Schema to ER Diagram
The Lectern tool has no database. It runs as a scraper: fetch postings from job boards, run them through a filter, write the results to JSON and CSV files. There is no SQL, and no schema file to find. To draw the ER diagram, we had to read the code.
Schema to ER Diagram
The Lectern tool has no database. It runs as a scraper: fetch postings from job boards, run them through a filter, write the results to JSON and CSV files. There is no SQL, and no schema file to find. To draw the ER diagram, we had to read the code.
Series: Figma for Educational AI Topic: FIGMA · SCHEMA
Chapters
0:00 Introduction 0:27 Introduction 1:44 Your Turn
YOUR TURN
Your turn. Paste this into Claude: paste your Python function — the one that writes JSON or CSV. Ask Claude: write the ER diagram for this data model as a Mermaid erDiagram, then list every field that is only written under a condition — which ones did the diagram leave out? Then check one more thing: take a conditional field the diagram missed, and
@NikBearBrown
Part of the Figma for Educational AI series.
Sources and citations: see the reel's SOURCES.md in the open textbook.
▶ The full series playlist — Brutalist https://www.youtube.com/playlist?list=PLG9H-C6rp5RU
What this video covers
- 0:00 — Introduction
- 0:27 — Introduction
- 1:44 — Your Turn
Transcript
A lightly formatted transcript of the video, with timestamps so you can jump to any section on YouTube.
Introduction
[0:00] Malo, this is Liam in forbear. Can an agent draw a systems data model? Better question. Can an agent draw a systems data model from code alone with no database to query? Three terms. An er diagram. A picture of a data model. Boxes for tables. Lines for how records connect. A mermaid diagram. A plain text format where a few typed lines become boxes and edges in a drawing tool. And an entity, one kind of thing stored in the system. Each row is one instance.
Introduction
[0:27] The lectern tool has no database. It runs as a scraper. Fetch postings from job boards, run them through a filter, write the results to JSON and CSV files. There is no SQL and no schema file to find. To draw the ER diagram, we had to read the code. We described it in Mermaid. Five entities, four relationships, source, the job board and the company it tracks. Posting one job ad, 10 fields match what passed the filter. materials evidence what a passing post showed about how instructors work and keyword config the version stamped set of rules that drove the filter. We created a named Fig Jam board. Then one call generate diagram. It read the mermaid and
Drew the shapes. Five tables, the field lists, and the four relationship lines. Get fig back. Five tables, four connectors, every field hole. The board matched the mermaid. Four fields were missing. The keyword config table lost keep rule and materials rule. The two rules that decide which postings pass the filter and which show that an instructor assigns materials. The posting table lost reject reason written only when a posting fails. and the raw original record. Those four were in the code. They were not in what we wrote. Middle. The tool draws what it is given. Getting the schema right. Reading every function. Finding the conditional fields. Deciding which entities matter. That part was ours and we got four
Your Turn
[1:44] Wrong. Your turn. Paste this into Claude. Paste your Python function. The one that writes JSON or CSV. Ask Claude. Write the ER diagram for this data model as a mermaid diagram. Then list every field that is only written under a condition. Which ones did the diagram leave out? Then check one more thing. Take a conditional field the diagram missed and put it in yourself. Does the diagram change how you design the system? Schema to er diagram at Nick Bear Brown.
