documente_shared 0.1.1__py3-none-any.whl → 0.1.3__py3-none-any.whl

documente_shared/domain/base_enum.py

@@ -0,0 +1,54 @@
+from enum import Enum
+from typing import Union, Optional
+
+
+class BaseEnum(Enum):
+    """Provides the common functionalties to multiple model choices."""
+
+    @classmethod
+    def get_members(cls):
+        return [tag for tag in cls if type(tag.value) in [int, str, float]]
+
+    @classmethod
+    def choices(cls):
+        """Generate choice options for models."""
+        return [
+            (option.value, option.value)
+            for option in cls
+            if type(option.value) in [int, str, float]
+        ]
+
+    @classmethod
+    def values(cls):
+        """Returns values from choices."""
+        return [option.value for option in cls]
+
+    def __str__(self):  # noqa: D105
+        return str(self.value)
+
+    def __repr__(self):
+        return self.__str__()
+
+    def __hash__(self):
+        return hash(self.value)
+
+    @classmethod
+    def as_list(cls):
+        """Returns properties as a list."""
+        return [
+            value
+            for key, value in cls.__dict__.items()
+            if isinstance(value, str) and not key.startswith('__')
+        ]
+
+    @classmethod
+    def from_value(
+        cls,
+        value: Union[str, int],
+    ) -> Optional['BaseEnum']:
+        for tag in cls:
+            if isinstance(tag.value, str) and str(tag.value).upper() == str(value).upper():
+                return tag
+            elif not isinstance(tag.value, str) and tag.value == value:
+                return tag
+        return None

documente_shared/domain/entities.py

@@ -0,0 +1,114 @@
+from dataclasses import dataclass
+from datetime import datetime
+from decimal import Decimal
+from typing import Optional
+
+from documente_shared.domain.enums import DocumentProcessStatus
+
+
+@dataclass
+class DocumentProcess(object):
+    digest: str
+    status: DocumentProcessStatus
+    file_path: str
+    processed_csv_path: str
+    processed_xlsx_path: str
+    processing_time: Optional[Decimal] = None
+    enqueued_at: Optional[datetime] = None
+    started_at: Optional[datetime] = None
+    failed_at: Optional[datetime] = None
+    processed_at: Optional[datetime] = None
+
+    @property
+    def is_pending(self) -> bool:
+        return self.status == DocumentProcessStatus.PENDING
+
+    @property
+    def is_enqueued(self) -> bool:
+        return self.status == DocumentProcessStatus.ENQUEUED
+
+    @property
+    def is_processing(self) -> bool:
+        return self.status == DocumentProcessStatus.PROCESSING
+
+    @property
+    def is_completed(self) -> bool:
+        return self.status == DocumentProcessStatus.COMPLETED
+
+    @property
+    def is_failed(self) -> bool:
+        return self.status == DocumentProcessStatus.FAILED
+
+    @property
+    def is_valid(self) -> bool:
+        return all([
+            self.digest,
+            self.status,
+            self.file_path,
+        ])
+
+    def enqueue(self):
+        self.status = DocumentProcessStatus.ENQUEUED
+        self.enqueued_at = datetime.now()
+
+    def processing(self):
+        self.status = DocumentProcessStatus.PROCESSING
+        self.started_at = datetime.now()
+
+    def failed(self):
+        self.status = DocumentProcessStatus.FAILED
+        self.failed_at = datetime.now()
+
+    def completed(self):
+        self.status = DocumentProcessStatus.COMPLETED
+        self.processed_at = datetime.now()
+
+
+
+    @property
+    def to_dict(self) -> dict:
+        return {
+            'digest': self.digest,
+            'status': self.status.value,
+            'file_path': self.file_path,
+            'processed_csv_path': self.processed_csv_path,
+            'processed_xlsx_path': self.processed_xlsx_path,
+            'processing_time': (
+                str(self.processing_time)
+                if self.processing_time else None
+            ),
+            'enqueued_at': self.enqueued_at.isoformat() if self.enqueued_at else None,
+            'started_at': self.started_at.isoformat() if self.started_at else None,
+            'failed_at': self.failed_at.isoformat() if self.failed_at else None,
+            'processed_at': self.processed_at.isoformat() if self.processed_at else None,
+        }
+
+    @classmethod
+    def from_dict(cls, data: dict) -> 'DocumentProcess':
+        return cls(
+            digest=data.get('digest'),
+            status=DocumentProcessStatus.from_value(data.get('status')),
+            file_path=data.get('file_path'),
+            processed_csv_path=data.get('processed_csv_path'),
+            processed_xlsx_path=data.get('processed_xlsx_path'),
+            processing_time=(
+                Decimal(data.get('processing_time'))
+                if data.get('processing_time') else None
+            ),
+            enqueued_at=(
+                datetime.fromisoformat(data.get('enqueued_at'))
+                if data.get('enqueued_at') else None
+            ),
+            started_at=(
+                datetime.fromisoformat(data.get('started_at'))
+                if data.get('started_at') else None
+            ),
+            failed_at=(
+                datetime.fromisoformat(data.get('failed_at'))
+                if data.get('failed_at') else None
+            ),
+            processed_at=(
+                datetime.fromisoformat(data.get('processed_at'))
+                if data.get('processed_at') else None
+            ),
+        )

