Improve AI stories generator

Background

Within the AquaData project, we're developing a tool to enrich the aquadata portal (a website) with stories that showcase the impact of WorldFish initiatives globally. A draft version of this tool, using a Langchain summarization process (details here), is already integrated into the aquadata.data.mapping repository via this code.

Objective

The main goal is to optimize this tool for greater flexibility and effectiveness. Specifically, we want the tool to:

1. Handle multiple text data formats efficiently.
2. Produce outputs that are stylistically tailored to fit the specific narrative and "branding" requirements of the AquaData portal.

Relation to Other Tasks

This issue is closely linked to Issue #9. A key component of the AI story generator involves efficiently scraping and processing information from various file formats.

Proposed Actions

1. Enhance Text Data Ingestion: Develop capabilities to ingest and process various text data formats (e.g., PDFs, web pages, plain text) effectively.
2. Customize Output Styles: Implement functionality that allows the tool to adjust the style and tone of the output to align with the AquaData portal's guidelines.
3. Integration with Scraping Tool: Ensure seamless integration with the scraping tool being developed under Issue #9, focusing on the coherent flow of data from source to story generation.
4. Testing and Iteration: Conduct thorough testing with different data sources and styles to refine the tool's performance and output quality.

Additional Resources

• Project Documentation: AquaData Project
• Related Issues: #9 - Developing a Scraping Tool