FEAT: add Platform API proxy with Elasticsearch and DynamoDB sync

- Add Platform API module to proxy requests to Platform API service
- Add Elasticsearch service for pipeline catalog sync (CRUD operations)
- Add DynamoDB service for storing pipeline tables/inputs configuration
- Sync pipeline creation/update/delete to Elasticsearch index
- Extract tables from jobs with connector-specific mappings:
  - JDBC: table_name, load_type, column_include_list, incremental_column_name
  - Singer: replication_method -> full_load/incremental
  - S3: same as Singer
- Map connector types to DynamoDB types (jdbc->database, singer->application, s3->file)
- Validate connector type is provided in job input
- Normalize pipeline IDs for Platform API (replace - with _), keep UUIDs for ES

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
Rafael
2025-12-10 12:42:18 -03:00
co-authored by Claude Opus 4.5
parent e789076ed4
commit df674dd441
15 changed files with 3277 additions and 295 deletions
@@ -0,0 +1,663 @@
import {
Controller,
Get,
Post,
Put,
Patch,
Delete,
Param,
Body,
Query,
Inject,
BadRequestException,
} from '@nestjs/common';
import { ApiTags, ApiOperation } from '@nestjs/swagger';
import { DadosferaLogger } from '@dadosfera/dadosfera-logs';
import {
Authenticated,
RequireAllPermissions,
} from '../../decorators/authentication.decorator';
import { User, RequestUser } from '../../decorators/user.decorator';
import { PlatformApiService } from './platform-api.service';
import { PERMISSIONS_GROUPS } from '../../authentication/permissions.enum';
import { ElasticsearchService } from '../../services/elasticsearch';
import { DynamoDBService } from '../../services/dynamodb';
@ApiTags('Platform API')
@Controller('platform')
export class PlatformApiController {
private logger: any;
constructor(
private readonly platformApiService: PlatformApiService,
private readonly elasticsearchService: ElasticsearchService,
private readonly dynamoDBService: DynamoDBService,
@Inject(DadosferaLogger) dadosferaLogger: DadosferaLogger,
) {
this.logger = dadosferaLogger.logger;
}
/**
* Normalize pipeline ID to match Platform-API format.
* Platform-API replaces '-' with '_' in pipeline IDs.
*/
private normalizePipelineId(id: string): string {
return id?.replace(/-/g, '_') || '';
}
private readonly VALID_CONNECTORS = ['jdbc', 'singer', 's3'];
/**
* Validate that connector is provided and is a valid type.
*/
private validateConnector(connector: string | undefined): void {
if (!connector || !this.VALID_CONNECTORS.includes(connector)) {
throw new BadRequestException(
`connector is required in job input and must be one of: ${this.VALID_CONNECTORS.join(', ')}`,
);
}
}
/**
* Map connector type to DynamoDB type.
* jdbc -> 'database', singer -> 'application', s3 -> 'file'
*/
private mapConnectorToDynamoType(connector: string): string {
switch (connector) {
case 'jdbc':
return 'database';
case 'singer':
return 'application';
case 's3':
return 'file';
default:
return connector;
}
}
/**
* Extract and transform tables from jobs for DynamoDB input.
* Maps connector-specific fields to a common table format.
*
* - JDBC: load_type, table_name, column_include_list (columns), incremental_column_name (reference_column)
* - Singer: type maps replication_method (FULL_TABLE -> full_load, INCREMENTAL -> incremental), no columns
* - S3: same mapping as Singer, no columns
*/
private extractTablesFromJobs(jobs: any[], connector: string): Array<{
name: string;
type: string;
columns?: string[];
reference_column?: { name: string; type: string };
}> {
if (!jobs || jobs.length === 0) return [];
const tables: Array<{
name: string;
type: string;
columns?: string[];
reference_column?: { name: string; type: string };
}> = [];
for (const job of jobs) {
const input = job.input;
if (!input) continue;
if (connector === 'jdbc') {
// JDBC: table_name, load_type, column_include_list, incremental_column_name
const table: {
name: string;
type: string;
columns?: string[];
reference_column?: { name: string; type: string };
} = {
name: input.table_name || '',
type: input.load_type || 'full_load',
};
if (input.column_include_list && input.column_include_list.length > 0) {
table.columns = input.column_include_list;
}
if (input.incremental_column_name) {
table.reference_column = {
name: input.incremental_column_name,
type: input.incremental_column_type || 'timestamp',
};
}
tables.push(table);
} else if (connector === 'singer' || connector === 's3') {
// Singer/S3: replication_method -> type mapping, no columns
let type = 'full_load';
if (input.replication_method === 'INCREMENTAL') {
type = 'incremental';
} else if (input.replication_method === 'FULL_TABLE') {
type = 'full_load';
}
tables.push({
name: input.table_name || '',
type,
});
}
}
return tables;
}
/**
* Build properties object for Elasticsearch based on connector type.
