add user-management

user-management/Makefile

@@ -0,0 +1,5 @@
+build:
+	sam build --use-container
+
+deploy: build
+	sam deploy --debug

user-management/app/boto3clients.py

@@ -0,0 +1,3 @@
+import boto3
+
+s3_client = boto3.client('s3')

user-management/app/config.py

@@ -0,0 +1 @@
+CHUNK_SIZE = 50

user-management/app/csv_utils.py

@@ -0,0 +1,83 @@
+import csv
+from typing import TextIO
+
+from smart_open import open
+
+
+def byte_ranges(
+    csvfile: str,
+    chunk_size: int = 100,
+    **kwargs,
+) -> list[tuple[int, int]]:
+    """Compute byte ranges for reading a CSV file in fixed-size line chunks.
+
+    Returns pairs (start_byte, end_byte) for each fixed-size group of lines.
+
+    Parameters
+    ----------
+    csvfile : str
+        Path to the CSV file, opened in binary mode internally.
+    chunk_size : int, optional
+        Number of lines per chunk. Default is 100.
+    **kwargs :
+        Extra options passed to `open()`, e.g., buffering.
+
+    Returns
+    -------
+    list of tuple[int, int]
+        Byte ranges covering each chunk of lines.
+
+    Example
+    -------
+    >>> byte_ranges("users.csv", chunk_size=500)
+    [(0, 3125), (3126, 6150), (6151, 9124)]
+    """
+    line_offsets = [0]
+
+    with open(csvfile, 'rb', **kwargs) as fp:
+        while True:
+            if not fp.readline():
+                break
+            line_offsets.append(fp.tell())
+
+    total_lines = len(line_offsets) - 1
+    byte_ranges = []
+
+    for start_line in range(1, total_lines + 1, chunk_size):
+        # Calculate the end line index, bounded by total lines
+        end_line = min(start_line + chunk_size - 1, total_lines)
+        # Get byte range for this chunk
+        start_byte = line_offsets[start_line - 1]
+        end_byte = line_offsets[end_line] - 1
+
+        byte_ranges.append((start_byte, end_byte))
+
+    return byte_ranges
+
+
+def detect_delimiter(sample: TextIO) -> str:
+    """Detect the delimiter character used in a CSV file.
+
+    Parameters
+    ----------
+    sample : TextIO
+        A file-like object opened in text mode (e.g., from `open('file.csv')`).
+        Must be readable and at position 0.
+
+    Returns
+    -------
+    str
+        The detected delimiter character (e.g., ',', ';', '\\t').
+
+    Raises
+    ------
+    csv.Error
+        If the file cannot be parsed as CSV or delimiter detection fails.
+    ValueError
+        If the file is empty or contains no detectable delimiter.
+    """
+    sniffer = csv.Sniffer()
+    dialect = sniffer.sniff(sample.read())
+    sample.seek(0)
+
+    return dialect.delimiter

user-management/app/events/batch/csv_into_chunks.py

@@ -0,0 +1,20 @@
+from aws_lambda_powertools.utilities.data_classes import (
+    EventBridgeEvent,
+    event_source,
+)
+from aws_lambda_powertools.utilities.typing import LambdaContext
+
+from boto3clients import s3_client
+from config import CHUNK_SIZE
+from csv_utils import byte_ranges
+
+transport_params = {'client': s3_client}
+
+
+@event_source(data_class=EventBridgeEvent)
+def lambda_handler(event: EventBridgeEvent, context: LambdaContext) -> bool:
+    new_image = event.detail['new_image']
+    csvfile = new_image['s3uri']
+    pairs = byte_ranges(csvfile, CHUNK_SIZE, transport_params=transport_params)
+
+    return True

user-management/app/events/batch/excel_to_csv.py

@@ -0,0 +1,14 @@
+from aws_lambda_powertools.utilities.data_classes import (
+    EventBridgeEvent,
+    event_source,
+)
+from aws_lambda_powertools.utilities.typing import LambdaContext
+
+from boto3clients import s3_client
+
+transport_params = {'client': s3_client}
+
+
+@event_source(data_class=EventBridgeEvent)
+def lambda_handler(event: EventBridgeEvent, context: LambdaContext) -> bool:
+    return True

