Data Harmonisation for Drug Development Pipeline

[necibeahat]

A new bedrock agent that dynamically collects and harmonises pipeline data from sample pharmaceutical companies. It has an agent that can answer questions based on the enriched data.

[hasanp87]

@necibeahat thanks for the new agent PR, some initial comments below :

• Great use case for reviewing the drug pipeline across pharma companies

• What is not clear is the separation between the pre-requisite data processing vs the agent capability itself. Noted that the initial data processing was performed by Q cli, can this be something that the agent can perform natively as well ? If not, can you make it more clear in the README by highlighting the capabilities of the agent

• Any reason to have two separate cfn templates for the agent and its components, can you simplify to combine into a single cfn template

• Can you include the ontologies used for data harmonization in the README as well

[necibeahat]

In the project README.MD file, I've added

• Prerequisite subsection describing the data collection process with Q CLI and Fetch MCP. It states that you can alternatively use Strands Agents SDK or Bedrock Inline Agents with the MCP.
• A new section on the key ontologies used for the data harmonisation and standardisation.
• Description on agent capabilities such as using a knowledgebase as a tool
• I've left the two CFN as they are. The first CFN is needed to create AOSS. There is a manual step to create the index for the vector. The second CFN creates the Agent and KB. I wasn't able to successfully create the vector index viw CFN. All the examples from AWS include the manual step also used in this project.

[necibeahat]

Changes:

• created simplified CFN template
• clear description of ontologies, agent capabilities and next steps.