Monorepo architecture

This app is divided in multiple packages :

• backend is responsible for the user authentication, pdf generation and some static data serving
• frontend contains the main app logic

DB Workflow

erDiagram
    chip {
        text key PK
        text text
        text udap_id PK
        text value PK
    }

    clause {
        text id PK
        text label
        text value
    }

    delegation {
        text createdBy PK,FK
        text delegatedTo PK,FK
    }

    report {
        text applicantAddress
        text applicantName
        text contacts
        timestamp_without_time_zone createdAt
        text createdByEmail FK
        text decision
        text furtherInformation
        text id PK
        timestamp_without_time_zone meetDate
        text precisions
        text projectCadastralRef
        text projectDescription
        text projectSpaceType
        text redactedBy
        text serviceInstructeur
        text title
    }

    report_to_clause {
        text clauseId FK
        text id PK
        text reportId FK
    }

    udap {
        text address
        text city
        text completeCoords
        text department
        text email
        text id PK
        text name
        text phone
        boolean visible
        text zipCode
    }

    user {
        text email PK
        text name
        text password
        text temporaryLink
        text temporaryLinkExpiresAt
        text udap_id FK
    }

    report_to_clause }o--|| clause : "clauseId"
    delegation }o--|| user : "createdBy"
    delegation }o--|| user : "delegatedTo"
    report }o--|| user : "createdByEmail"
    report_to_clause }o--|| report : "reportId"
    user }o--|| udap : "udap_id"

Loading

Live data

PowerSync allows users to be notified when data changes.

To achieve this, every table containing live values must be part of the "powersync" publication.

ALTER PUBLICATION powersync ADD TABLE "table";

PowerSync documentation

Migrations

• Write SQL migrations in db/migrations/
• pnpm migration:up

DB Clients

• Both packages use Kysely
• pnpm backend pull-types generated a single .d.ts file describing the database structure
• frontend/src/db/AppSchema.ts contains the schema that will be used against indexed-db

  It is quite similar to the generated .d.ts file except it uses SQLite types (so TIMESTAMP or JSONB becomes TEXT)

Scripts

• clearDb.sh clears postgres db
• frontend/createEnvFile.ts used in prod to inject env vars starting with VITE_ at runtime
• frontend/generatePandaDS.ts used in dev to generate PandaCSS tokens from DSFR