Run Scripts Across All 3 Environments (Local, Dev, Stg)

For Claude: REQUIRED SUB-SKILL: Use superpowers-extended-cc:executing-plans to implement this plan task-by-task.

Goal: Execute all necessary batch scripts against all 3 MongoDB databases (local, dev, stg) to ensure consistent data across environments.

Architecture: Scripts are run via yarn script-dev which reads script/config.json for the script sequence and script/.env for the DB connection. We switch environments by uncommenting the appropriate MONGO_CONNECTION_STRING in script/.env, run the required scripts, then move to the next environment.

Tech Stack: Node.js, Yarn 4 PnP, MongoDB Atlas, Mongoose

---

Environment Connection Strings

All stored in script/.env (toggle by commenting/uncommenting):

Env	Connection String
Local	mongodb+srv://demo:GhKYzU37veEYEmrc@energyaudit.0l6duna.mongodb.net/
Dev	mongodb+srv://dgi-ai-widgets-demo-dev:FEQ8AKdlea4rmkHM@dev.e45p6au.mongodb.net/
Stg	mongodb+srv://dgi-ai-widgets-demo-stg:tiuYWzxKOWnvNXwj@stg.ivmjjjb.mongodb.net/

DB name is dtcea-mumbai-demo for all 3.

---

Scripts Inventory & Execution Order

Scripts have dependencies — they must run in this order:

#	Script	Entity	What it does	Dependencies
1	POST /seed/ai-audit-test-data	seed (API)	Seeds demo TCs, installations, users, AI audit results for section S01	None (standalone seed)
2	transformer_backfillAuditStatus	transformer	Sets auditStatus = AUDITED/AUDIT_FAILED on all TCs based on lossPercentage	Requires TCs to exist (after seeds + migration data)
3	transformer_fixInvalidAuditedTCs	transformer	Fixes TCs that have lossPercentage but missing tcConsumption/installationConsumption	After backfillAuditStatus
4	installation_backfillAvgConsumption	installation	Computes 6-month avg consumption for remarked installations	Requires installations to exist
5	transformer_backfillHealthScore	transformer	Computes healthScore + healthGrade on all TCs	After audit status is finalized
6	auditSummary_computeAuditSummary	auditSummary	Computes section→MESCOM level audit summaries	Must run LAST — depends on final TC audit states

---

Execution Strategy

For each environment, we run all 6 scripts in order. The approach:

1. Update script/.env to point to the target environment
2. Update script/config.json with the script(s) to run
3. Run yarn script-dev
4. Verify output logs
5. Repeat for next script or next environment

Order: Local → Dev → Stg (local first so we can verify before touching shared envs)

---

Task 1: Run all scripts on LOCAL environment

Files:

• Modify: script/.env (ensure Local connection string is active)
• Modify: script/config.json (update scriptRunArray for each script)

Step 1: Verify script/.env points to Local

Ensure the active (uncommented) connection string is:

MONGO_CONNECTION_STRING=mongodb+srv://demo:GhKYzU37veEYEmrc@energyaudit.0l6duna.mongodb.net/

Step 2: Seed via API

POST /dtcea-mumbai-demo/v1/seed/ai-audit-test-data
Body: { "clearExisting": true }

Expected: Response showing seed data inserted for TCs, installations, users, AI audit results

Step 3: Backfill audit status

Update script/config.json:

{
  "scriptRunArray": ["transformer_backfillAuditStatus"],
  "input": {
    "transformer": {
      "backfillAuditStatus": {}
    }
  },
  "printTagsToScriptMapping": false
}

Run: yarn script-dev Expected: Logs with Marked AUDITED: N and Marked AUDIT_FAILED: N

Step 4: Fix invalid audited TCs

Update script/config.json:

{
  "scriptRunArray": ["transformer_fixInvalidAuditedTCs"],
  "input": {
    "transformer": {
      "fixInvalidAuditedTCs": {}
    }
  },
  "printTagsToScriptMapping": false
}

Run: yarn script-dev Expected: Either "Nothing to fix" or logs showing affected TCs corrected

Step 5: Backfill avg consumption

Update script/config.json:

{
  "scriptRunArray": ["installation_backfillAvgConsumption"],
  "input": {
    "installation": {
      "backfillAvgConsumption": {}
    }
  },
  "printTagsToScriptMapping": false
}

Run: yarn script-dev Expected: Per-month logs showing averages computed and merged

Step 6: Backfill health scores

Update script/config.json:

{
  "scriptRunArray": ["transformer_backfillHealthScore"],
  "input": {
    "transformer": {
      "backfillHealthScore": {}
    }
  },
  "printTagsToScriptMapping": false
}

