For RAG tools, AWS Bedrock is supported as LLM, but not as embedder

[mintomohan]

I am attempting to perform RAG (Retrieval-Augmented Generation) using a CSV file, with AWS Bedrock as the provider.

Here is the code I used:

tool = CSVSearchTool(csv=r'path/to/file.csv',
            config=dict(
                llm=dict(
                    provider="aws_bedrock", 
                    config=dict(
                        model="amazon.titan-text-express-v1",
                    ),
                ),
                embedder=dict(
                    provider="aws_bedrock", 
                    config=dict(
                        model="amazon.titan-embed-text-v2:0",
                    ),
                ),
            )
        )

However, it seems the embedding model on Bedrock is not supported.

Logs:

\path\to\venv\lib\site-packages\langchain\_api\module_import.py:87: LangChainDeprecationWarning: Importing GuardrailsOutputParser from langchain.output_parsers is deprecated. Please replace the import with the following:
from langchain_community.output_parsers.rail_parser import GuardrailsOutputParser
  warnings.warn(
Traceback (most recent call last):
  File "\path\to\venv\lib\site-packages\schema.py", line 542, in validate
    return s.validate(data, **kwargs)
  File "\path\to\venv\lib\site-packages\schema.py", line 219, in validate
    raise SchemaError(
schema.SchemaError: Or('openai', 'gpt4all', 'huggingface', 'vertexai', 'azure_openai', 'google', 'mistralai', 'nvidia') did not validate 'aws_bedrock'
'openai' does not match 'aws_bedrock'
'gpt4all' does not match 'aws_bedrock'
'huggingface' does not match 'aws_bedrock'
'vertexai' does not match 'aws_bedrock'
'azure_openai' does not match 'aws_bedrock'
'google' does not match 'aws_bedrock'
'mistralai' does not match 'aws_bedrock'
'nvidia' does not match 'aws_bedrock'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "\path\to\venv\lib\site-packages\schema.py", line 483, in validate
    nvalue = Schema(
  File "\path\to\venv\lib\site-packages\schema.py", line 544, in validate
    raise SchemaError(
schema.SchemaError: Or('openai', 'gpt4all', 'huggingface', 'vertexai', 'azure_openai', 'google', 'mistralai', 'nvidia') did not validate 'aws_bedrock'
'openai' does not match 'aws_bedrock'
'gpt4all' does not match 'aws_bedrock'
'huggingface' does not match 'aws_bedrock'
'vertexai' does not match 'aws_bedrock'
'azure_openai' does not match 'aws_bedrock'
'google' does not match 'aws_bedrock'
'mistralai' does not match 'aws_bedrock'
'nvidia' does not match 'aws_bedrock'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "\path\to\venv\lib\site-packages\schema.py", line 483, in validate
    nvalue = Schema(
  File "\path\to\venv\lib\site-packages\schema.py", line 489, in validate
    raise SchemaError(
schema.SchemaError: Key 'provider' error:
Or('openai', 'gpt4all', 'huggingface', 'vertexai', 'azure_openai', 'google', 'mistralai', 'nvidia') did not validate 'aws_bedrock'
'openai' does not match 'aws_bedrock'
'gpt4all' does not match 'aws_bedrock'
'huggingface' does not match 'aws_bedrock'
'vertexai' does not match 'aws_bedrock'
'azure_openai' does not match 'aws_bedrock'
'google' does not match 'aws_bedrock'
'mistralai' does not match 'aws_bedrock'
'nvidia' does not match 'aws_bedrock'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "\path\to\src\simple_bot.py", line 19, in <module>
    csv_cost_tool = CSVSearchTool(csv=r'path/to/file.csv',
  File "\path\to\venv\lib\site-packages\crewai_tools\tools\csv_search_tool\csv_search_tool.py", line 32, in __init__
    super().__init__(**kwargs)
  File "\path\to\venv\lib\site-packages\pydantic\main.py", line 176, in __init__
    self.__pydantic_validator__.validate_python(data, self_instance=self)
  File "\path\to\venv\lib\site-packages\crewai_tools\tools\rag\rag_tool.py", line 47, in _set_default_adapter
    app = App.from_config(config=self.config) if self.config else App()
  File "\path\to\venv\lib\site-packages\embedchain\app.py", line 363, in from_config
    validate_config(config_data)
  File "\path\to\venv\lib\site-packages\embedchain\utils\misc.py", line 503, in validate_config
    return schema.validate(config_data)
  File "\path\to\venv\lib\site-packages\schema.py", line 489, in validate
    raise SchemaError(
schema.SchemaError: Key 'embedder' error:
Key 'provider' error:
Or('openai', 'gpt4all', 'huggingface', 'vertexai', 'azure_openai', 'google', 'mistralai', 'nvidia') did not validate 'aws_bedrock'
'openai' does not match 'aws_bedrock'
'gpt4all' does not match 'aws_bedrock'
'huggingface' does not match 'aws_bedrock'
'vertexai' does not match 'aws_bedrock'
'azure_openai' does not match 'aws_bedrock'
'google' does not match 'aws_bedrock'
'mistralai' does not match 'aws_bedrock'
'nvidia' does not match 'aws_bedrock'

Due to this issue, AWS Bedrock seems to be unusable in this context. Is there any workaround for this problem?