documente_shared/{enums.py → domain/enums.py}

@@ -1,4 +1,4 @@
-from domain.base_enum import BaseEnum
+from documente_shared.domain.base_enum import BaseEnum
 
 
 class DocumentProcessStatus(BaseEnum):

documente_shared/domain/repositories.py

@@ -0,0 +1,14 @@
+from abc import ABC, abstractmethod
+
+from documente_shared.domain.entities import DocumentProcess
+
+
+class DocumentProcessRepository(ABC):
+
+    @abstractmethod
+    def persist(self, instance: DocumentProcess) -> DocumentProcess:
+        raise NotImplementedError
+
+    @abstractmethod
+    def remove(self, instance: DocumentProcess):
+        raise NotImplementedError

documente_shared/infrastructure/dynamo_repositories.py

@@ -0,0 +1,17 @@
+from documente_shared.domain.entities import DocumentProcess
+from documente_shared.domain.repositories import DocumentProcessRepository
+from documente_shared.infrastructure.dynamo_table import DynamoDBTable
+
+
+class DynamoDocumentProcessRepository(
+    DynamoDBTable,
+    DocumentProcessRepository,
+):
+
+    def persist(self, instance: DocumentProcess) -> DocumentProcess:
+        self.put(instance.to_dict)
+        return instance
+
+
+    def remove(self, instance: DocumentProcess):
+        self.delete(key=instance.digest)

documente_shared/infrastructure/dynamo_table.py

@@ -0,0 +1,75 @@
+from dataclasses import dataclass
+
+import boto3
+from boto3.dynamodb.conditions import Key
+
+
+RETURN_VALUES = 'UPDATED_NEW'
+
+@dataclass
+class DynamoDBTable(object):
+    table_name: str
+
+    def __post_init__(self):
+        self._table = boto3.resource('dynamodb').Table(self.table_name)
+
+    def get(self, key: str):
+        return self._table.get_item(Key=key).get('Item')
+
+    def get_all(self):
+        return self._table.scan().get('Items')
+
+    def upsert(self, key, attributes):
+        return self.put({**key, **attributes})
+
+
+    def filter_by(self, attribute, target_value):
+        return self._table(
+            FilterExpression=Key(attribute).eq(target_value),
+        ).get('Items')
+
+    def put(self, attributes: dict, condition: dict = None):
+        extra_args = {}
+        if condition:
+            extra_args['ConditionExpression'] = condition
+        return self._table.put_item(Item=attributes, **extra_args)
+
+
+    def update(self, key: str, attributes: dict):
+        return self._table.update_item(
+            Key=key,
+            UpdateExpression=self._update_expression(attributes),
+            ExpressionAttributeNames=self._expression_attribute_names(attributes),
+            ExpressionAttributeValues=self._expression_attribute_values(attributes),
+            ReturnValues=RETURN_VALUES,
+        )
+
+    def delete(self, key: str):
+        return self._table.delete_item(Key=key)
+
+    def count(self) -> int:
+        return self._table.item_count
+
+
+    @classmethod
+    def _update_expression(cls, attributes):
+        return 'SET {param}'.format(
+            param=','.join(
+                '#{key}=:{key}'.format(
+                    key=key,
+                )
+                for key in attributes
+            ),
+        )
+
+    @classmethod
+    def _expression_attribute_names(cls, attributes):
+        return {
+            '#{key}'.format(key=key): key for key in attributes
+        }
+
+    @classmethod
+    def _expression_attribute_values(cls, attributes):
+        return {
+            ':{key}'.format(key=key): attr for key, attr in attributes.items()
+        }