* Different connectors have different property structures.
*
* Note: In pi-factory flow, properties come pre-built from frontend.
* In Maestro proxy flow, we reconstruct from job input fields.
*/
private buildPipelineProperties(jobInput: any): Record<string, any> {
if (!jobInput) return {};
const connector = jobInput.connector;
const properties: Record<string, any> = {};
// Determine credentials_type
if (jobInput.auth_parameters?.credentials_type) {
properties.credentials_type = jobInput.auth_parameters.credentials_type;
} else {
// Default based on connector type
// S3 connector typically uses iam_user, others use basic_auth
properties.credentials_type = connector === 's3' ? 'iam_user' : 'basic_auth';
}
if (connector === 'jdbc') {
// JDBC connectors: schema comes from table_schema
if (jobInput.table_schema) {
properties.schema = jobInput.table_schema;
}
} else if (connector === 'singer') {
// Singer connectors: merge config fields (dates, selected_competitions, etc.)
if (jobInput.config) {
Object.assign(properties, jobInput.config);
}
} else if (connector === 's3') {
// S3 connector
if (jobInput.engine) properties.engine = jobInput.engine;
if (jobInput.source_bucket) properties.source_bucket = jobInput.source_bucket;
if (jobInput.source_prefix) properties.source_prefix = jobInput.source_prefix;
if (jobInput.file_format_params) properties.file_format_params = jobInput.file_format_params;
}
return properties;
}
// ==================== PIPELINE ROUTES ====================
@Post('pipeline')
@ApiOperation({ summary: 'Create a new pipeline' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.CREATE)
async createPipeline(@Body() body: any, @User() user: RequestUser) {
// Validate that pipeline has jobs
if (!body.jobs || body.jobs.length === 0) {
throw new BadRequestException('Pipeline must have at least one job');
}
// Validate connector before proceeding
const firstJob = body.jobs[0]?.input;
this.validateConnector(firstJob?.connector);
// Inject customer_id (actually customer_name) and normalized ID into body for Platform-API
// Note: Platform-API was created before customer_id existed, so it expects customer_name in the customer_id field
const enrichedBody = {
...body,
id: this.normalizePipelineId(body.id),
customer_id: user.customer_name,
};
const result = await this.platformApiService.proxy('POST', '/pipeline', user, enrichedBody);
// Sync to Elasticsearch and DynamoDB
try {
const plugin = firstJob?.plugin;
const connectorType = firstJob?.connector;
const connectionId = firstJob?.auth_parameters?.config_id;
// Build properties based on connector type
const properties = this.buildPipelineProperties(firstJob);
const connector = plugin
? await this.elasticsearchService.getConnectorByPlugin(plugin)
: null;
// Extract tables from jobs and create DynamoDB input
const tables = this.extractTablesFromJobs(body.jobs, connectorType);
let inputId: string | undefined;
if (tables.length > 0) {
const inputDoc = await this.dynamoDBService.createInput(
user.customer_id,
user.user_id,
{
name: body.name,
description: body.description,
plugin: plugin || '',
type: this.mapConnectorToDynamoType(connectorType),
tables,
},
);
inputId = inputDoc.id;
this.logger.info('Created DynamoDB input for tables config', {
inputId,
pipelineId: body.id,
tablesCount: tables.length,
});
}
this.logger.info('Syncing pipeline to Elasticsearch', {
customerName: user.customer_name,
pipelineId: body.id,
plugin,
connector: connectorType,
properties,
inputId,
});
// Keep original UUID format for Elasticsearch (not normalized)
await this.elasticsearchService.createPipeline(
user.customer_name,
body.id,
{
name: body.name,
description: body.description,
user_id: user.user_id,
username: user.username,
customer_id: user.customer_id,
plugin,
connection_id: connectionId,
cron: body.cron,
tables: inputId,
properties,
},
connector,
);
} catch (error) {
this.logger.error('Failed to sync pipeline creation to Elasticsearch/DynamoDB', {
pipelineId: body.id,
customerName: user.customer_name,
error: error.message,
errorName: error.name,
});
}
return result;
}
@Get('pipelines')
@ApiOperation({ summary: 'List all pipelines for customer' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.GET)
async getPipelines(
@User() user: RequestUser,
@Query() query: Record<string, string>,
) {
return this.platformApiService.proxy(
'GET',
'/pipelines',
user,
undefined,
query,
);
}
@Get('pipeline/:pipelineId')
@ApiOperation({ summary: 'Get pipeline by ID' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.GET)
