The data structure for multimodal data

> :arrow_up: **DocArray v2**: We are currently working on v2 of DocArray. Keep reading here if you are interested in the > current (stable) version, or check out the [v2 alpha branch](https://github.com/docarray/docarray/tree/feat-rewrite-v2#readme) > and [v2 roadmap](https://github.com/docarray/docarray/issues/780)! DocArray is a library for nested, unstructured, multimodal data in transit, including text, image, audio, video, 3D mesh, etc. It allows deep-learning engineers to efficiently process, embed, search, recommend, store, and transfer multimodal data with a Pythonic API. 🚪 **Door to multimodal world**: super-expressive data structure for representing complicated/mixed/nested text, image, video, audio, 3D mesh data. The foundation data structure of [Jina](https://github.com/jina-ai/jina), [CLIP-as-service](https://github.com/jina-ai/clip-as-service), [DALL·E Flow](https://github.com/jina-ai/dalle-flow), [DiscoArt](https://github.com/jina-ai/discoart) etc. 🧑‍🔬 **Data science powerhouse**: greatly accelerate data scientists' work on embedding, k-NN matching, querying, visualizing, evaluating via Torch/TensorFlow/ONNX/PaddlePaddle on CPU/GPU. 🚡 **Data in transit**: optimized for network communication, ready-to-wire at anytime with fast and compressed serialization in Protobuf, bytes, base64, JSON, CSV, DataFrame. Perfect for streaming and out-of-memory data. 🔎 **One-stop k-NN**: Unified and consistent API for mainstream vector databases that allows nearest neighbor search including Elasticsearch, Redis, AnnLite, Qdrant, Weaviate. 👒 **For modern apps**: GraphQL support makes your server versatile on request and response; built-in data validation and JSON Schema (OpenAPI) help you build reliable web services. 🐍 **Pythonic experience**: as easy as a Python list. If you can Python, you can DocArray. Intuitive idioms and type annotation simplify the code you write. 🛸 **IDE integration**: pretty-print and visualization on Jupyter notebook and Google Colab; comprehensive autocomplete and type hints in PyCharm and VS Code. Read more on [why should you use DocArray](https://docarray.jina.ai/get-started/what-is/) and [comparison to alternatives](https://docarray.jina.ai/get-started/what-is/#comparing-to-alternatives). DocArray was released under the open-source [Apache License 2.0](https://github.com/docarray/docarray/blob/main/LICENSE) in January 2022. It is currently a sandbox project under [LF AI & Data Foundation](https://lfaidata.foundation/). ## [Documentation](https://docarray.jina.ai) ## Install Requires Python 3.7+ ```shell pip install docarray ``` or via Conda: ```shell conda install -c conda-forge docarray ``` [Commonly used features](https://docarray.jina.ai/#install) can be enabled via `pip install "docarray[common]"`. ## Get Started DocArray consists of three simple concepts: - **Document**: a data structure for easily representing nested, unstructured data. - **DocumentArray**: a container for efficiently accessing, manipulating, and understanding multiple Documents. - **Dataclass**: a high-level API for intuitively representing multimodal data. Let's see DocArray in action with some examples. ### Example 1: represent multimodal data in a dataclass You can easily represent the following news article card with `docarray.dataclass` and type annotation:

```python from docarray import dataclass, Document from docarray.typing import Image, Text, JSON @dataclass class WPArticle: banner: Image headline: Text meta: JSON a = WPArticle( banner='https://.../cat-dog-flight.png', headline='Everything to know about flying with pets, ...', meta={ 'author': 'Nathan Diller', 'Column': 'By the Way - A Post Travel Destination', }, ) d = Document(a) ```

### Example 2: text matching in 10 lines Let's search for top-5 similar sentences of she smiled too much in "Pride and Prejudice": ```python from docarray import Document, DocumentArray d = Document(uri='https://www.gutenberg.org/files/1342/1342-0.txt').load_uri_to_text() da = DocumentArray(Document(text=s.strip()) for s in d.text.split('\n') if s.strip()) da.apply(Document.embed_feature_hashing, backend='process') q = ( Document(text='she smiled too much') .embed_feature_hashing() .match(da, metric='jaccard', use_scipy=True) ) print(q.matches[:5, ('text', 'scores__jaccard__value')]) ``` ```text [['but she smiled too much.', '_little_, she might have fancied too _much_.', 'She perfectly remembered everything that had passed in', 'tolerably detached tone. While she spoke, an involuntary glance', 'much as she chooses.”'], [0.3333333333333333, 0.6666666666666666, 0.7, 0.7272727272727273, 0.75]] ``` Here the feature embedding is done by simple [feature hashing](https://en.wikipedia.org/wiki/Feature_hashing) and distance metric is [Jaccard distance](https://en.wikipedia.org/wiki/Jaccard_index). You have better embeddings? Of course you do! We look forward to seeing your results! ### Example 3: external storage for out-of-memory data When your data is too big, storing in memory is not the best idea. DocArray supports [multiple storage backends](https://docarray.jina.ai/advanced/document-store/) such as SQLite, Weaviate, Qdrant and AnnLite. They're all unified under **the exact same user experience and API**. Take the above snippet: you only need to change one line to use SQLite: ```python da = DocumentArray( (Document(text=s.strip()) for s in d.text.split('\n') if s.strip()), storage='sqlite', ) ``` The code snippet can still run **as-is**. All APIs remain the same, the subsequent code then runs in an "in-database" manner. Besides saving memory, you can leverage storage backends for persistence and faster retrieval (e.g. on nearest-neighbor queries). ### Example 4: complete workflow of visual search Let's use DocArray and the [Totally Looks Like](https://sites.google.com/view/totally-looks-like-dataset) dataset to build a simple meme image search. The dataset contains 6,016 image-pairs stored in `/left` and `/right`. Images that share the same filename appear similar to the human eye. For example:

left/00018.jpg	right/00018.jpg	left/00131.jpg	right/00131.jpg

Given an image from `/left`, can we find the most-similar image to it in `/right`? (without looking at the filename). ### Load images First we load images. You *can* go to [Totally Looks Like](https://sites.google.com/view/totally-looks-like-dataset)'s website, unzip and load images as below: ```python from docarray import DocumentArray left_da = DocumentArray.from_files('left/*.jpg') ``` Or you can simply pull it from Jina AI Cloud: ```python left_da = DocumentArray.pull('jina-ai/demo-leftda', show_progress=True) ``` **Note** If you have more than 15GB of RAM and want to try using the whole dataset instead of just the first 1,000 images, remove `[:1000]` when loading the files into the DocumentArrays `left_da` and `right_da`. You'll see a progress bar to indicate how much has downloaded. To get a feeling of the data, we can plot them in one sprite image. You need matplotlib and torch installed to run this snippet: ```python left_da.plot_image_sprites() ```

### Apply preprocessing Let's do some standard computer vision pre-processing: ```python from docarray import Document def preproc(d: Document): return ( d.load_uri_to_image_tensor() # load .set_image_tensor_normalization() # normalize color .set_image_tensor_channel_axis(-1, 0) ) # switch color axis for the PyTorch model later left_da.apply(preproc) ``` Did I mention `apply` works in parallel? ### Embed images Now let's convert images into embeddings using a pretrained ResNet50: ```python import torchvision model = torchvision.models.resnet50(pretrained=True) # load ResNet50 left_da.embed(model, device='cuda') # embed via GPU to speed up ``` This step takes ~30 seconds on GPU. Beside PyTorch, you can also use TensorFlow, PaddlePaddle, or ONNX models in `.embed(...)`. ### Visualize embeddings You can visualize the embeddings via tSNE in an interactive embedding projector. You will need to have pydantic, uvicorn and FastAPI installed to run this snippet: ```python left_da.plot_embeddings(image_sprites=True) ```

Fun is fun, but our goal is to match left images against right images, and so far we have only handled the left. Let's repeat the same procedure for the right:

Pull from Cloud	Download, unzip, load from local
```python right_da = ( DocumentArray.pull('jina-ai/demo-rightda', show_progress=True) .apply(preproc) .embed(model, device='cuda')[:1000] ) ```	```python right_da = ( DocumentArray.from_files('right/*.jpg')[:1000] .apply(preproc) .embed(model, device='cuda') ) ```

