Add Tutorials for Amazon Rekognition #3290
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Mingshi Liu <mingshl@amazon.com>
mingshl
requested review from
b4sjoo,
dhrubo-os,
jngz-es,
model-collapse,
rbhavna,
ylwu-amzn,
zane-neo,
Zhangxunmt,
austintlee,
HenryL27 and
xinyual
as code owners
mingshl
temporarily deployed
to
ml-commons-cicd-env-require-approval
GitHub Actions
Inactive
mingshl
temporarily deployed
to
ml-commons-cicd-env-require-approval
GitHub Actions
Inactive
brianf-aws
suggested changes
Dec 23, 2024
| { | ||
| "parameters": { | ||
| "response_filter": "$.TextDetections.*.DetectedText", | ||
| "image_bytes": "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" |
There was a problem hiding this comment.
I think it would be nice if you added a mini explanation of what the image was.
There was a problem hiding this comment.
| - [2] Invokes the Amazon Rekognition DetectText API providing the image_bytes parameter. | ||
| - [3] Extracts values from the DetectText API response with JSON path. | ||
| - [4] Inserts the extracted values into the text field. | ||
| - [5] Removes the original image field. |
There was a problem hiding this comment.
Removes original image field may confuse users to think ML Inference Processor can do that. Maybe refactoring to
Step 6: Create Ingest pipeline
Explanation that the pipeline has two processors:
- ML Inference processor:
- Extracts values from the image field and passes the values to the image_bytes parameter.
- Invokes the Amazon Rekognition DetectText API providing the image_bytes parameter.
- Extracts values from the DetectText API response with JSON path.
- Inserts the extracted values from the API into a new field within the same document.
- Remove processor:
- Removes the base64 string field in the original document for clarity
For an idea
| Sample response: | ||
| ```json | ||
| { | ||
| "connector_id": "o52l5pMB6Ebhud5_ypxu" |
There was a problem hiding this comment.
As I look at other tutorials within the same folder Im seeing they follow the format of showing the connector id and model id. Maybe we should make them placeholders? such as your_connector_id
What do you think?
|
|
||
| ## Detect text with DetectText API | ||
|
|
||
| This tutorial demonstrates how to create an OpenSearch connector for an Amazon Rekognition model that detects text in images using the DetectText API. |
There was a problem hiding this comment.
You wrote that this is for the blueprint but also added the ML Inference processor. Maybe add that here too? Other tutorials just mention how to deploy and then run inference so this important info could get lost.
There was a problem hiding this comment.
good idea. will add this part
mingshl
temporarily deployed
to
ml-commons-cicd-env-require-approval
GitHub Actions
Inactive
ylwu-amzn
reviewed
Dec 24, 2024
| { | ||
| "persistent": { | ||
| "plugins.ml_commons.trusted_connector_endpoints_regex": [ | ||
| "^https://rekognition\\..*[a-z0-9-]\\.amazonaws\\.com$" |
There was a problem hiding this comment.
Can you cut a PR to add this to trusted URL setting?
There was a problem hiding this comment.
adding to my queue. raised an issue to track this #3308
ylwu-amzn
reviewed
Dec 24, 2024
| "session_token": "your_session_token" | ||
| }, | ||
| "parameters": { | ||
| "region": "us-west-2", |
There was a problem hiding this comment.
Suggested change
| "region": "us-west-2", | |
| "region": "your_aws_region_like_us-west-2", |
ylwu-amzn
reviewed
Dec 24, 2024
| POST _plugins/_ml/models/pp2n5pMB6Ebhud5_oJwF/_predict | ||
| { | ||
| "parameters": { | ||
| "response_filter": "$.TextDetections.*.DetectedText", |
There was a problem hiding this comment.
Suggest not add response_filter to show the raw model response first. Then introduce how to use response_filter to filter target parts.
ylwu-amzn
reviewed
Dec 30, 2024
Comment on lines
+252
to
+255
| "Platform", | ||
| "Search", | ||
| "for", | ||
| "anything" |
There was a problem hiding this comment.
These four results are words detected. They are duplicate with the first two results which are detected lines. Can we just return detected lines ?
There was a problem hiding this comment.
I was looking at the the documentation example and it looks like the API returns lines and words lines represent how we read it (search for anything) and the words are individual parts (search, for, anything).
I asked chatgpt if it were possible to filter (using only LINE) on the fly not sure if this works but worth a try $..TextDetections[?(@.Type=='LINE')].DetectedText
There was a problem hiding this comment.
Rekognition will detect the word or line of text recognized. Even though the words seems duplicated, the individual words can be used for tokenized word in search, they can be handy.
We can prefer keeping it this way, and suggest users to refer to rekognition API if they would like to do further filtering.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Add Tutorials for Amazon Rekognition
Check List
--signoff.By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.