Checking to see if there is anything in the prompts.

Question

In my aws lambda function i have prompts that i recieve from Android Studio in Kotlin. Sometimes there will be one prompt, sometimes there will be three. if there is only one, in the "shots" section i only need that one prompt. if theres 2 i need both in there. The thing is, is i need to send all 3, whether there is any actaul prompt in all 3 or not, it could just be one like i said. so down in the shots section, i need the prompts that are actually prompts to be there. my codes below will help show this.

import json
import boto3

bedrock_runtime = boto3.client("bedrock-runtime")

def lambda_handler(event, context):

body = json.loads(event['body'])
    prompt = body['prompt']
    prompt2 = body['prompt2']
    prompt3 = body['prompt3']

model_input = {
    "taskType": "MULTI_SHOT_MANUAL",
    "multiShotManualParams": {
        "shots": 
        [
            {"text": prompt},
            {"text": prompt2},
            {"text": prompt3},
        ]
    },
    "videoGenerationConfig": {
        "seed": 33,
        "fps": 24,  # Must be 24
        "dimension": "1280x720",  # Must be "1280x720"
    },
}

try:
        # Start the asynchronous video generation job
        invocation = bedrock_runtime.start_async_invoke(
            modelId="amazon.nova-reel-v1:1",
            modelInput=model_input,
            outputDataConfig={
                "s3OutputDataConfig": {
                    "s3Uri": "s3://novabucket"
                }
            }
        )

# Extract the invocation ID and construct the future object key
        invocation_id = invocation['invocationArn'].split('/')[-1]
        future_object_key = f"{invocation_id}/output.mp4"

return {
            "statusCode": 200,
            "body": json.dumps({
                "invocationArn": invocation.get('invocationArn'),
                "requestId": invocation.get('requestId'),
                "futureObjectKey": f"{invocation_id}/output.mp4"
            })
        }

except Exception as e:
        message = str(e)
        return {
            "statusCode": 500,
            "body": json.dumps({"error": message})
        }

re:Post Agent · Answer

To handle varying numbers of prompts in your AWS Lambda function, you can modify your code to dynamically build the "shots" section based on the non-empty prompts received. Here's how you can adjust your lambda_handler function:

1. First, create a list of non-empty prompts:

```python
prompts = [body['prompt'], body['prompt2'], body['prompt3']]
non_empty_prompts = [p for p in prompts if p]
```

2. Then, dynamically build the "shots" section:

```python
shots = [{"text": prompt} for prompt in non_empty_prompts]
```

3. Update your model_input dictionary:

```python
model_input = {
    "taskType": "MULTI_SHOT_MANUAL",
    "multiShotManualParams": {
        "shots": shots
    },
    "videoGenerationConfig": {
        "seed": 33,
        "fps": 24,
        "dimension": "1280x720",
    },
}
```

This approach ensures that only non-empty prompts are included in the "shots" section, while still allowing for up to three prompts. If there's only one prompt, the "shots" list will contain only that prompt. If there are two or three, it will include all of them.

Remember to handle potential errors if no prompts are provided:

```python
if not non_empty_prompts:
    return {
        "statusCode": 400,
        "body": json.dumps({"error": "At least one prompt must be provided"})
    }
```

This modification allows your Lambda function to flexibly handle varying numbers of prompts without sending empty ones to the Bedrock runtime.
      
**Sources**  
[Changing AWS Lambda Function prompt value with an event instead of changing environment variable. | AWS re:Post](https://repost.aws/questions/QUPXwYdqAoSw6Qt67TqBR4JA/changing-aws-lambda-function-prompt-value-with-an-event-instead-of-changing-environment-variable)

Checking to see if there is anything in the prompts.

Añade tu respuesta

Contenido relevante