### Match nearest neighbors Now we can match the left to the right and take the top-9 results. ```python left_da.match(right_da, limit=9) ``` Let's inspect what's inside `left_da` matches now: ```python for m in left_da[0].matches: print(d.uri, m.uri, m.scores['cosine'].value) ``` ```text left/02262.jpg right/03459.jpg 0.21102 left/02262.jpg right/02964.jpg 0.13871843 left/02262.jpg right/02103.jpg 0.18265384 left/02262.jpg right/04520.jpg 0.16477376 ... ``` Or shorten the loop to a one-liner using the element and attribute selector: ```python print(left_da['@m', ('uri', 'scores__cosine__value')]) ``` Better see it. ```python ( DocumentArray(left_da[8].matches, copy=True) .apply( lambda d: d.set_image_tensor_channel_axis( 0, -1 ).set_image_tensor_inv_normalization() ) .plot_image_sprites() ) ```

Here we reversed the preprocessing steps (i.e. switching axis and normalizing) on the copied matches, so you can visualize them using image sprites. ### Quantitative evaluation Serious as you are, visual inspection is surely not enough. Let's calculate the recall@K. First we construct the groundtruth matches: ```python groundtruth = DocumentArray( Document(uri=d.uri, matches=[Document(uri=d.uri.replace('left', 'right'))]) for d in left_da ) ``` Here we created a new DocumentArray with real matches by simply replacing the filename, e.g. `left/00001.jpg` to `right/00001.jpg`. That's all we need: if the predicted match has the identical `uri` as the groundtruth match, then it is correct. Now let's check recall rate from 1 to 5 over the full dataset: ```python for k in range(1, 6): print( f'recall@{k}', left_da.evaluate( groundtruth, hash_fn=lambda d: d.uri, metric='recall_at_k', k=k, max_rel=1 ), ) ``` ```text recall@1 0.02726063829787234 recall@2 0.03873005319148936 recall@3 0.04670877659574468 recall@4 0.052194148936170214 recall@5 0.0573470744680851 ``` You can also use other metrics like `precision_at_k`, `ndcg_at_k`, `hit_at_k`. If you think a pretrained ResNet50 is good enough, let me tell you with [Finetuner](https://github.com/jina-ai/finetuner) you can do much better with [just another ten lines of code](https://finetuner.jina.ai/notebooks/image_to_image/). ### Save results You can save a DocumentArray to binary, JSON, dict, DataFrame, CSV or Protobuf message with/without compression. In its simplest form: ```python left_da.save('left_da.bin') ``` To reuse that DocumentArray's data, use `left_da = DocumentArray.load('left_da.bin')`. If you want to transfer a DocumentArray from one machine to another or share it with your colleagues, you can do: ```python left_da.push('my_shared_da') ``` Now anyone who knows the token `my_shared_da` can pull and work on it. ```python left_da = DocumentArray.pull('/my_shared_da') ``` Intrigued? That's only scratching the surface of what DocArray is capable of. [Read our docs to learn more](https://docarray.jina.ai). ## Support - Join our [Discord server](https://discord.gg/WaMp6PVPgR) and chat with other community members about ideas. > DocArray is a trademark of LF AI Projects, LLC