[DCP - Terraform] Initializes the APIs, Service Account, Cloud Run Service and Spanner Instance and DB.

.gitignore

@@ -9,4 +9,29 @@ venv
 node_modules
 .ruff_cache
 .coverage
-htmlcov/
+htmlcov/
+
+# Terraform
+.terraform/
+*.tfstate
+*.tfstate.*
+crash.log
+crash.log.*
+*.tfvars
+*.tfvars.json
+override.tf
+override.tf.json
+_override.tf
+_override.tf.json
+.terraformrc
+terraform.rc
+tfplan
+infra/dcp/backend.tf
+.terraform.lock.hcl
+
+# Configuration
+infra/dcp/.env
+
+*.pyc
+.venv/
+uv.lock

infra/dcp/README.md

@@ -0,0 +1,64 @@
+# Data Commons Platform (DCP) Infrastructure
+
+This directory contains the Terraform configuration to deploy the Data Commons Platform on Google Cloud Platform (GCP).
+
+## Prerequisites
+*   **GCP Project**: A GCP project with billing enabled.
+*   **Terraform**: Terraform installed locally (>= 1.0.0).
+*   **gcloud CLI**: GCP CLI installed and authenticated.
+
+## Setup
+
+1.  **Configure Local Variables**:
+    Copy the example variable file and fill in your project details.
+    ```bash
+    cp terraform.tfvars.example terraform.tfvars
+    ```
+    Edit `terraform.tfvars` with your `project_id` and other preferred settings.
+
+2.  **Run Setup Script**:
+    The `setup.sh` script creates a GCS bucket for your Terraform state and initializes the backend.
+    ```bash
+    ./setup.sh
+    ```
+
+## Deployment
+
+1.  **Initialize**:
+    Initialize Terraform (if not already done by setup.sh).
+    ```bash
+    terraform init
+    ```
+
+2.  **Plan**:
+    Review the changes Terraform will make.
+    ```bash
+    terraform plan
+    ```
+
+3.  **Apply**:
+    Provision the infrastructure.
+    ```bash
+    terraform apply
+    ```
+
+4.  **Teardown**:
+    Destroy all resources.
+    ```bash
+    terraform destroy
+    ```
+
+## Architecture
+
+This setup uses an **Orchestrator Pattern**:
+- `infra/dcp/main.tf`: The root entrypoint that calls modules.
+- `infra/dcp/modules/dcp/`: The new Data Commons Platform stack (Cloud Run + Spanner).
+- `infra/dcp/modules/cdc/`: The legacy Custom Data Commons stack (Cloud Run + MySQL + Redis).
+
+Each module is independent and can be toggled via the root variables in `terraform.tfvars`.
+
+## Troubleshooting
+*   **Deletion Errors**: If you get a "cannot destroy... deletion_protection" error, ensure `deletion_protection = false` in your `terraform.tfvars`, run `terraform apply`, and then try `terraform destroy` again. Alternatively, use the helper command:
+    ```bash
+    make force-destroy
+    ```

