DocArray

The data structure for multimodal data · Neural Search · Vector Search · Document Store

[DocArray @docarray]

OMG finally 😻 what a merge! It's done, we pressed the release button for DocArray v2 🎉

If you're building a machine learning application that deals with multimodal data, then DocArray is the way to go 😼

https://t.co/hQy2EqKmKF

[DocArray @docarray]

We've just released DocArray 0.40. See the release notes here.

https://t.co/8rQV6wyrKa

[DocArray @docarray]

We've just released DocArray 0.39.1. See the release notes here.

https://t.co/C0eVrW6Q2P

[DocArray @docarray]

@pydantic See the full release notes here https://t.co/wqNfcqmDNE

[DocArray @docarray]

🚨DocArray now fully supports @pydantic v2 !

We have been working on supporting both versions of Pydantic and we finally released it today with DocArray v0.39.

It brings better performance overall of about +20% and more than +240% faster serialization of complex tensor!

[LF AI & Data Foundation @LFAIDataFdn]

🎉@docarray is now an Incubation Stage project!

Their dynamic schemas, vectorDBs integration, and seamless framework compatibility are fueling their journey of success.

Join us in celebrating their achievements! 🌟

Read more: https://t.co/GtsPADMaIs

[DocArray @docarray]

We've just released DocArray v0.37, featuring:

👉 @milvusio integration
👉 Filtering in HnswDocumentIndex
👉 Pre-filtering in InMemoryExactNNIndex
👉 Choosing your own tensor format in DocVec
👉 And more!

Read the release notes now: https://t.co/wlCMB0kUPR

[DocArray @docarray]

DocArray 0.36 is out! 🥳

New features:
👉 JAX Integration
👉 Redis Integration

Plus faster searching with HnswDocumentIndex and multiple bug fixes!

https://t.co/OnWCKj5YTH

[Jina AI @JinaAI_]

We've updated Jina to 3.19 with new features:

👉 Support for all versions of @DocArray
👉 Hubble adapts to the local DocArray version
👉 Support for pulling from private registries using Kubernetes secrets
👉 Rewrote documentation for new DocArray

https://t.co/s8RRZKLqHa

[JoanFM @JoanFM_]

Finally we have managed to bring all the new benefits introduced with @docarray refactor into https://t.co/zTTv1uOF09.

Now you can serve your models and applications with your own types and schemas. Jina adapts to your data instead of forcing you to adapt to Jina expected schema

[DocArray @docarray]

We all love when a perfect couple finds each other, like DocArray and Otter.

DocArray brings native multi-modal data structures to Otter, a native multi-modal model trained for in-context and few-shot learning. See how in this article from Jina AI.

https://t.co/yfs1ccHetM

[DocArray @docarray]

DocArray 0.35.0 is out! 🥳

New features:
👉DocVec now has the same serialization interface as DocList
👉File format validation on URLs
👉New methods to create BaseDoc instances from schemas

https://t.co/lLfGEKB9av

[DocArray @docarray]

Love @langchain 🦜⛓️?
Love DocArray?

Langchain just launched their new integrations hub and DocArray is right there in the mix - twice!

The best part: You can even give us a ❤️
https://t.co/Lv2Ji2ocD2

[DocArray @docarray]

⚡️Deserialize DocVec directly to numpy, torch, or tf

When you deserialize a DocVec from Protobuf, you can now choose if its tensors should be loaded to numpy, pytorch, or tensorflow:

[DocArray @docarray]

DocArray 0.34.0 is out! 🥳

🚨 BREAKING CHANGE: Requires Python 3.8 or higher!

New features:

🔍 Nested search in InMemoryExactNNIndex
�� Easily check if a Document is already in a DocumentIndex
⚡️ Deserialize DocVec directly to numpy, torch, or tf

https://t.co/CGw5Be0P0V

[DocArray @docarray]

💡 Easily check if a Document is already in a DocumentIndex

Just use the `in` keyword to see if a Document has already been indexed.

[Saba Sturua @jupyterjazz]

🎉 Excited to introduce DocArrayRetriever in @langchain!

Looking to link external data to LLMs? @docarray is your key to:
1⃣ Representing data flexibly
2⃣ Numerous data storage options
3⃣ Smooth integration into Langchain

Follow this thread for a step-by-step guide 🧵

1/5

[LlamaIndex 🦙 @llama_index]

[1] @supabase integration: https://t.co/WBplTJUGbY

Tair (@AlibabaGroup) integration: https://t.co/MxYy3wi86M

In-memory stores using @docarray:
https://t.co/xkXgihaJui

https://t.co/DEDnctRTZM

[samsja @samsja19]

This is a big deal.

Now @jinaAI_ fully support @docarray v2 and allows to build multi-modal AI API with an experience similar to @FastAPI

Here is how it looks like now

[Saba Sturua @jupyterjazz]

Thrilled to see @docarray in action within @DeepLearningAI_'s Langchain course, delivered by @AndrewYNg and @hwchase17!

DocArrayInMemorySearch really is an excellent foundation for kickstarting your own LLM applications with @langchain