> ## Documentation Index
> Fetch the complete documentation index at: https://docs.loqate.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Verify Helm Configuration

> Configure country-specific deployments, certified datasets, and resource allocation

## Docker Image Versions

Specify image tags for Verify components.

**Helmfile (Unix/Linux):**

```bash theme={null}
export LOQATE_INSTALLMANAGER_VERSION="image-tag"
export LOQATE_QUERY_COORDINATOR_VERSION="image-tag"
export LOQATE_SPATIAL_API_VERSION="image-tag"
```

**Helmfile (Windows):**

```powershell theme={null}
$env:LOQATE_INSTALLMANAGER_VERSION="image-tag"
$env:LOQATE_QUERY_COORDINATOR_VERSION="image-tag"
$env:LOQATE_SPATIAL_API_VERSION="image-tag"
```

**Helm:**

```bash theme={null}
helm install component loqate/component --set image.tag=image-tag
```

Contact your Loqate representative for available image tags.

## Country-Specific Deployments

Deploy dedicated Spatial-API instances for high-traffic countries to optimize performance and enable independent scaling.

### Why Use Country-Specific Deployments

* **Better performance:** Dedicated resources for frequently queried countries
* **Flexible scaling:** Scale each country independently based on traffic
* **Cost optimization:** Allocate resources where needed

### Configuration

Set `verify.dataset` to the ISO 3166-2 country code. Keep a `row` (rest of world) deployment for other countries.

**Example: United Kingdom deployment**

```bash theme={null}
helm install -n loqate sa-gb loqate/spatial-api \
  --set imageCredentials.username=docker-user \
  --set imageCredentials.password=docker-pass \
  --set app.memberlistService=memberlist.loqate.svc \
  --set verify.dataset=gb
```

**Example with custom storage path:**

```bash theme={null}
helm install -n loqate sa-gb loqate/spatial-api \
  --set imageCredentials.username=docker-user \
  --set imageCredentials.password=docker-pass \
  --set storage.path=/custom/data/path \
  --set app.memberlistService=memberlist.loqate.svc \
  --set verify.dataset=gb
```

**Common Country Codes:**

| Country                   | Code  |
| ------------------------- | ----- |
| United States             | `us`  |
| United Kingdom            | `gb`  |
| Canada                    | `ca`  |
| Australia                 | `au`  |
| Germany                   | `de`  |
| France                    | `fr`  |
| Rest of World (catch-all) | `row` |

## Certified Datasets

Verify supports postal authority-certified validation:

* **CASS** - United States (Coding Accuracy Support System)
* **SERP** - Canada (Software Evaluation and Recognition Program)
* **AMAS** - Australia (Address Matching Approval System)

### Requirements

* License key with certified features enabled
* Certified libraries (downloaded automatically to `lib64` folder)
* Spatial-API configured for the certified country

<Warning>
  **CASS Restriction:** CASS data cannot be downloaded from locations outside the United States due to legal requirements.
</Warning>

### Example: AMAS (Australia)

```bash theme={null}
helm install -n loqate spatial-api-au loqate/spatial-api \
  --set imageCredentials.username=docker-user \
  --set imageCredentials.password=docker-pass \
  --set app.memberlistService=memberlist.loqate.svc \
  --set verify.dataset=au
```

The certified engine activates automatically when licensed libraries are present.

## Premium Datasets

Premium datasets provide enhanced data quality for:

* **France**
* **Ireland**
* **United States**

### Deployment Options

**Option 1: Dedicated deployments (recommended for high volume)**

```bash theme={null}
# France
helm install -n loqate sa-fr loqate/spatial-api \
  --set verify.dataset=fr \
  # ... other settings

# Ireland
helm install -n loqate sa-ie loqate/spatial-api \
  --set verify.dataset=ie \
  # ... other settings
```

**Option 2: Single rest-of-world deployment**

Set `verify.dataset=row` to handle all premium countries with one deployment. Simpler but less flexible scaling.

### Premium US Resources

The premium US dataset requires more resources:

```yaml theme={null}
resources:
  requests:
    memory: "8Gi"
    cpu: "2"
  limits:
    memory: "16Gi"
    cpu: "4"
```

Apply via Helm:

```bash theme={null}
--set resources.requests.memory=8Gi \
--set resources.requests.cpu=2 \
--set resources.limits.memory=16Gi \
--set resources.limits.cpu=4
```

## Resource Allocation

Configure CPU and memory based on expected traffic.

### Recommended Sizing

| Traffic Level               | Memory Request | CPU Request | Memory Limit | CPU Limit |
| --------------------------- | -------------- | ----------- | ------------ | --------- |
| Low (under 100 req/min)     | 2Gi            | 0.5         | 4Gi          | 1         |
| Medium (under 1000 req/min) | 4Gi            | 1           | 8Gi          | 2         |
| High (under 5000 req/min)   | 8Gi            | 2           | 16Gi         | 4         |
| Premium US                  | 8Gi            | 2           | 16Gi         | 4         |

### Example Configuration

**Via values file:**

```yaml theme={null}
resources:
  requests:
    memory: "4Gi"
    cpu: "1"
  limits:
    memory: "8Gi"
    cpu: "2"
```

**Via Helm:**

```bash theme={null}
--set resources.requests.memory=4Gi \
--set resources.requests.cpu=1 \
--set resources.limits.memory=8Gi \
--set resources.limits.cpu=2
```

Start conservative and adjust based on Prometheus metrics.

## AI Parser (Optional)

Enable AI-enhanced parsing for improved accuracy with unstructured addresses.

### Requirements

* Separate persistent storage (at least 4Gi)
* Additional resources for TensorFlow Serving and AI Parser components

### Enable AI Parser

**In helmfile, add to QueryCoordinator values:**

```yaml theme={null}
app:
  allow_aiparser: true
```

**Configure AI storage:**

```bash theme={null}
# Unix
export LOQATE_AIPARSER_STORAGE_PATH="/opt/loqate/ai-models"

# Windows
$env:LOQATE_AIPARSER_STORAGE_PATH="C:\loqate\ai-models"
```

**Uncomment AI Parser sections in helmfile and apply:**

```bash theme={null}
helmfile apply  # Unix
helmfile sync   # Windows
```

See the AI Parser section in helmfile.yaml for detailed configuration.

## Next Steps

* [Usage](/helm/usage) - Learn management and monitoring commands
* [Troubleshooting](/helm/troubleshooting) - Resolve common issues
* [Quick Start](/helm/quick-start) - Installation guide