learning-extractor

To extract marketing patterns, you provide weekly diagnosis.json, calculated_metrics.json, and content used to the extractor. It derives evidence-based strategic insights and outputs structured JSON entries ready for your knowledge base.

You can append new patterns without duplicating existing entries because the extractor checks the current knowledge_base_marketing.json. It avoids overwriting existing patterns and only appends newly identified evidence-based findings.

You need calculated_metrics.json and diagnosis.json files to extract marketing patterns. The extractor requires these inputs alongside content used and your existing knowledge base to generate structured new_kb_entries JSON for appends.

A limitation of automated pattern extraction is that all newly generated findings must retain a tentative status. The extractor requires evidence-based claims from diagnosis runs and does not overwrite existing validated patterns in the knowledge base.