user-management/app/events/batch/read_csv_chunk.py

@@ -0,0 +1,49 @@
+import csv
+from io import StringIO
+
+from aws_lambda_powertools.utilities.data_classes import (
+    EventBridgeEvent,
+    event_source,
+)
+from aws_lambda_powertools.utilities.typing import LambdaContext
+
+from boto3clients import s3_client
+
+transport_params = {'client': s3_client}
+
+
+@event_source(data_class=EventBridgeEvent)
+def lambda_handler(event: EventBridgeEvent, context: LambdaContext) -> bool:
+    new_image = event.detail['new_image']
+    csvfile = new_image['s3_uri']
+
+    data = _get_s3_object_range(
+        csvfile,
+        start_byte=new_image['start_byte'],
+        end_byte=new_image['end_byte'],
+        s3_client=s3_client,
+    )
+    reader = csv.reader(data)
+
+    for x in reader:
+        print(x)
+
+    return True
+
+
+def _get_s3_object_range(
+    s3_uri: str,
+    *,
+    start_byte: int,
+    end_byte: int,
+    s3_client,
+) -> StringIO:
+    bucket, key = s3_uri.replace('s3://', '').split('/', 1)
+
+    response = s3_client.get_object(
+        Bucket=bucket,
+        Key=key,
+        Range=f'bytes={start_byte}-{end_byte}',
+    )
+
+    return StringIO(response['Body'].read().decode('utf-8'))

user-management/app/events/email_receiving.py

@@ -0,0 +1,45 @@
+import urllib.parse as urllib_parse
+from email.utils import parseaddr
+from typing import Any, Iterator
+
+from aws_lambda_powertools import Logger
+from aws_lambda_powertools.utilities.data_classes import SESEvent, event_source
+from aws_lambda_powertools.utilities.typing import LambdaContext
+
+logger = Logger(__name__)
+
+
+@logger.inject_lambda_context
+@event_source(data_class=SESEvent)
+def lambda_handler(event: SESEvent, context: LambdaContext) -> dict:
+    ses = event.record.ses
+    to = urllib_parse.unquote(ses.receipt.recipients[0]).lower()
+    name, email_from = parseaddr(get_header_value(ses.mail.headers, 'from'))
+    subject = get_header_value(
+        ses.mail.headers,
+        'subject',
+        default='',
+        raise_on_missing=False,
+    )
+
+    if email_from == 'sergio@somosbeta.com.br':
+        return {'disposition': 'CONTINUE'}
+
+    return {'disposition': 'STOP_RULE_SET'}
+
+
+def get_header_value(
+    headers: Iterator,
+    header_name: str,
+    *,
+    default: Any = None,
+    raise_on_missing: bool = True,
+) -> str:
+    for header in headers:
+        if header.name.lower() == header_name:
+            return header.value
+
+    if raise_on_missing:
+        raise ValueError(f'{header_name} not found.')
+
+    return default

user-management/cf.py

@@ -0,0 +1,56 @@
+# /// script
+# dependencies = [
+#   "cloudflare"
+# ]
+# ///
+
+from cloudflare import Cloudflare
+
+CLOUDFLARE_ACCOUNT_ID = '5436b62470020c04b434ad31c3e4cf4e'
+CLOUDFLARE_API_TOKEN = 'gFndkBJCzH4pRX7mKXokdWfw1xhm8-9FHfvLfhwa'
+
+
+client = Cloudflare(api_token=CLOUDFLARE_API_TOKEN)
+
+assistant = """
+You are a data analysis assistant specialized in identifying Brazilian
+personal data from CSV files.
+
+These CSV files may or may not include headers.
+
+Your task is to analyze the content and identify only three possible
+data types: 'name', 'cpf', and 'email'.
+
+Ignore all other fields.
+"""
+
+csv_content = """
+Sérgio Rafael de Siqueira,10,07879819908,osergiosiqueria@gmail.com,cipa
+Tiago Maciel,12,086.790.049-01,tiago@somosbeta.com.br,nr 10
+"""
+
+prompt = f"""
+Here is a CSV sample:
+
+{csv_content}
+
+Your task is to:
+- Detect which columns most likely contain "name", "cpf", or "email".
+- Skip any category that is not present in the data.
+- Return ONLY a valid Python list of tuples, like:
+[('name', index), ('cpf', index), ('email', index)]
+- Use the column index that most likely matches each data type,
+based on frequency and data format.
+- Don't include explanations, code, or any additional text.
+"""
+
+r = client.ai.run(
+    model_name='@cf/meta/llama-3-8b-instruct',
+    account_id=CLOUDFLARE_ACCOUNT_ID,
+    messages=[
+        {'role': 'system', 'content': assistant},
+        {'role': 'user', 'content': prompt},
+    ],
+)
+
+print(r)

