operator.NOT,
|
||||
arguments=[
|
||||
Comparison(comparator=Comparator.LT, attribute="foo", value=2),
|
||||
Comparison(comparator=Comparator.EQ, attribute="bar", value="baz"),
|
||||
Comparison(comparator=Comparator.LT, attribute="abc", value="4"),
|
||||
],
|
||||
)
|
||||
|
||||
try:
|
||||
DEFAULT_TRANSLATOR.visit_operation(op)
|
||||
except ValueError as e:
|
||||
assert str(e) == '"not" can have only one argument in Milvus'
|
||||
else:
|
||||
assert False, "Expected exception not raised" # No exception -> test failed
|
||||
|
||||
|
||||
def test_visit_structured_query() -> None:
|
||||
query = "What is the capital of France?"
|
||||
structured_query = StructuredQuery(
|
||||
query=query,
|
||||
filter=None,
|
||||
)
|
||||
expected: Tuple[str, Dict] = (query, {})
|
||||
|
||||
actual = DEFAULT_TRANSLATOR.visit_structured_query(structured_query)
|
||||
assert expected == actual
|
||||
|
||||
comp = Comparison(comparator=Comparator.LT, attribute="foo", value=454)
|
||||
structured_query = StructuredQuery(
|
||||
query=query,
|
||||
filter=comp,
|
||||
)
|
||||
|
||||
expected = (
|
||||
query,
|
||||
{"expr": "( foo < 454 )"},
|
||||
)
|
||||
|
||||
actual = DEFAULT_TRANSLATOR.visit_structured_query(structured_query)
|
||||
assert expected == actual
|
||||
|
||||
op = Operation(
|
||||
operator=Operator.AND,
|
||||
arguments=[
|
||||
Comparison(comparator=Comparator.LT, attribute="foo", value=2),
|
||||
Comparison(comparator=Comparator.EQ, attribute="bar", value="baz"),
|
||||
Comparison(comparator=Comparator.LT, attribute="abc", value=50),
|
||||
],
|
||||
)
|
||||
|
||||
structured_query = StructuredQuery(
|
||||
query=query,
|
||||
filter=op,
|
||||
)
|
||||
|
||||
expected = (
|
||||
query,
|
||||
{"expr": "(( foo < 2 ) " 'and ( bar == "baz" ) ' "and ( abc < 50 ))"},
|
||||
)
|
||||
|
||||
actual = DEFAULT_TRANSLATOR.visit_structured_query(structured_query)
|
||||
assert expected == actual
|
Loading…
Reference in New Issue