async getPipeline(
@Param('pipelineId') pipelineId: string,
@User() user: RequestUser,
) {
const normalizedId = this.normalizePipelineId(pipelineId);
return this.platformApiService.proxy('GET', `/pipeline/${normalizedId}`, user);
}
@Patch('pipeline/:pipelineId')
@ApiOperation({ summary: 'Update pipeline by ID' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async updatePipeline(
@Param('pipelineId') pipelineId: string,
@Body() body: any,
@User() user: RequestUser,
) {
const normalizedId = this.normalizePipelineId(pipelineId);
const result = await this.platformApiService.proxy(
'PATCH',
`/pipeline/${normalizedId}`,
user,
body,
);
// Sync to Elasticsearch (use original UUID, not normalized)
try {
await this.elasticsearchService.updatePipeline(
user.customer_name,
pipelineId,
{
name: body.name,
description: body.description,
cron: body.cron,
status: body.status,
},
);
} catch (error) {
this.logger.error('Failed to sync pipeline update to Elasticsearch', {
pipelineId,
error: error.message,
});
}
return result;
}
@Delete('pipeline/:pipelineId')
@ApiOperation({ summary: 'Delete pipeline by ID' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.DELETE)
async deletePipeline(
@Param('pipelineId') pipelineId: string,
@User() user: RequestUser,
) {
const normalizedId = this.normalizePipelineId(pipelineId);
const result = await this.platformApiService.proxy(
'DELETE',
`/pipeline/${normalizedId}`,
user,
);
// Sync to Elasticsearch (use original UUID, not normalized)
try {
await this.elasticsearchService.deletePipeline(
user.customer_name,
pipelineId,
);
} catch (error) {
this.logger.error('Failed to sync pipeline deletion to Elasticsearch', {
pipelineId,
error: error.message,
});
}
return result;
}
@Post('pipeline/execute')
@ApiOperation({ summary: 'Execute a pipeline' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async executePipeline(@Body() body: any, @User() user: RequestUser) {
return this.platformApiService.proxy('POST', '/pipeline/execute', user, body);
}
@Post('pipeline/pause')
@ApiOperation({ summary: 'Pause a pipeline' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async pausePipeline(@Body() body: any, @User() user: RequestUser) {
return this.platformApiService.proxy('POST', '/pipeline/pause', user, body);
}
@Post('pipeline/unpause')
@ApiOperation({ summary: 'Unpause a pipeline' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async unpausePipeline(@Body() body: any, @User() user: RequestUser) {
return this.platformApiService.proxy('POST', '/pipeline/unpause', user, body);
}
@Put('pipeline/:pipelineId/memory')
@ApiOperation({ summary: 'Update pipeline memory configuration' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async updatePipelineMemory(
@Param('pipelineId') pipelineId: string,
@Body() body: any,
@User() user: RequestUser,
) {
return this.platformApiService.proxy(
'PUT',
`/pipeline/${pipelineId}/memory`,
user,
body,
);
}
// ==================== PIPELINE METADATA ROUTES ====================
@Put('pipeline/:pipelineId/metadata')
@ApiOperation({ summary: 'Update pipeline metadata' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async updatePipelineMetadata(
@Param('pipelineId') pipelineId: string,
@Body() body: any,
@User() user: RequestUser,
) {
return this.platformApiService.proxy(
'PUT',
`/pipeline/${pipelineId}/metadata`,
user,
body,
);
}
@Get('pipelines/metadata')
@ApiOperation({ summary: 'Get all pipelines metadata' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.GET)
async getPipelinesMetadata(
@User() user: RequestUser,
@Query() query: Record<string, string>,
) {
return this.platformApiService.proxy(
'GET',
'/pipelines/metadata',
user,
undefined,
query,
);
}
// ==================== PIPELINE RUN ROUTES ====================
@Post('pipeline/pipeline_run')
@ApiOperation({ summary: 'Create a pipeline run' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async createPipelineRun(@Body() body: any, @User() user: RequestUser) {
return this.platformApiService.proxy(
'POST',
'/pipeline/pipeline_run',
user,
body,
);
}
@Get('pipeline/:pipelineId/pipeline_run')
@ApiOperation({ summary: 'Get pipeline runs for a pipeline' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.GET)
async getPipelineRuns(
@Param('pipelineId') pipelineId: string,
@User() user: RequestUser,
@Query() query: Record<string, string>,
) {
return this.platformApiService.proxy(
'GET',
`/pipeline/${pipelineId}/pipeline_run`,
user,
undefined,
query,
);
}
