Compare commits

...
Author SHA1 Message Date
Marcos Rodrigues Silva 1175591521 Merge pull request #295 from dadosfera/feature/pii-rules
Feature/pii rules
2025-04-29 14:04:06 -03:00
marcos-silva-rodrigues 04accb6689 FIX: correct stg urls 2025-04-29 14:02:24 -03:00
marcos-silva-rodrigues 58e3c7be94 FEAT: pii 2025-04-29 14:00:23 -03:00
marcos-silva-rodrigues 7a6db36554 FEAT: generate pii reporter 2025-04-27 13:00:49 -03:00
Marcos Rodrigues Silva 9e0495fc26 Merge pull request #293 from dadosfera/feat/enforce-mfa
Feat/enforce mfa
2025-04-22 14:50:39 -03:00
14 changed files with 1138 additions and 37 deletions
+11 -1
View File
@@ -4,9 +4,19 @@ FROM base_image AS build_base
WORKDIR /app
RUN apk update
# needed packages to build dependencies from source
RUN apk add --no-cache aws-cli
RUN apk add --no-cache \
aws-cli \
chromium \
nss \
freetype \
harfbuzz \
ca-certificates \
ttf-freefont
COPY package*.json ./
ENV PUPPETEER_SKIP_CHROMIUM_DOWNLOAD=true \
PUPPETEER_EXECUTABLE_PATH=/usr/bin/chromium-browser
# run aws cli without mounting secret, because CI already has AWS credentials
FROM build_base AS ci_image
+46
View File
@@ -4963,6 +4963,52 @@
]
}
},
"/catalog/pii-reporter": {
"get": {
"operationId": "CatalogController_getPiiReporter",
"parameters": [
{
"name": "dadosfera-lang",
"in": "header",
"required": false,
"schema": {
"enum": [
"pt-br",
"en-us"
],
"type": "string"
}
},
{
"name": "type",
"required": true,
"in": "query",
"schema": {
"type": "string"
}
}
],
"responses": {
"200": {
"description": ""
}
},
"tags": [
"Catalog"
],
"security": [
{
"access-token": []
},
{
"access-token": []
},
{
"access-token": []
}
]
}
},
"/terms-of-use/token": {
"get": {
"operationId": "TermsOfUseController_getToken",
+2 -2
View File
@@ -9,9 +9,9 @@ charts:
- name: hostname
value: maestro.stg.dadosfera.ai
- name: maestro.pi_factory_url
value: pi-factory.dadosfera.ai
value: pi-factory.stg.dadosfera.ai
- name: maestro.in_factory_url
value: in-factory.dadosfera.ai
value: in-factory.stg.dadosfera.ai
- name: maestro.tr_factory_url
value: in-factory.dadosfera.ai
- name: maestro.open_customer_id
+750 -9
View File
File diff suppressed because it is too large Load Diff
+4 -1
View File
@@ -30,7 +30,7 @@
"@aws-sdk/client-secrets-manager": "^3.414.0",
"@dadosfera/dadosfera-logs": "^1.0.0-beta.4",
"@dadosfera/protospack": "2.5.3",
"@dadosfera/protospack-v2": "3.37.0-beta.22",
"@dadosfera/protospack-v2": "3.37.0-beta.23",
"@grpc/grpc-js": "^1.9.3",
"@grpc/proto-loader": "^0.7.9",
"@nestjs/cli": "^9.5.0",
@@ -48,8 +48,10 @@
"class-transformer": "^0.5.1",
"class-validator": "^0.14.0",
"cron-parser": "^4.9.0",
"csv": "^6.3.11",
"dotenv": "^14.3.2",
"elastic-apm-node": "^3.50.0",
"handlebars": "^4.7.8",
"helmet": "^5.1.1",
"jsonwebtoken": "^9.0.2",
"jwk-to-pem": "^2.0.5",
@@ -61,6 +63,7 @@
"passport-google-oauth20": "^2.0.0",
"passport-hubspot-oauth2": "^1.0.3",
"passport-mailchimp": "^1.1.0",
"puppeteer": "^24.7.2",
"reflect-metadata": "^0.1.13",
"rimraf": "^3.0.2",
"rxjs": "^7.5.5",
+2 -1
View File
@@ -613,7 +613,8 @@ export interface DadosferaModule {
export const DADOSFERA_MODULES_KEYS = {
LOG_DASHBOARD: 'logs-dashboard',
ACCESS_DASHBOARD: 'access-dashboard',
DANGER_ZONE: 'danger-zone'
DANGER_ZONE: 'danger-zone',
PII: 'pii'
}
export const DADOSFERA_MODULES: Array<DadosferaModule> = [
+75 -23
View File
@@ -15,6 +15,7 @@ import {
UseFilters,
HttpException,
HttpStatus,
Res,
} from '@nestjs/common';
import {
ApiCreatedResponse,
@@ -25,14 +26,14 @@ import {
import {
Authenticated,
RequireAllPermissions,
RequireModule,
RequireSomePermission,
} from '../../decorators/authentication.decorator';
import { PERMISSIONS_GROUPS } from '../../authentication/permissions.enum';
import { DADOSFERA_MODULES_KEYS, PERMISSIONS_GROUPS } from '../../authentication/permissions.enum';
