Table of Contents
These examples showcase Jina in action and provide sample code for you to work from. We suggest you read Jina 101 and What is Neural Search? to get a conceptual overview.
❗If you'd like to run our examples on Windows, please follow this instruction.
To learn more about how to use Jina, please refer to our docs.
| Brand new to neural search? See a simple text-search example to understand how Jina works | |
| Index more effectively by adding incremental indexing to your Wikipedia search | |
| Get a better understanding of chunks by searching a lyrics database. Now with shiny front-end! | |
| Use SOTA visual representation for searching Pokémon! | |
| Detect, index and query similar objects | |
| A demo of neural search for audio data based Vggish model. | |
| Use prefetching and sharding to improve the performance of your index and query flow when searching animated GIFs. |
| Build a vector search engine that finds the closest vector in the database to a query. | |
| Use one modality (text) to search another (images) | |
| Use more than one modality (image+text) to search images | |
| Create separate indexes and queries for different clothing in Fashion-MNIST |
Want to add your own example? Please check our guidelines!
| Opinionated QA passage retrieval with BERT-based reranker |
You can run the perf-script.sh in order to run all of the examples on your machine. Make sure this is done in a separate python virtualenv.
This measures QPS for indexing and querying.
This will store the results in performance.txt.
Note that a lot of the examples are not optimized (or configured for scaling). They are provided as is, for the basic functionality.
You are highly encouraged to add a test for your example so that we will be alerted if it breaks in the future:
- Put your test data in the
testsfolder. The test data can be a few text sentences, images or audio samples - Create
test_[your_example].pyin thetestsfolder. Add your test cases to thetestsfile with meaningful asserts depending on example input and output - Run the test locally to confirm before pushing with pytest
- Add your example folder name to the
pathvariable inmatrixof.github/worflows/ci.yml. This will trigger your example test on creating a pull request.
- For reference, check out the
testsfolder from South Park example if your data is about text and object search example for images. - Try using the original example function by importing them to the test. Avoid any modifications to original Flow or logic.
- Use the pytest fixture
tmpdirfor temporary directory
- Slack channel - a communication platform for developers to discuss Jina
- LinkedIn - get to know Jina AI as a company and find job opportunities
- follow us and interact with us using hashtag
#JinaSearch- Company - know more about our company, we are fully committed to open-source!
Copyright (c) 2021 Jina AI Limited. All rights reserved.
Jina is licensed under the Apache License, Version 2.0. See LICENSE for the full license text.