user-management/pyproject.toml

@@ -0,0 +1,31 @@
+[project]
+name = "user-management"
+version = "0.1.0"
+description = ""
+readme = ""
+requires-python = ">=3.13"
+dependencies = ["layercake"]
+
+[dependency-groups]
+dev = [
+    "pytest>=8.3.4",
+    "pytest-cov>=6.0.0",
+    "ruff>=0.9.1",
+]
+
+[tool.pytest.ini_options]
+pythonpath = ["app/"]
+addopts = "--cov --cov-report html -v"
+
+[tool.ruff]
+target-version = "py311"
+src = ["app"]
+
+[tool.ruff.format]
+quote-style = "single"
+
+[tool.ruff.lint]
+select = ["E", "F", "I"]
+
+[tool.uv.sources]
+layercake = { path = "../layercake" }

user-management/pyrightconfig.json

@@ -0,0 +1,3 @@
+{
+  "extraPaths": ["app/"]
+}

user-management/samconfig.toml

@@ -0,0 +1,9 @@
+version = 0.1
+[default.deploy.parameters]
+stack_name = "saladeaula-user-management"
+resolve_s3 = true
+s3_prefix = "user_management"
+region = "sa-east-1"
+confirm_changeset = false
+capabilities = "CAPABILITY_IAM"
+image_repositories = []

user-management/template.yaml

@@ -0,0 +1,110 @@
+AWSTemplateFormatVersion: 2010-09-09
+Transform: AWS::Serverless-2016-10-31
+
+Parameters:
+  BucketName:
+    Type: String
+    Default: saladeaula.digital
+  UserTable:
+    Type: String
+    Default: betaeducacao-prod-users_d2o3r5gmm4it7j
+
+Globals:
+  Function:
+    CodeUri: app/
+    Runtime: python3.13
+    Tracing: Active
+    Architectures:
+      - x86_64
+    Layers:
+      - !Sub arn:aws:lambda:sa-east-1:336641857101:layer:layercake:72
+    Environment:
+      Variables:
+        TZ: America/Sao_Paulo
+        LOG_LEVEL: DEBUG
+        POWERTOOLS_LOGGER_SAMPLE_RATE: 0.1
+        POWERTOOLS_LOGGER_LOG_EVENT: true
+        USER_TABLE: !Ref UserTable
+
+Resources:
+  EventLog:
+    Type: AWS::Logs::LogGroup
+    Properties:
+      RetentionInDays: 90
+
+  EventCsvChunksFunction:
+    Type: AWS::Serverless::Function
+    Properties:
+      Handler: events.batch.csv_chunks.lambda_handler
+      LoggingConfig:
+        LogGroup: !Ref EventLog
+      Policies:
+        - S3CrudPolicy:
+            BucketName: !Ref BucketName
+      Events:
+        DynamoDBEvent:
+          Type: EventBridgeRule
+          Properties:
+            Pattern:
+              resources: [betaeducacao-prod-users_d2o3r5gmm4it7j]
+              detail:
+                new_image:
+                  sk:
+                    - prefix: batch_jobs#
+
+  EventEmailReceivingFunction:
+    Type: AWS::Serverless::Function
+    Properties:
+      Handler: events.email_receiving.lambda_handler
+      LoggingConfig:
+        LogGroup: !Ref EventLog
+
+  LambdaInvokePermission:
+    Type: AWS::Lambda::Permission
+    Properties:
+      FunctionName: !GetAtt EventEmailReceivingFunction.Arn
+      Action: lambda:InvokeFunction
+      Principal: ses.amazonaws.com
+      SourceArn: !Sub arn:aws:ses:${AWS::Region}:${AWS::AccountId}:receipt-rule-set/*
+
+  BucketPolicy:
+    Type: AWS::S3::BucketPolicy
+    Properties:
+      Bucket: !Ref BucketName
+      PolicyDocument:
+        Version: "2012-10-17"
+        Statement:
+          - Effect: Allow
+            Principal:
+              Service: ses.amazonaws.com
+            Action: s3:PutObject
+            Resource: !Sub arn:aws:s3:::${BucketName}/*
+            Condition:
+              StringEquals:
+                aws:SourceAccount: !Ref AWS::AccountId
+              StringLike:
+                aws:SourceArn: !Sub arn:aws:ses:${AWS::Region}:${AWS::AccountId}:receipt-rule-set/*
+
+  EmailReceiptRuleSet:
+    Type: AWS::SES::ReceiptRuleSet
+    Properties:
+      RuleSetName: users.noreply.saladeaula.digital
+
+  EmailReceiptRule:
+    Type: AWS::SES::ReceiptRule
+    DependsOn:
+      - LambdaInvokePermission
+      - BucketPolicy
+    Properties:
+      RuleSetName: !Ref EmailReceiptRuleSet
+      Rule:
+        Name: lambda
+        Enabled: true
+        Actions:
+          - LambdaAction:
+              FunctionArn: !GetAtt EventEmailReceivingFunction.Arn
+              InvocationType: RequestResponse
+          - S3Action:
+              BucketName: !Ref BucketName
+              ObjectKeyPrefix: "mailbox"
+        ScanEnabled: true