import { CatalogService } from './catalog.service';
import { DadosferaLogger } from '@dadosfera/dadosfera-logs';
import { PackTheMetadata } from 'src/utils/ PackTheMetadata';
import { RequestUser, User } from 'src/decorators/user.decorator';
import { lastValueFrom } from 'rxjs';
import {
BatchRemoveRlsRulesRequest,
GetDatasetCatalogTaskRes,
@@ -59,6 +60,7 @@ import {
GetRlsRulesRequest,
RegisterDatasetWithMetatadaRequest,
} from '@dadosfera/protospack-v2/dist/lib/Catalog/interfaces/messages';
import { Response } from 'express';
@ApiTags('Catalog')
@ApiHeaders([{ name: 'dadosfera-lang', enum: LanguageEnum, required: false }])
@@ -344,7 +346,7 @@ export class CatalogController {
@Language() language: LanguageEnum,
@Param('id') id: string,
): Promise<IPreviewResponse> {
const { customer_name, customer_id, user_id, username } = user;
const { customer_name, customer_id, user_id, username, customer_modules } = user;
this.logger.info(`/catalog - ON GET DATA DOCS ROUTE`, {
user_id,
@@ -357,6 +359,7 @@ export class CatalogController {
user_id,
username,
language,
is_mask: customer_modules.some(mod => mod === 'pii')
});
const preview = await this.catalogService.getDatasetPreview(id, metadata);
@@ -748,25 +751,23 @@ export class CatalogController {
return JSON.parse(dashboards);
}
@Post('register-dataset')
@RequireSomePermission(
PERMISSIONS_GROUPS.CATALOG.permissions.CREATE,
PERMISSIONS_GROUPS.CATALOG.permissions.DATA_MANAGER
PERMISSIONS_GROUPS.CATALOG.permissions.DATA_MANAGER,
)
async registerDatasetWithMetadataRequest(
@Body() body: RegisterDatasetWithMetatadaRequest,
@User() user: RequestUser,
) {
const { customer_id, customer_name, user_id, username } = user;
const logMetadata = {
customer_name: customer_name,
user_id: user_id,
method: 'POST',
path: '/catalog/register-dataset',
};
const { customer_id, customer_name, user_id, username } = user;
const logMetadata = {
customer_name: customer_name,
user_id: user_id,
method: 'POST',
path: '/catalog/register-dataset',
};
try {
const metadata = PackTheMetadata({
customer_id,
customer_name,
@@ -774,18 +775,20 @@ export class CatalogController {
username,
});
this.logger.log(`Request from user ${user_id} for customer ${customer_name}`, logMetadata);
this.logger.log(
`Request from user ${user_id} for customer ${customer_name}`,
logMetadata,
);
// Create table metadata
const tableMetadataBody = {
table_metadata: body.table_metadata,
info: {
customer: customer_name,
},
logMetadata: logMetadata
logMetadata: logMetadata,
};
const table_metadata_id = await this.catalogService.createTableMetadata(
tableMetadataBody,
);
@@ -798,8 +801,12 @@ export class CatalogController {
customer: customer_name,
},
};
this.logger.info(`Creating column metadata for table ${table_metadata_id}`, logMetadata);
const column_metadata_ids = await this.catalogService.createColumnMetadata(columnMetadataBody)
this.logger.info(
`Creating column metadata for table ${table_metadata_id}`,
logMetadata,
);
const column_metadata_ids =
await this.catalogService.createColumnMetadata(columnMetadataBody);
// Create data preview
const dataPreviewBody = {
@@ -808,15 +815,24 @@ export class CatalogController {
customer: customer_name,
},
};
this.logger.debug(`Creating data preview for table ${table_metadata_id}`, logMetadata);
this.logger.debug(
`Creating data preview for table ${table_metadata_id}`,
logMetadata,
);
const data_preview_id = await this.catalogService.createDataPreview(
dataPreviewBody,
);
// Catalog dataset item
this.logger.info(`Cataloging dataset item for table ${table_metadata_id}`, logMetadata);
this.logger.info(
`Cataloging dataset item for table ${table_metadata_id}`,
logMetadata,
);
await this.catalogService.catalogDatasetItem(table_metadata_id, metadata);