infra/dcp/main.tf

@@ -0,0 +1,123 @@
+terraform {
+  required_providers {
+    google = {
+      source  = "hashicorp/google"
+      version = ">= 5.0"
+    }
+    null = {
+      source  = "hashicorp/null"
+      version = ">= 3.0"
+    }
+  }
+}
+
+provider "google" {
+  project = var.project_id
+  region  = var.region
+}
+
+# Enable required APIs for both stacks
+resource "google_project_service" "apis" {
+  for_each = toset(concat([
+    "run.googleapis.com",
+    "iam.googleapis.com",
+    "sqladmin.googleapis.com",
+    "redis.googleapis.com",
+    "secretmanager.googleapis.com",
+    "vpcaccess.googleapis.com",
+    "artifactregistry.googleapis.com",
+    "compute.googleapis.com"
+  ], var.enable_dcp ? ["spanner.googleapis.com"] : []))
+
+  service            = each.key
+  disable_on_destroy = false
+}
+
+# --- Data Commons Platform (DCP) Stack ---
+module "dcp" {
+  source = "./modules/dcp"
+  count  = var.enable_dcp ? 1 : 0
+
+  project_id               = var.project_id
+  namespace                = var.namespace
+  region                   = var.region
+  image_url                = var.dcp_image_url
+  service_name             = var.dcp_service_name
+  service_account_name     = var.dcp_service_account_name
+  create_spanner_instance  = var.dcp_create_spanner_instance
+  create_spanner_db        = var.dcp_create_spanner_db
+  spanner_instance_id      = var.dcp_spanner_instance_id
+  spanner_database_id      = var.dcp_spanner_database_id
+  spanner_processing_units = var.dcp_spanner_processing_units
+  service_cpu              = var.dcp_service_cpu
+  service_memory           = var.dcp_service_memory
+  service_min_instances    = var.dcp_service_min_instances
+  service_max_instances    = var.dcp_service_max_instances
+  service_concurrency      = var.dcp_service_concurrency
+  service_timeout_seconds  = var.dcp_service_timeout_seconds
+  deletion_protection      = var.deletion_protection
+
+  depends_on = [google_project_service.apis]
+}
+
+# --- Custom Data Commons (CDC) Legacy Stack ---
+module "cdc" {
+  source = "./modules/cdc"
+  count  = var.enable_cdc ? 1 : 0
+
+  project_id                    = var.project_id
+  namespace                     = var.namespace
+  dc_api_key                    = var.cdc_dc_api_key
+  maps_api_key                  = var.cdc_maps_api_key
+  disable_google_maps           = var.cdc_disable_google_maps
+  region                        = var.region
+  google_analytics_tag_id       = var.cdc_google_analytics_tag_id
+  gcs_data_bucket_name          = var.cdc_gcs_data_bucket_name
+  gcs_data_bucket_input_folder  = var.cdc_gcs_data_bucket_input_folder
+  gcs_data_bucket_output_folder = var.cdc_gcs_data_bucket_output_folder
+  gcs_data_bucket_location      = var.cdc_gcs_data_bucket_location
+  mysql_instance_name           = var.cdc_mysql_instance_name
+  mysql_database_name           = var.cdc_mysql_database_name
+  mysql_database_version        = var.cdc_mysql_database_version
+  mysql_cpu_count               = var.cdc_mysql_cpu_count
+  mysql_memory_size_mb          = var.cdc_mysql_memory_size_mb
+  mysql_storage_size_gb         = var.cdc_mysql_storage_size_gb
+  mysql_user                    = var.cdc_mysql_user
+  mysql_deletion_protection     = var.deletion_protection
+  dc_web_service_image          = var.cdc_web_service_image
+  dc_web_service_min_instance_count = var.cdc_web_service_min_instance_count
+  dc_web_service_max_instance_count = var.cdc_web_service_max_instance_count
+  dc_web_service_cpu            = var.cdc_web_service_cpu
+  dc_web_service_memory         = var.cdc_web_service_memory
+  make_dc_web_service_public    = var.cdc_make_dc_web_service_public
+  dc_data_job_image             = var.cdc_data_job_image
+  dc_data_job_cpu               = var.cdc_data_job_cpu
+  dc_data_job_memory            = var.cdc_data_job_memory
+  dc_data_job_timeout           = var.cdc_data_job_timeout
+  dc_search_scope               = var.cdc_search_scope
+  enable_mcp                   = var.cdc_enable_mcp
+  vpc_network_name              = var.cdc_vpc_network_name
+  vpc_network_subnet_name       = var.cdc_vpc_network_subnet_name
+  enable_redis                  = var.cdc_enable_redis
+  redis_instance_name           = var.cdc_redis_instance_name
+  redis_memory_size_gb          = var.cdc_redis_memory_size_gb
+  redis_tier                    = var.cdc_redis_tier
+  redis_location_id             = var.cdc_redis_location_id
+  redis_alternative_location_id = var.cdc_redis_alternative_location_id
+  redis_replica_count           = var.cdc_redis_replica_count
+  vpc_connector_cidr            = var.cdc_vpc_connector_cidr
+  use_spanner                   = var.enable_dcp
+  spanner_instance_id           = var.enable_dcp ? module.dcp[0].spanner_instance_id : ""
+  spanner_database_id           = var.enable_dcp ? module.dcp[0].spanner_database_id : ""
+  deletion_protection           = var.deletion_protection
+
+  depends_on = [google_project_service.apis]
+}
+
+# Ensure Spanner instance ID is provided when not creating a new one
+check "spanner_instance_id_provided" {
+  assert {
+    condition     = !var.enable_dcp || var.dcp_create_spanner_instance || var.dcp_spanner_instance_id != ""
+    error_message = "dcp_spanner_instance_id must be provided when reusing an existing instance (dcp_create_spanner_instance = false)."
+  }
+}

