gw-svelte/scripts/clean-legacy-onboarding.mjs

#!/usr/bin/env node
// Cleans the legacy Gravity Forms onboarding CSV export into a structured JSON
// file ready for later import into Postgres.
//
// Input:  dog-enrolment-form-2026-05-20.csv (repo root)
// Output: data/legacy-onboarding.json
//
// Run:    node scripts/clean-legacy-onboarding.mjs

import { readFileSync, writeFileSync, mkdirSync } from 'node:fs';
import { dirname, resolve } from 'node:path';
import { fileURLToPath } from 'node:url';

const __dirname = dirname(fileURLToPath(import.meta.url));
const ROOT = resolve(__dirname, '..');
const INPUT = resolve(ROOT, 'dog-enrolment-form-2026-05-20.csv');
const OUTPUT = resolve(ROOT, 'data/legacy-onboarding.json');

function parseCsv(text) {
  const rows = [];
  let cur = [];
  let val = '';
  let inQ = false;
  for (let i = 0; i < text.length; i++) {
    const c = text[i];
    if (inQ) {
      if (c === '"') {
        if (text[i + 1] === '"') { val += '"'; i++; }
        else inQ = false;
      } else val += c;
    } else {
      if (c === '"') inQ = true;
      else if (c === ',') { cur.push(val); val = ''; }
      else if (c === '\n') { cur.push(val); rows.push(cur); cur = []; val = ''; }
      else if (c === '\r') { /* skip */ }
      else val += c;
    }
  }
  if (val.length || cur.length) { cur.push(val); rows.push(cur); }
  return rows;
}

const trimOrNull = (v) => {
  const s = (v ?? '').trim();
  if (!s) return null;
  if (s.toLowerCase() === 'select') return null;
  return s;
};

const yesNo = (v) => {
  const s = (v ?? '').trim().toLowerCase();
  if (!s || s === 'select') return null;
  if (s === 'yes') return true;
  if (s === 'no') return false;
  return null;
};

// reactive-to-animals style: 'Yes'/'No'/'' or a free-text species ('Cats', 'Birds', 'Other').
const reactiveField = (v) => {
  const s = (v ?? '').trim();
  if (!s) return { reactive: null, detail: null };
  const lower = s.toLowerCase();
  if (lower === 'yes') return { reactive: true, detail: null };
  if (lower === 'no') return { reactive: false, detail: null };
  if (lower === 'select') return { reactive: null, detail: null };
  return { reactive: true, detail: s };
};

const parseFloatOrNull = (v) => {
  const s = (v ?? '').trim();
  if (!s) return null;
  const n = Number(s);
  return Number.isFinite(n) ? n : null;
};

// Best-effort NZ phone normalisation. Keeps the raw, adds an e164 when we can
// confidently produce one. Doesn't try to invent missing data.
function normalizePhone(raw) {
  const original = (raw ?? '').trim();
  if (!original) return { raw: null, e164: null };
  let digits = original.replace(/[^\d+]/g, '');
  if (digits.startsWith('+')) {
    return { raw: original, e164: '+' + digits.slice(1).replace(/\D/g, '') };
  }
  digits = digits.replace(/\D/g, '');
  if (!digits) return { raw: original, e164: null };
  if (digits.startsWith('64')) return { raw: original, e164: '+' + digits };
  if (digits.startsWith('0')) return { raw: original, e164: '+64' + digits.slice(1) };
  return { raw: original, e164: null };
}

const isoDateOrNull = (v) => {
  const s = (v ?? '').trim();
  if (!s) return null;
  // CSV already uses YYYY-MM-DD or 'YYYY-MM-DD HH:MM:SS'
  return s;
};

function emailGuess(ownerName, ownerSurname) {
  // Legacy CSV doesn't include owner email. Leaving null so the Postgres
  // import treats these as needing email-claim before they can sign in.
  return null;
}

const raw = readFileSync(INPUT, 'utf8').replace(/^/, '');
const rows = parseCsv(raw);
const headers = rows[0];
const data = rows.slice(1).filter((r) => r.some((cell) => (cell ?? '').trim() !== ''));