this.logger.info(`Dataset registration completed successfully for table ${table_metadata_id}`, logMetadata);
this.logger.info(
`Dataset registration completed successfully for table ${table_metadata_id}`,
logMetadata,
);
return {
message: 'Dataset registered successfully',
table_metadata_id: table_metadata_id,
@@ -824,7 +840,10 @@ export class CatalogController {
data_preview_id: data_preview_id,
};
} catch (error) {
this.logger.error(`Failed to register dataset. The following error occurred: ${error.response.data}`, logMetadata);
this.logger.error(
`Failed to register dataset. The following error occurred: ${error.response.data}`,
logMetadata,
);
throw new HttpException(
{
@@ -837,4 +856,37 @@ export class CatalogController {
);
}
}
@Get('pii-reporter')
@RequireSomePermission(
PERMISSIONS_GROUPS.USERS.permissions.ADMIN
)
@RequireModule(
DADOSFERA_MODULES_KEYS.PII
)
async getPiiReporter(@User() user: RequestUser, @Res() res: Response, @Query('type') type: "html" | "csv") {
this.logger.info('GET pii-reporter');
const metadata = PackTheMetadata(user);
try {
const reporter = await this.catalogService.getPiiReporter(metadata, type);
if (type === "html") {
res.setHeader('Content-Type', 'text/html');
} else if (type === "csv") {
res.setHeader('Content-Type', 'text/csv');
} else if (type === "pdf") {
res.setHeader('Content-Type', 'application/pdf');
}
const timestamp = new Date().toISOString().replace(/[:.]/g, '-');
const filename = `relatorio-pii-${timestamp}.${type}`;
res.setHeader('Content-Disposition', `attachment; filename="${filename}"`);
return res.send(reporter);
} catch (error) {
this.logger.error(error);
}
}
}
+49
View File
@@ -34,6 +34,12 @@ import {
GetNimbusDashboardsRequest,
GetRlsRulesRequest,
} from '@dadosfera/protospack-v2/dist/lib/Catalog/interfaces/messages';
import { writeFileSync } from 'fs';
import path from 'path';
import { HtmlParser } from 'src/utils/FileParser/html-parser';
import { PiiDto } from './dtos/pii.dto';
import { CsvParser } from 'src/utils/FileParser/csv-parser';
import { PDFParser } from 'src/utils/FileParser/pdf-parser';
class CatalogService implements OnModuleInit {
catalogReadService: ReadService.CatalogReadServices;
@@ -80,6 +86,49 @@ class CatalogService implements OnModuleInit {
)}.dadosfera.ai`;
}
async getPiiReporter(metadata: Metadata, type: "html"| "csv") {
try {
const {
data
} = await lastValueFrom(
this.catalogWriteService.GetPiiReporter({}, metadata)
)
this.logger.info("Finish grpc call")
// 3. Ler o template HTML
const templatePath = path.join(
__dirname,
'../../../src/template/pii-reporter.html',
);
this.logger.info("templatePath: " + templatePath);
if (type === "html") {
const htmlParser = new HtmlParser<PiiDto>(templatePath);
return htmlParser.parse(data);
} else if (type === "csv") {
const csvParser = new CsvParser<PiiDto>();
return await csvParser.parse(data)
} else if (type === "pdf") {
const htmlParser = new HtmlParser<PiiDto>(templatePath);
const pdfParser = new PDFParser();
const pdf = await pdfParser.parse(htmlParser.parse(data));
const pdfPath = path.join(
__dirname,
'../../../src/template/teste.pdf',
);
writeFileSync(pdfPath, pdf);
return pdf
}
throw new Error("Unsupported Type: " + type)
} catch (error) {
this.logger.error(error);
}
}
async createDataAsset(data: Messages.CreateDataAssetRequest, metadata) {
this.logger.info('CatalogService - Manage Data assets permissions');
if (!data.embed) data.embed = undefined;
+8
View File
@@ -0,0 +1,8 @@
export class PiiDto {
database_name: string;
table_schema: string;
table_name: string;
column_name: string;
data_type: string;
pii_rules: string;
}
+112
View File
@@ -0,0 +1,112 @@
<!DOCTYPE html>
<html lang="pt-br">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Dadosfera Relatório de PII</title>