user-management/tests/conftest.py

@@ -0,0 +1,16 @@
+from dataclasses import dataclass
+
+import pytest
+
+
+@dataclass
+class LambdaContext:
+    function_name: str = 'test'
+    memory_limit_in_mb: int = 128
+    invoked_function_arn: str = 'arn:aws:lambda:eu-west-1:809313241:function:test'
+    aws_request_id: str = '52fdfc07-2182-154f-163f-5f0f9a621d72'
+
+
+@pytest.fixture
+def lambda_context() -> LambdaContext:
+    return LambdaContext()

user-management/tests/events/batch/test_csv_into_chunks.py

@@ -0,0 +1,13 @@
+import events.batch.csv_into_chunks as app
+
+
+def test_chunk_csv(lambda_context):
+    event = {
+        'detail': {
+            'new_image': {
+                's3uri': 's3://saladeaula.digital/samples/large_users.csv',
+            },
+        },
+    }
+
+    app.lambda_handler(event, lambda_context)  # type: ignore

user-management/tests/events/test_email_receiving.py

@@ -0,0 +1,136 @@
+from aws_lambda_powertools.utilities.typing import LambdaContext
+
+import events.email_receiving as app
+
+event = {
+    'Records': [
+        {
+            'eventSource': 'aws:ses',
+            'eventVersion': '1.0',
+            'ses': {
+                'mail': {
+                    'timestamp': '2025-05-29T15:50:41.604Z',
+                    'source': 'sergio@somosbeta.com.br',
+                    'messageId': '2994higq3tr7efijr3lj65etntffapgg1q7hea81',
+                    'destination': [
+                        'org+35980592000130@users.noreply.saladeaula.digital'
+                    ],
+                    'headersTruncated': False,
+                    'headers': [
+                        {'name': 'Return-Path', 'value': '<sergio@somosbeta.com.br>'},
+                        {
+                            'name': 'Received',
+                            'value': 'from mail-lf1-f54.google.com (mail-lf1-f54.google.com [209.85.167.54]) by inbound-smtp.sa-east-1.amazonaws.com with SMTP id 2994higq3tr7efijr3lj65etntffapgg1q7hea81 for org+35980592000130@users.noreply.saladeaula.digital; Thu, 29 May 2025 15:50:41 +0000 (UTC)',
+                        },
+                        {'name': 'X-SES-Spam-Verdict', 'value': 'PASS'},
+                        {'name': 'X-SES-Virus-Verdict', 'value': 'PASS'},
+                        {
+                            'name': 'Received-SPF',
+                            'value': 'pass (spfCheck: domain of somosbeta.com.br designates 209.85.167.54 as permitted sender) client-ip=209.85.167.54; envelope-from=sergio@somosbeta.com.br; helo=mail-lf1-f54.google.com;',
+                        },
+                        {
+                            'name': 'Authentication-Results',
+                            'value': 'amazonses.com; spf=pass (spfCheck: domain of somosbeta.com.br designates 209.85.167.54 as permitted sender) client-ip=209.85.167.54; envelope-from=sergio@somosbeta.com.br; helo=mail-lf1-f54.google.com; dkim=pass header.i=@somosbeta.com.br; dmarc=none header.from=somosbeta.com.br;',
+                        },
+                        {
+                            'name': 'X-SES-RECEIPT',
+                            'value': 'AEFBQUFBQUFBQUFHVWpuODdPY2tGUlordE5YWkVEUlZNWWZFYkpDMU5MUURyaHNVSldnTGhEVWhCQzd5UGpzWHI4LzJoS1VaN0lOU0FkMzJFU0h6MjVuUzk2c09KUXlzbUJQdHd6T0d0Y2ptZXhRVk1KY3RkOXpRamZMb3hwSGJIVlFla2tBcmZvRmYwQS9WU3hBVlBqcUpDYm00eTdiRnRqNW45ek9ld0ZyTGJKV3k2TXRpc0J6aGhBdmFvZDFDQ000Zm9QTng3VHljNXArM0hjT2ZsYkhtM3RCZnpRV1NOczU2RDdmL0RKclJOcDNvY2ZxV1hmajNYMkczVHpsWEZCMm40Z2pQM29udkMyb01vN3JwU0p2TUI1WGorN2JPd2RPYW5lUDN3T3RMRlhsdEpGbGNCa3c9PQ==',