// Column index lookups (defensive in case header order shifts).
const idx = Object.fromEntries(headers.map((h, i) => [h.trim(), i]));
const col = (row, name) => row[idx[name]] ?? '';

const records = data.map((row) => {
  const hasFoodAllergy = yesNo(col(row, 'Does your dog have any food allergies?'));
  const foodAllergyDetail = trimOrNull(col(row, 'Specify'));
  // Two columns are labelled 'Specify'. The first is food, then env, then diet,
  // then medication — pick them by position.
  const specifyCols = headers
    .map((h, i) => ({ h, i }))
    .filter((x) => x.h.trim() === 'Specify')
    .map((x) => x.i);
  const [foodSpecifyIdx, envSpecifyIdx, dietSpecifyIdx, medSpecifyIdx] = specifyCols;

  const phone = normalizePhone(col(row, 'Owner Contact'));
  const emergencyPhone = normalizePhone(col(row, 'Emergency Contact Number'));
  const vetPhone = normalizePhone(col(row, 'Vet Contact Number'));

  const dogBirth = isoDateOrNull(col(row, "Dog's date of birth"));
  const signedOn = isoDateOrNull(col(row, 'Date'));
  const entryDate = isoDateOrNull(col(row, 'Entry Date'));
  const updatedDate = isoDateOrNull(col(row, 'Date Updated'));

  const ownerFirst = trimOrNull(col(row, 'Owner Name'));
  const ownerLast = trimOrNull(col(row, 'Owner Surname'));