<link href="https://fonts.googleapis.com/css2?family=Quicksand:wght@400;500;700&display=swap" rel="stylesheet">
<style>
@page {
size: A4;
margin: 20mm 15mm 20mm 15mm;
}
body {
font-family: 'Quicksand', sans-serif;
color: #5c5c5c;
margin: 0;
padding: 20px;
}
.container {
margin: 0 24px;
}
.header {
display: flex;
align-items: center;
margin-bottom: 30px;
}
.logo {
max-width: 200px;
height: auto;
}
h1 {
color: #0d003b;
font-weight: 700;
margin-left: 20px;
}
table {
width: 100%;
border-collapse: collapse;
margin-top: 20px;
box-shadow: 0 2px 8px rgba(0,0,0,0.1);
border: 1px solid #d0d0d0; /* Borda externa da tabela */
}
th {
background-color: #1700a2;
color: white;
font-weight: bold;
text-align: left;
padding: 12px 15px;
border: 1px solid #3a26b8; /* Borda um pouco mais escura para o header */
}
td {
padding: 10px 15px;
border: 1px solid #d0d0d0; /* Borda clara da cor do texto */
}
tr:nth-child(even) {
background-color: #f9f9f9;
}
tr:nth-child(odd) {
background-color: white;
}
.info-section {
margin-top: 30px;
color: #5c5c5c;
}
.timestamp {
font-style: italic;
text-align: right;
margin-top: 20px;
font-size: 0.9em;
}
</style>
</head>
<body>
<div class="container">
<div class="header">
<img src="https://dadosfera.ai/wp-content/webp-express/webp-images/uploads/2022/06/Logo-Dadosfera1-1.png.webp" alt="Logo Dadosfera" class="logo">
<h1>Relatório de PII</h1>
</div>
<div class="info-section">
<p>Este relatório apresenta a estrutura de tabelas e suas as seguintes características de PII identificadas.</p>
</div>
<table>
<thead>
<tr>
<th>Database</th>
<th>Schema</th>
<th>Tabela</th>
<th>Coluna</th>
<th>Tipo de Dado</th>
<th>Regras PII</th>
</tr>
</thead>
<tbody>
{{#each dados}}
<tr>
<td>{{database_name}}</td>
<td>{{table_schema}}</td>
<td>{{table_name}}</td>
<td>{{column_name}}</td>
<td>{{data_type}}</td>
<td>{{pii_rules}}</td>
</tr>
{{/each}}
</tbody>
</table>
<p class="timestamp">Gerado em: {{dataGeracao}}</p>
</div>
</body>
</html>
+1
View File
@@ -14,6 +14,7 @@ interface IMetadata {
host?: string;
origin?: string;
custom_host?: string;
is_mask?: boolean;
}
export function PackTheMetadata(info: IMetadata): Metadata {
+14
View File
@@ -0,0 +1,14 @@
import { stringify } from 'csv'
export class CsvParser<T> {
async parse(data: T[]) {
return new Promise((resolve, reject) => {
stringify(data, { header: true }, function (err, output) {
if (err) reject(err);
console.log(output)
resolve(output)
});
})
}
}
+16
View File
@@ -0,0 +1,16 @@
import { readFileSync } from 'fs';
import Handlebars from 'handlebars';
export class HtmlParser<T> {
constructor(private readonly templatePath: string) {}
parse(data: T[]) {
const templateSource = readFileSync(this.templatePath, 'utf8');
const template = Handlebars.compile(templateSource);
return template({
dados: data,
dataGeracao: new Date().toLocaleString('pt-BR'),
});
}
}
+48
View File
@@ -0,0 +1,48 @@
import puppeteer from "puppeteer";
export class PDFParser<T extends string> {
async parse(data: T) {
try {
const browser = await puppeteer.launch({
executablePath: process.env.PUPPETEER_EXECUTABLE_PATH,
args: ['--no-sandbox', '--disable-setuid-sandbox', '--disable-dev-shm-usage']
});
const page = await browser.newPage();
page.setDefaultNavigationTimeout(0);
await page.setContent(data, {
waitUntil: 'networkidle0',
timeout: 30000
});
// Configurações adicionais para garantir um PDF válido
const pdfBuffer = await page.pdf({
format: 'A4',
printBackground: true,
margin: {
top: '20mm',
right: '15mm',
bottom: '20mm',
left: '15mm'
},
preferCSSPageSize: true,
displayHeaderFooter: false,
headerTemplate: '',
footerTemplate: '',
omitBackground: false,
pageRanges: '',
tagged: true,
outline: false,
timeout: 30000
});
await browser.close();
return pdfBuffer;
} catch (error) {
console.error('Error generating PDF:', error);
throw error;
}
}
}