+                        },
+                        {
+                            'name': 'X-SES-DKIM-SIGNATURE',
+                            'value': 'a=rsa-sha256; q=dns/txt; b=KPtFiBwsOTBl1YVLRTSfaZ+X6h7uSSOu/i1Cw6Pd+wvMBHRWy9EYcWUjyDjsLG/uYHShLW4+LHsSg9HiqrAP2jVJSAawrIwZr1wPQo7ovQvWuZfHQN/StgXIgBU+L7Bp6GSR26LRufxjj7q9YBmEeirjJ3d0G8E/rF2QqeITlpo=; c=relaxed/simple; s=bm3ypaoivbtdzmy3b6w37fzb5voa2uru; d=amazonses.com; t=1748533842; v=1; bh=kTUCV1DQAazu4FsUi1MrelD2QvSfHGArZ/c6A79t3/E=; h=From:To:Cc:Bcc:Subject:Date:Message-ID:MIME-Version:Content-Type:X-SES-RECEIPT;',
+                        },
+                        {
+                            'name': 'Received',
+                            'value': 'by mail-lf1-f54.google.com with SMTP id 2adb3069b0e04-54b0d638e86so1570269e87.1 for <org+35980592000130@users.noreply.saladeaula.digital>; Thu, 29 May 2025 08:50:40 -0700 (PDT)',
+                        },
+                        {
+                            'name': 'DKIM-Signature',
+                            'value': 'v=1; a=rsa-sha256; c=relaxed/relaxed; d=somosbeta.com.br; s=google; t=1748533838; x=1749138638; darn=users.noreply.saladeaula.digital; h=to:subject:message-id:date:from:in-reply-to:references:mime-version:from:to:cc:subject:date:message-id:reply-to; bh=kTUCV1DQAazu4FsUi1MrelD2QvSfHGArZ/c6A79t3/E=; b=Qi8gk/kTpwXCLDM7FPS7ULTy+9gO/4WsGL9zY1xEDw0Rp38f4rVR8L95hIhwK2daA27mq3pv9TdrK3XKQQIuSvRVvaM0b/evkZD8QhaT9tCmL0eKEBB4czGB0OSS3Q4qP34GFWMmXIaxoKIo1td76JnXbto9ZQvjUTBr3GGlF3Lm/MPTaAHs1b3dalv2diTvyj1tzoeb4wGePKsqLh5LKGOxbbWsxPeHEJ8sLM4LyJjxoqSOO0wgKdH5S/ZNpHWcJtXBntjiDUZNeQ5ucEn8ZLbADCObZZV/gH9i/cB1BmlSvJP3D07uJTAEBqyepd+W9fIW2mox/+fmOb3OEHRthQ==',
+                        },
+                        {
+                            'name': 'X-Google-DKIM-Signature',
+                            'value': 'v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1748533838; x=1749138638; h=to:subject:message-id:date:from:in-reply-to:references:mime-version :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=kTUCV1DQAazu4FsUi1MrelD2QvSfHGArZ/c6A79t3/E=; b=UnL/uRXRahnH5uStZ276LH4kqpHigngw4iql9GHKmFaIKxJ8hLGn/wu7ie4ljnw8m/ I4CvhDKH4TVtIWPS81fm06PMgeqQYRX3jLhHvIltROCNVX6ZFzbAXgiAlk0NS1npvDYJ evVgSaPPco4D/8pMWZX4fUjU+8me32ChKxHsklEAts/LiD+MvTuHCHZovSEj1aXAz91b yGZe1bx2+phuqzUZyIOeheKjl4TNjEBx83omOzf9HtClKhjzCwHjfZ8uk2lhJ10ogKZ0 GNQ5OlnPkgdAg0/+HsifvGR6xfkFsiunIDyinBWoOhMU1o0+DiicxOIjY8QEayF3MLUt REoA==',
+                        },
+                        {
+                            'name': 'X-Gm-Message-State',
+                            'value': 'AOJu0Yxw0icQkFV090vn5hx/hKp0ePH78Wr0iqi4V3x4mpVXrRX8te2o 30aBYeZRPwn8SRSrq/kbn4bLcs5mPDB+iRP9IGFxS7KLSQi+KG4PQeDHyW3R/AgOPHACUUXUUyz Vcwny029WGY5PVhxlikAYdDfhNdO8GM2DKMV1+Oxy/a+qmt5LZeuy',
+                        },
+                        {
+                            'name': 'X-Gm-Gg',
+                            'value': 'ASbGncslCMPPU/pax0+RNy/cQR/Y/wUroSJMvI2DCCMq6Qld+Ih1jG4+HnhQPqn3nTK EEW6/99tqazq+SKy+31AB77ajVczvJQTElRSW/+bhd42l7by2hicTKElcR3GWivlrqd1TywUZOB DkB9J/vupSV0PDCJfZVi+7Tb9Pb61nnxaU+SQ=',
+                        },
+                        {
+                            'name': 'X-Google-Smtp-Source',
+                            'value': 'AGHT+IFYi41KmJjGcfQmUvWJDdTAzGIv2JlL9XAwBpAb53mMOOm3tttzkhbvfuiKh/DI9NjITHuO3xuEPqnPI9lpum8=',
+                        },
+                        {
+                            'name': 'X-Received',