@Get('pipeline/:pipelineId/pipeline_run/:runId')
@ApiOperation({ summary: 'Get specific pipeline run' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.GET)
async getPipelineRun(
@Param('pipelineId') pipelineId: string,
@Param('runId') runId: string,
@User() user: RequestUser,
) {
return this.platformApiService.proxy(
'GET',
`/pipeline/${pipelineId}/pipeline_run/${runId}`,
user,
);
}
@Get('pipeline/pipeline_run/:runId/logs')
@ApiOperation({ summary: 'Get pipeline run logs' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.GET)
async getPipelineRunLogs(
@Param('runId') runId: string,
@User() user: RequestUser,
@Query() query: Record<string, string>,
) {
return this.platformApiService.proxy(
'GET',
`/pipeline/pipeline_run/${runId}/logs`,
user,
undefined,
query,
);
}
// ==================== JOBS - COLUMN EDITING ROUTES ====================
@Put('jobs/:jobId/input')
@ApiOperation({ summary: 'Update job input columns' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async updateJobInput(
@Param('jobId') jobId: string,
@Body() body: any,
@User() user: RequestUser,
) {
return this.platformApiService.proxy(
'PUT',
`/jobs/${jobId}/input`,
user,
body,
);
}
@Patch('jobs/:jobId/input')
@ApiOperation({ summary: 'Partial update job input columns' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async patchJobInput(
@Param('jobId') jobId: string,
@Body() body: any,
@User() user: RequestUser,
) {
return this.platformApiService.proxy(
'PATCH',
`/jobs/${jobId}/input`,
user,
body,
);
}
@Put('jobs/:jobId/memory')
@ApiOperation({ summary: 'Update job memory configuration' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async updateJobMemory(
@Param('jobId') jobId: string,
@Body() body: any,
@User() user: RequestUser,
) {
return this.platformApiService.proxy(
'PUT',
`/jobs/${jobId}/memory`,
user,
body,
);
}
@Post('jobs/:jobId/reset-state')
@ApiOperation({ summary: 'Reset job state' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async resetJobState(
@Param('jobId') jobId: string,
@Body() body: any,
@User() user: RequestUser,
) {
return this.platformApiService.proxy(
'POST',
`/jobs/${jobId}/reset-state`,
user,
body,
);
}
// ==================== JOBS - JDBC SYNC MODE ROUTES ====================
@Get('jobs/jdbc/:jobId')
@ApiOperation({ summary: 'Get JDBC job details' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.GET)
async getJdbcJob(@Param('jobId') jobId: string, @User() user: RequestUser) {
return this.platformApiService.proxy('GET', `/jobs/jdbc/${jobId}`, user);
}
@Put('jobs/jdbc/:jobId/sync-mode')
@ApiOperation({ summary: 'Update JDBC job sync mode' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async updateJdbcSyncMode(
@Param('jobId') jobId: string,
@Body() body: any,
@User() user: RequestUser,
) {
return this.platformApiService.proxy(
'PUT',
`/jobs/jdbc/${jobId}/sync-mode`,
user,
body,
);
}
@Get('jobs/jdbc/configs/allowed_datatypes')
@ApiOperation({ summary: 'Get allowed datatypes for JDBC' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.GET)
async getJdbcAllowedDatatypes(@User() user: RequestUser) {
return this.platformApiService.proxy(
'GET',
'/jobs/jdbc/configs/allowed_datatypes',
user,
);
}
// ==================== JOBS - SINGER REPLICATION ROUTES ====================
@Get('jobs/singer/:jobId')
@ApiOperation({ summary: 'Get Singer job details' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.GET)
async getSingerJob(@Param('jobId') jobId: string, @User() user: RequestUser) {
return this.platformApiService.proxy('GET', `/jobs/singer/${jobId}`, user);
}
@Put('jobs/singer/:jobId/sync-mode')
@ApiOperation({ summary: 'Update Singer job sync mode' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.UPDATE)
async updateSingerSyncMode(
@Param('jobId') jobId: string,
@Body() body: any,
@User() user: RequestUser,
) {
return this.platformApiService.proxy(
'PUT',
`/jobs/singer/${jobId}/sync-mode`,
user,
body,
);
}
// ==================== JOBS - S3 ROUTES ====================
@Get('jobs/s3/:jobId')
@ApiOperation({ summary: 'Get S3 job details' })
@RequireAllPermissions(PERMISSIONS_GROUPS.PIPELINE.permissions.GET)
async getS3Job(@Param('jobId') jobId: string, @User() user: RequestUser) {
return this.platformApiService.proxy('GET', `/jobs/s3/${jobId}`, user);
}
// ==================== HEALTH ROUTE ====================
@Get('health')
@ApiOperation({ summary: 'Platform API health check' })
@Authenticated()
async healthCheck(@User() user: RequestUser) {
return this.platformApiService.proxy('GET', '/health', user);
}
}