  return {
    legacy: {
      entryId: trimOrNull(col(row, 'Entry Id')),
      entryDate,
      dateUpdated: updatedDate,
      createdByUserId: trimOrNull(col(row, 'Created By (User Id)')),
      sourceUrl: trimOrNull(col(row, 'Source Url')),
      signatureUrl: trimOrNull(col(row, 'Signature')),
      pdfUrl: trimOrNull(col(row, 'PDF: DogOnboardingForm')),
      userAgent: trimOrNull(col(row, 'User Agent')),
      userIp: trimOrNull(col(row, 'User IP')),
    },
    owner: {
      firstName: ownerFirst,
      lastName: ownerLast,
      fullName: [ownerFirst, ownerLast].filter(Boolean).join(' ') || null,
      email: emailGuess(ownerFirst, ownerLast),
      phone: phone.e164,
      phoneRaw: phone.raw,
      address: null, // not collected in legacy form
    },
    emergencyContact: {
      name: trimOrNull(col(row, 'Emergency Contact Name')),
      phone: emergencyPhone.e164,
      phoneRaw: emergencyPhone.raw,
    },
    dog: {
      name: trimOrNull(col(row, 'Dog Name')),
      surname: trimOrNull(col(row, 'Dog Surname')),
      dateOfBirth: dogBirth,
      breed: null, // not collected in legacy form
    },
    vet: {
      name: trimOrNull(col(row, 'Vet Name')),
      address: trimOrNull(col(row, 'Vet Address')),
      phone: vetPhone.e164,
      phoneRaw: vetPhone.raw,
    },
    health: {
      vaccinated: yesNo(col(row, 'Is your dog vaccinated?')),
      foodAllergy: {
        present: hasFoodAllergy,
        detail: hasFoodAllergy === true ? trimOrNull(row[foodSpecifyIdx]) : null,
      },
      environmentalAllergy: (() => {
        const present = yesNo(col(row, 'Does your dog have any environmental allergy?'));
        return {
          present,
          detail: present === true ? trimOrNull(row[envSpecifyIdx]) : null,
        };
      })(),
      specialDiet: (() => {
        const present = yesNo(col(row, 'Is your dog on a special diet?'));
        return {
          present,
          detail: present === true ? trimOrNull(row[dietSpecifyIdx]) : null,
        };
      })(),
      medication: (() => {
        const present = yesNo(col(row, 'Is your dog taking any medication that could put him at risk during a walk'));
        return {
          present,
          detail: present === true ? trimOrNull(row[medSpecifyIdx]) : null,
        };
      })(),
    },
    behaviour: {
      wellSocialised: yesNo(col(row, 'Is your dog well socialised?')),
      dogsInteractedWeekly: trimOrNull(col(row, 'How many dogs does your dog interact with weekly (excluding your family dogs)')),
      beachNotes: trimOrNull(col(row, 'Does your dog visit the beach?')),
      dogParkNotes: trimOrNull(col(row, 'Does your dog visit dog parks frequently - How many times a week?')),
      biteHistory: yesNo(col(row, 'Does your dog have a bite history?')),
      reactiveToDogs: yesNo(col(row, 'Is your dog reactive to other dogs?')),
      reactiveToAnimals: reactiveField(col(row, 'Is your dog reactive to other animals?')),
      reactiveToChildren: yesNo(col(row, 'Is your dog reactive to children?')),
      reactiveToPeople: yesNo(col(row, 'Is your dog reactive to other people?')),
      desexed: yesNo(col(row, 'Is your dog desexed?')),
      councilRegistered: yesNo(col(row, 'Is your dog registered?')),
      leashTrained: yesNo(col(row, 'Is your dog leash trained?')),
      recallRating: parseFloatOrNull(col(row, "Rate your dog's recall from 1 to 5, with one being the lowest score and 5 the highest.")),
      ranAwayBefore: yesNo(col(row, 'Has your dog ran away before?')),
      carBehaviour: trimOrNull(col(row, "Please describe your dog's behaviour in the car")),
      knownCommands: trimOrNull(col(row, 'List of commands your dog understands')),
    },
    misc: {
      additionalNotes: trimOrNull(col(row, "Anything else you'd like to let us know?")),
      socialMediaAccount: trimOrNull(col(row, 'Social Media Account Name')),
      howDidYouHear: trimOrNull(col(row, 'How did you hear about Goodwalk?')),
      referredBy: trimOrNull(col(row, "Person's name who reffered Goodwalk to you")),
    },
    declaration: {
      signedOn,
      signatureUrl: trimOrNull(col(row, 'Signature')),
      // Legacy submissions implicitly accepted T&Cs when submitted.
      termsAccepted: true,
      emergencyVetConsent: true,
    },
  };
});

mkdirSync(dirname(OUTPUT), { recursive: true });

const payload = {
  exportedAt: new Date().toISOString(),
  source: {
    file: 'dog-enrolment-form-2026-05-20.csv',
    rows: data.length,
    columns: headers.length,
  },
  notes: [
    'Legacy Gravity Forms export. Owner email/address were not collected by the old form.',
    'Dog breed not collected by the old form.',
    'Phone numbers retained as raw + best-effort E.164 (+64). Verify before dialling.',
    "Yes/No fields parsed to booleans; 'Select' and blanks become null.",
    "'Reactive to other animals' values like Cats/Birds/Other map to { reactive: true, detail }.",
  ],
  records,
};

writeFileSync(OUTPUT, JSON.stringify(payload, null, 2) + '\n', 'utf8');

// Print a small summary so the operator can sanity-check.
const summary = {
  records: records.length,
  withOwnerPhone: records.filter((r) => r.owner.phone).length,
  withVetPhone: records.filter((r) => r.vet.phone).length,
  withEmergencyPhone: records.filter((r) => r.emergencyContact.phone).length,
  withSignature: records.filter((r) => r.declaration.signatureUrl).length,
  vaccinatedTrue: records.filter((r) => r.health.vaccinated === true).length,
  councilRegistered: records.filter((r) => r.behaviour.councilRegistered === true).length,
  output: OUTPUT,
};
console.log(JSON.stringify(summary, null, 2));