+                            'value': 'by 2002:a05:6512:1392:b0:553:2f61:58f1 with SMTP id 2adb3069b0e04-5532f615a8dmr2268707e87.53.1748533837647; Thu, 29 May 2025 08:50:37 -0700 (PDT)',
+                        },
+                        {'name': 'MIME-Version', 'value': '1.0'},
+                        {
+                            'name': 'References',
+                            'value': '<CAMThe4mV9=1-BLiOi9MU3fAS=C6uYE9+3hKUjibrwxxngYNn2Q@mail.gmail.com>',
+                        },
+                        {
+                            'name': 'In-Reply-To',
+                            'value': '<CAMThe4mV9=1-BLiOi9MU3fAS=C6uYE9+3hKUjibrwxxngYNn2Q@mail.gmail.com>',
+                        },
+                        {
+                            'name': 'From',
+                            'value': 'Sérgio Rafael Siqueira <sergio@somosbeta.com.br>',
+                        },
+                        {'name': 'Date', 'value': 'Thu, 29 May 2025 12:50:26 -0300'},
+                        {
+                            'name': 'X-Gm-Features',
+                            'value': 'AX0GCFvofROqzf21KTgiIJq_AULCNljEuNFUJBk2xQGwVKmPjim_4slYIOP0WRw',
+                        },
+                        {
+                            'name': 'Message-ID',
+                            'value': '<CAMThe4=yMRJg4YOcACYAR509N1RyWyQgAghyVmr=NuSJnbondg@mail.gmail.com>',
+                        },
+                        {'name': 'Subject', 'value': 'Re: test'},
+                        {
+                            'name': 'To',
+                            'value': 'org+35980592000130@users.noreply.saladeaula.digital',
+                        },
+                        {
+                            'name': 'Content-Type',
+                            'value': 'multipart/alternative; boundary="00000000000045b8c206364842b3"',
+                        },
+                    ],
+                    'commonHeaders': {
+                        'returnPath': 'sergio@somosbeta.com.br',
+                        'from': ['"Sérgio Rafael Siqueira" <sergio@somosbeta.com.br>'],
+                        'date': 'Thu, 29 May 2025 12:50:26 -0300',
+                        'to': ['org+35980592000130@users.noreply.saladeaula.digital'],
+                        'messageId': '<CAMThe4=yMRJg4YOcACYAR509N1RyWyQgAghyVmr=NuSJnbondg@mail.gmail.com>',
+                        'subject': 'Re: test',
+                    },
+                },
+                'receipt': {
+                    'timestamp': '2025-05-29T15:50:41.604Z',
+                    'processingTimeMillis': 1105,
+                    'recipients': [
+                        'org+35980592000130@users.noreply.saladeaula.digital'
+                    ],
+                    'spamVerdict': {'status': 'PASS'},
+                    'virusVerdict': {'status': 'PASS'},
+                    'spfVerdict': {'status': 'PASS'},
+                    'dkimVerdict': {'status': 'PASS'},
+                    'dmarcVerdict': {'status': 'GRAY'},
+                    'action': {
+                        'type': 'Lambda',
+                        'functionArn': 'arn:aws:lambda:sa-east-1:336641857101:function:saladeaula-user-managemen-EventEmailReceivingFunct-LmnnEfi9tL2O',
+                        'invocationType': 'Event',
+                    },
+                },
+            },
+        }
+    ]
+}
+
+
+def test_email_receiving(lambda_context: LambdaContext):
+    assert app.lambda_handler(event, lambda_context) == {'disposition': 'CONTINUE'}

