feat(types): make apply and apply_batch self-returning

[hanxiao]

import onnxruntime

model = onnxruntime.InferenceSession('resnet.onnx')

from jina import DocumentArray, Document


def pre_proc(d: Document):
    return (
        d.load_uri_to_image_blob()
        .set_image_blob_shape((224, 224))
        .set_image_blob_normalization()
        .set_image_blob_channel_axis(-1, 0)
    )


def post_proc(d: Document):
    return d.set_image_blob_channel_axis(0, -1).set_image_blob_inv_normalization()


def embed(da):
    return da.embed(model)


(
    DocumentArray.from_files('/Users/hanxiao/Downloads/left/*.jpg')[:1000]
    .apply(pre_proc)
    .apply_batch(embed, batch_size=16)
    .apply(post_proc)
    .plot_embeddings(image_sprites=True)
)

[github-actions]

Latency summary

Current PR yields:

• 😶 index QPS at 1270, delta to last 2 avg.: -4%
• 🐢🐢 query QPS at 24, delta to last 2 avg.: -12%
• 🐢🐢 dam extend QPS at 35472, delta to last 2 avg.: -15%
• 🐢🐢 avg flow time within 1.5315 seconds, delta to last 2 avg.: +6%
• 😶 import jina within 0.4353 seconds, delta to last 2 avg.: +2%

Breakdown

Version	Index QPS	Query QPS	DAM Extend QPS	Avg Flow Time (s)	Import Time (s)
current	1270	24	35472	1.5315	0.4353
2.5.1	1529	33	51651	1.283	0.3385
2.5.0	1134	20	32393	1.5857	0.509

Backed by latency-tracking. Further commits will update this comment.

[codecov]

Codecov Report

Merging #4018 (750e3f0) into master (e5326e6) will decrease coverage by 4.10%.
The diff coverage is 28.57%.

@@            Coverage Diff             @@
##           master    #4018      +/-   ##
==========================================
- Coverage   79.46%   75.35%   -4.11%     
==========================================
  Files         183      183              
  Lines       12926    12934       +8     
==========================================
- Hits        10272     9747     -525     
- Misses       2654     3187     +533     

Flag	Coverage Δ
daemon	43.25% <28.57%> (-0.02%)	⬇️
jina	75.16% <28.57%> (-1.13%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
jina/types/arrays/mixins/embed.py	14.92% <0.00%> (-0.23%)	⬇️
jina/types/arrays/mixins/parallel.py	30.95% <22.22%> (-1.49%)	⬇️
jina/types/document/mixins/sugar.py	39.13% <33.33%> (-3.73%)	⬇️
jina/__init__.py	70.23% <100.00%> (ø)
jina/peapods/pods/k8slib/kubernetes_deployment.py	20.00% <0.00%> (-74.67%)	⬇️
jina/peapods/pods/k8slib/kubernetes_tools.py	20.22% <0.00%> (-64.05%)	⬇️
jina/parsers/deprecated.py	33.33% <0.00%> (-50.01%)	⬇️
jina/peapods/pods/k8slib/kubernetes_client.py	20.58% <0.00%> (-50.00%)	⬇️
jina/peapods/pods/k8s.py	33.05% <0.00%> (-46.78%)	⬇️
jina/types/document/generators.py	48.18% <0.00%> (-40.91%)	⬇️
... and 29 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b039379...750e3f0. Read the comment docs.