documente_shared/infrastructure/s3_bucket.py

@@ -0,0 +1,38 @@
+from dataclasses import dataclass
+from typing import Optional
+
+from python_layer.python import boto3
+
+def remove_none_values(data: dict) -> dict:  # noqa: WPS110
+    return {key: value for key, value in data.items() if value is not None}  # noqa: WPS110
+
+
+@dataclass
+class S3Bucket(object):
+    bucket_url: str
+
+    def __post_init__(self):
+        self._resource = boto3.resource('s3')
+
+
+    def get(self, file_name: str):
+        return self._resource.Object(self.bucket_url, file_name).get()
+
+    def upload(self, file_name, file_content, content_type: Optional[str] = None):
+        optional_params = {'ContentType': content_type}
+        return self._resource.Object(self.bucket_url, file_name).put(
+            Body=file_content,
+            **remove_none_values(optional_params),
+        )
+
+    def delete(self, file_name):
+        return self._resource.Object(self.bucket_url, file_name).delete()
+
+    def get_url(self, file_name):
+        return 'https://{bucket_url}.s3.amazonaws.com/{file_name}'.format(
+            bucket_url=self.bucket_url,
+            file_name=file_name,
+        )
+
+    def read(self, file_name):
+        return self.get(file_name)['Body'].read()

{documente_shared-0.1.1.dist-info → documente_shared-0.1.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: documente_shared
-Version: 0.1.1
+Version: 0.1.3
 Summary: Shared utilities for Documente AI projects
 License: MIT
 Author: Tech

documente_shared-0.1.3.dist-info/RECORD

@@ -0,0 +1,13 @@
+documente_shared/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+documente_shared/domain/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+documente_shared/domain/base_enum.py,sha256=DojAfn-zQdtjtImeHUpBzE6TBTm07XrbMOdW3h8RVd8,1449
+documente_shared/domain/entities.py,sha256=0Flg2yWYdVM34hAwZmarn8CUKGZ9LchtC6Imeikbthk,3726
+documente_shared/domain/enums.py,sha256=LF-gY3EyucQSmGBxXNy8z4Cb4Ry672muPtz5Xt7e9oo,227
+documente_shared/domain/repositories.py,sha256=h_nArptP5Xh7Jrhden-Ii9rZQqzgLFHhuzj1p78LgmU,365
+documente_shared/infrastructure/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+documente_shared/infrastructure/dynamo_repositories.py,sha256=RrJ3T1GsH5NNbaTqBwZ4_h34nWnoQ9mhcd5VDRiiv48,521
+documente_shared/infrastructure/dynamo_table.py,sha256=H4X_SlK3DzOAt4wPQ8u4wkURt00bsKr5mfeSQTTA1O8,2094
+documente_shared/infrastructure/s3_bucket.py,sha256=nJe6OCHhycPcl21gwazC_xxqVQ6y1cRS4T3LH14Hfjk,1180
+documente_shared-0.1.3.dist-info/METADATA,sha256=1sL4e7TIZ29d2o1EEKaOkMLGdWtNMnfKicSX09H5zh0,639
+documente_shared-0.1.3.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
+documente_shared-0.1.3.dist-info/RECORD,,

documente_shared-0.1.1.dist-info/RECORD

@@ -1,5 +0,0 @@
-documente_shared/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-documente_shared/enums.py,sha256=CABjh0FRrBjoD2Q2-CwxTupnVclXi-YWYfurjnbg1kQ,210
-documente_shared-0.1.1.dist-info/METADATA,sha256=8iurH4PWXg_QYTByZwVIysyb16E4mw7bKr8s7jioRcQ,639
-documente_shared-0.1.1.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
-documente_shared-0.1.1.dist-info/RECORD,,