user-management/tests/samples/users.csv

@@ -0,0 +1,28 @@
+CADASTRO DE COLABORADOR,,,,
+,NOME COMPLETO,EMAIL (letra minúscula),CPF,TREINAMENTO
+,ANDRE HENRIQUE LOPES ZAFALON,henrique.zafalon@fanucamerica.com,261.955.138-22,NR-35 (RECICLAGEM)
+,SERGIO DA SILVA CUPERTINO,sergio.cupertino@fanucamerica.com,066.945.708-64,NR-10 (RECICLAGEM)
+,SERGIO DA SILVA CUPERTINO,sergio.cupertino@fanucamerica.com,066.945.708-64,NR-35 (RECICLAGEM)
+,ROVANE CAMPOS,rovane.campos@fanucamerica.com,095.958.578-82,NR-10 (RECICLAGEM)
+,ROVANE CAMPOS,rovane.campos@fanucamerica.com,095.958.578-82,NR-35 (RECICLAGEM)
+,MARCIO ATSUSHI KANEKO MASUDA,marcio.masuda@fanucamerica.com,293.042.798-10,NR-10 (RECICLAGEM)
+,FABIO AKIRA HARAGUCHI,fabio.haraguchi@fanucamerica.com,287.018.428-03,NR-10 (RECICLAGEM)
+,EMIDIO YOITI MOCHIZUKI,emidio.mochizuki@fanucamerica.com,268.579.208-26,NR-10 (RECICLAGEM)
+,EMIDIO YOITI MOCHIZUKI,emidio.mochizuki@fanucamerica.com,268.579.208-26,NR-35 (RECICLAGEM)
+,ERIC HIDEKI MORIKIO,eric.morikio@fanucamerica.com,417.359.838-61,NR-10 (RECICLAGEM)
+,HENRIQUE DE FIGUEIREDO BASTOS FERRAZ,henrique.ferraz@fanucamerica.com,417.059.788-51,NR-10 (RECICLAGEM)
+,LAYS MORETTI DA SILVA,lays.silva@fanucamerica.com,013.107.662-07,NR-10 (RECICLAGEM)
+,LAYS MORETTI DA SILVA,lays.silva@fanucamerica.com,013.107.662-07,NR-12
+,ANDRE DE SOUZA,andre.souza@fanucamerica.com,290.688.648-31,NR-10 (RECICLAGEM)
+,ANDRE DE SOUZA,andre.souza@fanucamerica.com,290.688.648-31,NR-12
+,RAFAEL TOSHIO BURATO MAEDA,rafael.maeda@fanucamerica.com,394.153.268-59,NR-10 (RECICLAGEM)
+,RAFAEL TOSHIO BURATO MAEDA,rafael.maeda@fanucamerica.com,394.153.268-59,NR-12
+,RAFAEL TOSHIO BURATO MAEDA,rafael.maeda@fanucamerica.com,394.153.268-59,NR-35 (RECICLAGEM)
+,RICARDO GALLES BONET,ricardo.bonet@fanucamerica.com,424.430.528-93,NR-10 (RECICLAGEM)
