[BUG] ML inference processor error in foreach processor

[ylwu-amzn]

What is the bug?
Run ML inference processor in foreach processor , and see errors.

How can one reproduce the bug?
Steps to reproduce the behavior:

1. Create bedrock embedding model, model id 1Fn9rI8BcrAWgSPyeLjY
2. Create ingest pipeline

PUT _ingest/pipeline/test_pipeline
{
  "description": "test embeddings",
  "processors": [
    {
      "foreach": {
        "field": "books",
        "processor": {
          "ml_inference": {
            "model_id": "1Fn9rI8BcrAWgSPyeLjY",
            "input_map": [
              {
                "input": "_ingest._value.title"
              }
            ],
            "output_map": [
              {
                "_ingest._value.title_embedding": "$.embedding"
              }
            ],
            "ignore_missing": true,
            "ignore_failure": false
          }
        }
      }
    }
  ]
}

3. Run pipeline and see error

{
  "docs": [
    {
      "error": {
        "root_cause": [
          {
            "type": "runtime_exception",
            "reason": "the prediction field: $.embedding is an array in size of 1536 but the document field array from field _ingest._value.title_embedding is in size of 0"
          }
        ],
        "type": "runtime_exception",
        "reason": "the prediction field: $.embedding is an array in size of 1536 but the document field array from field _ingest._value.title_embedding is in size of 0"
      }
    }
  ]
}

What is the expected behavior?
It should generate embedding and generate such result

{
    "books": [{
        "title": "first book",
        "title_embedding": [
            -1.1015625,
            0.65234375,
            0.7578125
            ...
        ],
        "description": "This is first book"
    }]
}

What is your host/environment?

• OS: [e.g. iOS]
• Version [e.g. 22]
• Plugins

Do you have any screenshots?
If applicable, add screenshots to help explain your problem.

Do you have any additional context?
Add any other context about the problem.