This walkthrough summarizes how the Notion Life Tracker was built, in the order the pieces were added and refined.

• Created two databases: Blood Sugar Log and Food Log.
• Blood sugar properties: Entry (Title), Created time (Created time), Blood Sugar Level (Number).
• Food properties: food (Title/Text), Created time (Created time), plus macro fields like calories, protein, carbs, fats, fiber, sugar, sodium.

• Created a Notion integration, shared it with both databases.
• Stored secrets for Val Town:
  • NOTION_TOKEN
  • NOTION_BLOOD_SUGAR_DB_ID
  • NOTION_FOOD_DB_ID

• Implemented shared/notion.ts to query pages and update page properties.
• Added minimal Notion types and error handling for invalid responses.

• Implemented shared/date.ts for weekly ranges, date lists, and streak counts.
• Centralized UTC date logic for consistent weekly rollups.

• Implemented shared/blood_sugar_logic.ts:
  • Parse Notion pages into normalized entries.
  • Group by date and compute weekly stats (avg/min/max, completion rate, streaks).
  • Render email-friendly summaries.
• Added XP and badge logic for motivational feedback.

• Implemented collectors/blood_sugar_report.cron.tsx:
  • Query the last 7 days from Notion.
  • Build a summary report and send it via Val Town email.
  • Persist weekly rollups to SQLite.

• Added timezone-aware date extraction for blood sugar entries (America/New_York).
• Prevented late-night readings from rolling into the next UTC day.

• Implemented shared/food_enrich.ts:
  • Parse food entries and read existing macros from Notion.
  • Only enrich entries missing core macros.
  • Build update payloads for Notion.
  • Compute weekly food rollups and macro summaries.

• Implemented shared/fatsecret.ts:
  • OAuth token retrieval.
  • Food search and macro parsing from FatSecret responses.

• Implemented collectors/food_report.cron.tsx:
  • Fetch weekly food entries.
  • Normalize food names with OpenAI.
  • Fetch nutrition data and write macros back to Notion.
  • Persist weekly rollups to SQLite.

• Added fiber, sugar, and sodium parsing from FatSecret.
• Summed those fields across multi-item meals when all items provide values.

• Added a food.get call to pull detailed nutrition data when available.
• Kept food_description parsing as a fallback.

• Updated OpenAI normalization to emit { name, servings } per item.
• Scaled FatSecret macros by the inferred servings.
• Defaulted to one serving for typical meals (e.g., "shrimp lo mein").

• Fixed TypeScript inference edge cases and unused params.
• Verified deno lint passes across the codebase.

• Clarified required Notion fields.
• Documented the OpenAI normalization + FatSecret detail workflow.
• Added serving size examples for clarity.

If you want, I can expand this into a full contributor guide with diagrams or link each step to the exact files and functions.