+,RULIO SIEFERT SERA,rulio.sera@fanucamerica.com,063.916.859-08,NR-10 (RECICLAGEM)
+,MACIEL FERREIRA BOMFIM,maciel.bomfim@fanucamerica.com,334.547.088-85,NR-10 (RECICLAGEM)
+,JAIME EDUARDO GALVEZ AVILES,jaime.galvez@fanucamerica.com,280.238.818-50,NR-12
+,JAIME EDUARDO GALVEZ AVILES,jaime.galvez@fanucamerica.com,280.238.818-50,NR-35 (RECICLAGEM)
+,HIGOR MACHADO SILVA,higor.silva@fanucamerica.com,419.879.878-88,NR-12
+,LÁZARO SOUZA DIAS,lazaro.dias@fanucamerica.com,067.179.825-19,NR-12
+,JOÃO PEDRO AGUIAR GALASSO,joao.pedro@fanucamerica.com,570.403.588-40,NR-12

user-management/tests/test_csv_utils.py

@@ -0,0 +1,29 @@
+from csv_utils import byte_ranges, detect_delimiter
+
+
+def test_detect_delimiter():
+    with open('tests/samples/users.csv') as fp:
+        assert detect_delimiter(fp) == ','
+
+
+def test_byte_ranges():
+    csvpath = 'tests/samples/users.csv'
+    ranges = byte_ranges(csvpath, 10)
+    *_, pair = ranges
+    start_byte, end_byte = pair
+
+    assert ranges == [(0, 808), (809, 1655), (1656, 2303)]
+
+    expected = """,RICARDO GALLES BONET,ricardo.bonet@fanucamerica.com,424.430.528-93,NR-10 (RECICLAGEM)
+,RULIO SIEFERT SERA,rulio.sera@fanucamerica.com,063.916.859-08,NR-10 (RECICLAGEM)
+,MACIEL FERREIRA BOMFIM,maciel.bomfim@fanucamerica.com,334.547.088-85,NR-10 (RECICLAGEM)
+,JAIME EDUARDO GALVEZ AVILES,jaime.galvez@fanucamerica.com,280.238.818-50,NR-12
+,JAIME EDUARDO GALVEZ AVILES,jaime.galvez@fanucamerica.com,280.238.818-50,NR-35 (RECICLAGEM)
+,HIGOR MACHADO SILVA,higor.silva@fanucamerica.com,419.879.878-88,NR-12
+,LÁZARO SOUZA DIAS,lazaro.dias@fanucamerica.com,067.179.825-19,NR-12
+,JOÃO PEDRO AGUIAR GALASSO,joao.pedro@fanucamerica.com,570.403.588-40,NR-12"""
+
+    with open(csvpath, 'rb') as f:
+        f.seek(start_byte)
+        data = f.read(end_byte - start_byte + 1)
+        assert data.decode('utf-8') == expected