infra/dcp/modules/cdc/locals.tf

@@ -0,0 +1,109 @@
+# Local variable definitions
+
+locals {
+  # Data Commons Data Bucket
+  name_prefix = var.namespace != "" ? "${var.namespace}-" : ""
+  gcs_data_bucket_name = var.gcs_data_bucket_name != "" ? var.gcs_data_bucket_name : "${local.name_prefix}datacommons-data-${var.project_id}"
+
+  # Use var.maps_api_key if set, otherwise use generated Maps API key
+  maps_api_key = var.maps_api_key != null ? var.maps_api_key : google_apikeys_key.maps_api_key[0].key_string
+
+  # Data Commons API hostname
+  dc_api_hostname = "api.datacommons.org"
+
+  # Data Commons API protocol
+  dc_api_protocol = "https"
+
+  # Data Commons API root URL
+  dc_api_root = "${local.dc_api_protocol}://${local.dc_api_hostname}"
+
+  # Optionally-configured Redis instance
+  redis_instance = var.enable_redis ? google_redis_instance.redis_instance[0] : null
+
+
+  # Shared environment variables used by the Data Commons web service and the Data
+  # Commons data loading job
+  cloud_run_shared_env_variables = [
+    {
+      name  = "USE_CLOUDSQL"
+      value = var.use_spanner ? "false" : "true"
+    },
+    {
+      name  = "CLOUDSQL_INSTANCE"
+      value = var.use_spanner ? "" : google_sql_database_instance.mysql_instance[0].connection_name
+    },
+    {
+      name  = "DB_NAME"
+      value = var.mysql_database_name
+    },
+    {
+      name  = "DB_USER"
+      value = var.mysql_user
+    },
+    {
+      name  = "DB_HOST"
+      value = ""
+    },
+    {
+      name  = "DB_PORT"
+      value = "3306"
+    },
+    {
+      name  = "OUTPUT_DIR"
+      value = "gs://${local.gcs_data_bucket_name}/${var.gcs_data_bucket_output_folder}"
+    },
+    {
+      name  = "FORCE_RESTART"
+      value = "${timestamp()}"
+    },
+    {
+      name  = "REDIS_HOST"
+      value = try(local.redis_instance.host, "")
+    },
+    {
+      name  = "REDIS_PORT"
+      value = try(local.redis_instance.port, "")
+    },
+    {
+      name  = "GCP_SPANNER_INSTANCE_ID"
+      value = var.spanner_instance_id
+    },
+    {
+      name  = "GCP_SPANNER_DATABASE_NAME"
+      value = var.spanner_database_id
+    }
+  ]
+
+  # Shared environment variables containing secret refs used by the Data Commons
+  # web service and the Data Commons data loading job
+  cloud_run_shared_env_variable_secrets = concat([
+    {
+      name = "DC_API_KEY"
+      value_source = {
+        secret_key_ref = {
+          secret  = google_secret_manager_secret.dc_api_key.secret_id
+          version = "latest"
+        }
+      }
+    },
+    {
+      name = "MAPS_API_KEY"
+      value_source = {
+        secret_key_ref = {
+          secret  = var.disable_google_maps ? "" : google_secret_manager_secret.maps_api_key[0].secret_id
+          version = "latest"
+        }
+      }
+    }
+  ], var.use_spanner ? [] : [
+    {
+      name = "DB_PASS"
+      value_source = {
+        secret_key_ref = {
+          secret  = google_secret_manager_secret.mysql_password_secret[0].id
+          version = "latest"
+        }
+      }
+    }
+  ])
+}