This project contains a Rust AWS Lambda function that uses the rust-bert library for sequence classification. It utilizes the Lambda Runtime and Simple Logger to facilitate interaction with AWS Lambda and logging capabilities, respectively.

The Lambda function processes input text for sequence classification using a pre-trained BERT model from the rust-bert library. The input is expected to be a JSON object containing a text field.

• Rust and Cargo (latest stable version)
• Docker
• AWS CLI configured with appropriate permissions

Clone the repository and navigate into the project directory:

git clone [URL_TO_REPOSITORY]
cd [PROJECT_DIRECTORY]

To build the project locally, run:

cargo build --release

To build the Lambda function for deployment, use the provided Dockerfile:

docker build -t rust-transformer-lambda .

This Dockerfile is configured to build the Rust project in a containerized environment, ensuring compatibility with the AWS Lambda execution environment.

Create an AWS Lambda function using the AWS CLI:

aws lambda create-function --function-name rust-transformer-lambda \
--handler doesnt.matter \
--zip-file fileb://./target/lambda/release/bootstrap.zip \
--runtime provided.al2 \
--role arn:aws:iam::[ACCOUNT_ID]:role/[LAMBDA_EXEC_ROLE] \
--environment Variables={RUST_BACKTRACE=1} \
--memory-size 1024 \
--timeout 15

To update the function code, build the Docker image and extract the binary to upload to AWS Lambda:

docker run --rm --entrypoint cat rust-transformer-lambda /home/rust/src/target/x86_64-unknown-linux-musl/release/bootstrap > bootstrap
zip lambda.zip bootstrap

aws lambda update-function-code --function-name rust-transformer-lambda \
--zip-file fileb://lambda.zip

To invoke the Lambda function, you can use the AWS CLI:

aws lambda invoke --function-name rust-transformer-lambda \
--payload '{"text":"example"}' response.json

Check response.json for the output.

The Lambda function my_handler expects an event with a JSON object containing a "text" key. It uses this text for sequence classification and returns the model's predictions as JSON.