Task: Design the data structure in haih-cms
Design the data structure in haih-cms
Design the target data model for haih-cms based on the needs of the new website, using the legacy database solely as a data source rather than an architectural template.
Goal
Design a clean target data model for the new website on haih-cms. The legacy Rails model is used only for understanding the source data and should not be copied one-to-one.
What to Do
- Based on the completed inventory, identify the real entities needed by the new website.
- Design fields and relationships, keeping future editing and project evolution in mind.
- Separately describe the mapping between legacy tables/fields and the new entities.
- Do not carry over technical fields and relationships that only served the old Rails implementation.
- Provide for SEO fields, slugs, publication statuses, and relationships between content items.
- Account for the import of media and related files.
Result
The target haih-cms schema, optimized for the new website, plus a clear legacy-to-new data transformation map.
Definition of Done
- All migrated entities have a target representation.
- The legacy structure is not copied unnecessarily.
- Data transformation rules for import are clear.
- The model is suitable for the future evolution of the website independently of the Ruby project.
Ворклоги
Progress: Target model conceptually defined
The legacy database analysis confirmed heavy duplication of the content infrastructure: different types of materials are distributed across specialized tables while repeating the same features—text, URL, SEO, publication, sorting, etc.
A target decision has been made: content entities of the new site are, wherever possible, reduced to a universal Concept. Differences between a case study, service, FAQ, page, news article, and other materials are expressed through content, type, URI, hierarchy, and relationships, rather than the mandatory creation of a new table and CRUD.
This decision is viewed as the foundation for AI-driven content: the fewer special schemas an agent has to understand, the easier mass automated editing and further project development will be.