Run: yarn script-dev Expected: Per-month logs showing TCs updated with healthScore + healthGrade

Step 7: Compute audit summaries (MUST BE LAST)

Update script/config.json:

{
  "scriptRunArray": ["auditSummary_computeAuditSummary"],
  "input": {
    "auditSummary": {
      "computeAuditSummary": {}
    }
  },
  "printTagsToScriptMapping": false
}

Run: yarn script-dev Expected: Per-month SECTION/SUB_DIVISION/DIVISION/CIRCLE/ZONE/MESCOM summaries inserted, invariant checks all "OK"

Step 8: Commit checkpoint

git add script/config.json
git commit -m "chore: run all scripts on local environment"

---

Task 2: Run all scripts on DEV environment

Files:

• Modify: script/.env (switch to Dev connection string)

Step 1: Switch script/.env to Dev

# Local
# MONGO_CONNECTION_STRING=mongodb+srv://demo:GhKYzU37veEYEmrc@energyaudit.0l6duna.mongodb.net/
# Dev
MONGO_CONNECTION_STRING=mongodb+srv://dgi-ai-widgets-demo-dev:FEQ8AKdlea4rmkHM@dev.e45p6au.mongodb.net/
# Stg
# MONGO_CONNECTION_STRING=mongodb+srv://dgi-ai-widgets-demo-stg:tiuYWzxKOWnvNXwj@stg.ivmjjjb.mongodb.net/

Steps 2-8: Run all 7 scripts in the same order as Task 1

Same config.json updates and yarn script-dev commands. Same expected outputs.

Step 9: Verify Dev data

Quick sanity check — connect to Dev DB and confirm:

• transformers collection has docs with auditStatus set
• audit-summaries collection has SECTION and MESCOM level docs
• installations collection has avgConsumption6Months on remarked docs

---

Task 3: Run all scripts on STG environment

Files:

• Modify: script/.env (switch to Stg connection string)

Step 1: Switch script/.env to Stg

# Local
# MONGO_CONNECTION_STRING=mongodb+srv://demo:GhKYzU37veEYEmrc@energyaudit.0l6duna.mongodb.net/
# Dev
# MONGO_CONNECTION_STRING=mongodb+srv://dgi-ai-widgets-demo-dev:FEQ8AKdlea4rmkHM@dev.e45p6au.mongodb.net/
# Stg
MONGO_CONNECTION_STRING=mongodb+srv://dgi-ai-widgets-demo-stg:tiuYWzxKOWnvNXwj@stg.ivmjjjb.mongodb.net/

Steps 2-8: Run all 7 scripts in the same order as Task 1

Same config.json updates and yarn script-dev commands. Same expected outputs.

Step 9: Verify Stg data

Same sanity checks as Dev.

---

Task 4: Restore Local config and clean up

Step 1: Switch script/.env back to Local

# Local
MONGO_CONNECTION_STRING=mongodb+srv://demo:GhKYzU37veEYEmrc@energyaudit.0l6duna.mongodb.net/
# Dev
# MONGO_CONNECTION_STRING=mongodb+srv://dgi-ai-widgets-demo-dev:FEQ8AKdlea4rmkHM@dev.e45p6au.mongodb.net/
# Stg
# MONGO_CONNECTION_STRING=mongodb+srv://dgi-ai-widgets-demo-stg:tiuYWzxKOWnvNXwj@stg.ivmjjjb.mongodb.net/

Step 2: Reset script/config.json to default

{
  "scriptRunArray": ["auditSummary_computeAuditSummary"],
  "input": {
    "auditSummary": {
      "computeAuditSummary": {}
    }
  },
  "printTagsToScriptMapping": false
}

Step 3: Verify local API still works

Run: yarn api-dev Expected: API starts successfully, connects to local MongoDB

---

Important Notes

• computeAuditSummary DELETES all existing audit summaries before recomputing — always run it last
• POST /seed/ai-audit-test-data is idempotent — pass clearExisting: true to clean up before inserting
• backfillAuditStatus only touches TCs with auditStatus: null — safe to re-run
• fixInvalidAuditedTCs is idempotent — only fixes TCs matching the bad state
• backfillAvgConsumption creates a temp collection and drops it after — safe to re-run
• backfillHealthScore overwrites healthScore/healthGrade — safe to re-run

Estimated Scale

• ~150K TCs × 7 months = ~1M transformer docs
• ~3.3M installation docs
• backfillAvgConsumption is the heaviest (aggregation + $merge on millions of docs) — expect 5-15 min per environment
• backfillHealthScore processes ~1M docs in batches of 500 — expect 3-8 min per environment
• computeAuditSummary aggregates ~1M TCs — expect 2-5 min per environment