can anyone suggest the timeline for release for milvus 2 support?

[rajeevkri]

can anyone suggest the timeline for release for milvus 2 support?

[bogdankostic]

Hi @rajeevkri! We do support Milvus 2 with our Milvus2DocumentStore (see here).

[bogdankostic]

With our latest release (version 1.0.0), you should be able to use the Milvus2DocumentStore. You can import it like this:

from haystack.document_stores.milvus2x import Milvus2DocumentStore

[rajeevkri]

I get below error when running above import statement, I have haystack and pymilvus client installed

`ModuleNotFoundError Traceback (most recent call last)
in
----> 1 from haystack.document_stores.milvus2x import Milvus2DocumentStore

~/.local/lib/python3.8/site-packages/haystack/init.py in
5 logging.getLogger("haystack").setLevel(logging.INFO)
6
----> 7 from haystack import pipelines
8 from haystack.schema import Document, Answer, Label, MultiLabel, Span
9 from haystack.nodes import BaseComponent

~/.local/lib/python3.8/site-packages/haystack/pipelines/init.py in
----> 1 from haystack.pipelines.base import Pipeline, RootNode, RayPipeline
2 from haystack.pipelines.standard_pipelines import (
3 BaseStandardPipeline,
4 DocumentSearchPipeline,
5 QuestionGenerationPipeline,

~/.local/lib/python3.8/site-packages/haystack/pipelines/base.py in
14 from networkx import DiGraph
15 from networkx.drawing.nx_agraph import to_agraph
---> 16 from haystack.nodes.evaluator.evaluator import calculate_em_str_multi, calculate_f1_str_multi, semantic_answer_similarity
17
18 try:

~/.local/lib/python3.8/site-packages/haystack/nodes/init.py in
1 from haystack.nodes.base import BaseComponent
----> 2 from haystack.nodes.answer_generator import BaseGenerator, RAGenerator, Seq2SeqGenerator
3 from haystack.nodes.connector import Crawler
4 from haystack.nodes.document_classifier import BaseDocumentClassifier, TransformersDocumentClassifier
5 from haystack.nodes.evaluator import EvalDocuments, EvalAnswers

~/.local/lib/python3.8/site-packages/haystack/nodes/answer_generator/init.py in
1 from haystack.nodes.answer_generator.base import BaseGenerator
----> 2 from haystack.nodes.answer_generator.transformers import RAGenerator, Seq2SeqGenerator

~/.local/lib/python3.8/site-packages/haystack/nodes/answer_generator/transformers.py in
12 from haystack.schema import Answer, Document
13 from haystack.nodes.answer_generator.base import BaseGenerator
---> 14 from haystack.nodes.retriever.dense import DensePassageRetriever
15
16

~/.local/lib/python3.8/site-packages/haystack/nodes/retriever/init.py in
----> 1 from haystack.nodes.retriever.base import BaseRetriever
2 from haystack.nodes.retriever.dense import DensePassageRetriever, EmbeddingRetriever, TableTextRetriever
3 from haystack.nodes.retriever.sparse import ElasticsearchRetriever, ElasticsearchFilterOnlyRetriever, TfidfRetriever
4 from haystack.nodes.retriever.text2sparql import Text2SparqlRetriever

~/.local/lib/python3.8/site-packages/haystack/nodes/retriever/base.py in
10 from haystack.schema import Document, MultiLabel
11 from haystack.nodes.base import BaseComponent
---> 12 from haystack.document_stores.base import BaseDocumentStore, BaseKnowledgeGraph
13
14

~/.local/lib/python3.8/site-packages/haystack/document_stores/init.py in
9 from haystack.document_stores.milvus2x import Milvus2DocumentStore as MilvusDocumentStore
10 else:
---> 11 from haystack.document_stores.milvus import MilvusDocumentStore # type: ignore
12
13 from haystack.document_stores.sql import SQLDocumentStore

~/.local/lib/python3.8/site-packages/haystack/document_stores/milvus.py in
9 from scipy.special import expit
10
---> 11 from milvus import IndexType, MetricType, Milvus, Status
12
13 from haystack.schema import Document

ModuleNotFoundError: No module named 'milvus'`

[rajeevkri]

[bogdankostic]

Can you try defining an environment variable called MILVUS2_ENABLED? This should solve this problem.

[rajeevkri]

Save and embedding is working fine for me but retrieval stucks in the search query
Can you please help with the same?

`import os
from haystack.retriever.dense import EmbeddingRetriever
from haystack.pipeline import FAQPipeline

os.environ["MILVUS2_ENABLED"] = "1"
print(os.getenv("MILVUS2_ENABLED"))

from haystack.document_stores.milvus2x import Milvus2DocumentStore

document_store_faq = Milvus2DocumentStore(index="test_data", port = "19530",
vector_dim =384)

retriever_faq = EmbeddingRetriever(document_store=document_store_faq, embedding_model='sentence-transformers/all-MiniLM-L6-v2', use_gpu=True)

pipeline = FAQPipeline(retriever=retriever_faq)

below query is stuck keeps on running indefinitely

search1 = pipeline.run(
query="What is the use of testing?"
)

print(search1)`