synapse-sdk 1.0.0a32__tar.gz → 1.0.0a33__tar.gz

{synapse_sdk-1.0.0a32/synapse_sdk.egg-info → synapse_sdk-1.0.0a33}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: synapse-sdk
-Version: 1.0.0a32
+Version: 1.0.0a33
 Summary: synapse sdk
 Author-email: datamaker <developer@datamaker.io>
 License: MIT

{synapse_sdk-1.0.0a32 → synapse_sdk-1.0.0a33}/synapse_sdk/clients/backend/__init__.py

@@ -1,6 +1,7 @@
 from synapse_sdk.clients.backend.annotation import AnnotationClientMixin
 from synapse_sdk.clients.backend.core import CoreClientMixin
 from synapse_sdk.clients.backend.dataset import DatasetClientMixin
+from synapse_sdk.clients.backend.hitl import HITLClientMixin
 from synapse_sdk.clients.backend.integration import IntegrationClientMixin
 from synapse_sdk.clients.backend.ml import MLClientMixin
 
@@ -11,6 +12,7 @@ class BackendClient(
     DatasetClientMixin,
     IntegrationClientMixin,
     MLClientMixin,
+    HITLClientMixin,
 ):
     name = 'Backend'
     token = None

{synapse_sdk-1.0.0a32 → synapse_sdk-1.0.0a33}/synapse_sdk/clients/backend/annotation.py

@@ -11,14 +11,14 @@ class AnnotationClientMixin(BaseClient):
         path = f'task_tags/{pk}/'
         return self._get(path)
 
-    def list_task_tags(self, data):
+    def list_task_tags(self, params):
         path = 'task_tags/'
-        return self._list(path, data=data)
+        return self._list(path, params=params)
 
-    def list_tasks(self, data, url_conversion=None, list_all=False):
+    def list_tasks(self, params=None, url_conversion=None, list_all=False):
         path = 'tasks/'
         url_conversion = get_default_url_conversion(url_conversion, files_fields=['files'])
-        return self._list(path, data=data, url_conversion=url_conversion, list_all=list_all)
+        return self._list(path, params=params, url_conversion=url_conversion, list_all=list_all)
 
     def create_tasks(self, data):
         path = 'tasks/'

synapse_sdk-1.0.0a33/synapse_sdk/clients/backend/hitl.py

@@ -0,0 +1,17 @@
+from synapse_sdk.clients.base import BaseClient
+from synapse_sdk.clients.utils import get_default_url_conversion
+
+
+class HITLClientMixin(BaseClient):
+    def get_assignment(self, pk):
+        path = f'assignments/{pk}/'
+        return self._get(path)
+
+    def list_assignments(self, params=None, url_conversion=None, list_all=False):
+        path = 'assignments/'
+        url_conversion = get_default_url_conversion(url_conversion, files_fields=['files'])
+        return self._list(path, params=params, url_conversion=url_conversion, list_all=list_all)
+
+    def set_tags_assignments(self, data, params=None):
+        path = 'assignments/set_tags/'
+        return self._post(path, payload=data, params=params)

{synapse_sdk-1.0.0a32 → synapse_sdk-1.0.0a33}/synapse_sdk/clients/backend/integration.py

@@ -1,3 +1,4 @@
+from synapse_sdk.clients.backend.models import Storage
 from synapse_sdk.clients.base import BaseClient
 from synapse_sdk.utils.file import convert_file_to_base64
 
@@ -79,5 +80,6 @@ class IntegrationClientMixin(BaseClient):
         return self._list(path, params=params, list_all=list_all)
 
     def get_storage(self, pk):
+        """Get specific storage data from synapse backend."""
         path = f'storages/{pk}/'
-        return self._get(path)
+        return self._get(path, pydantic_model=Storage)

synapse_sdk-1.0.0a33/synapse_sdk/clients/backend/models.py

@@ -0,0 +1,44 @@
+from enum import Enum
+from typing import Dict
+
+from pydantic import BaseModel
+
+
+class StorageCategory(str, Enum):
+    """Synapse Backend Storage Category Enum."""
+
+    INTERNAL = 'internal'
+    EXTERNAL = 'external'
+
+
+class StorageProvider(str, Enum):
+    """Synapse Backend Storage Provider Enum."""
+
+    AMAZON_S3 = 'amazon_s3'
+    AZURE = 'azure'
+    DIGITAL_OCEAN = 'digital_ocean'
+    FILE_SYSTEM = 'file_system'
+    FTP = 'ftp'
+    SFTP = 'sftp'
+    MINIO = 'minio'
+    GCP = 'gcp'
+
+
+class Storage(BaseModel):
+    """Synapse Backend Storage Model.
+
+    Attrs:
+        id (int): The storage pk.
+        name (str): The storage name.
+        category (str): The storage category. (ex: internal, external)
+        provider (str): The storage provider. (ex: s3, gcp)
+        configuration (Dict): The storage configuration.
+        is_default (bool): The storage is default for Synapse backend workspace.
+    """
+
+    id: int
+    name: str
+    category: StorageCategory
+    provider: StorageProvider
+    configuration: Dict
+    is_default: bool

{synapse_sdk-1.0.0a32 → synapse_sdk-1.0.0a33}/synapse_sdk/clients/base.py

@@ -85,26 +85,59 @@ class BaseClient:
         except ValueError:
             return response.text
 
-    def _get(self, path, url_conversion=None, **kwargs):
+    def _get(self, path, url_conversion=None, pydantic_model=None, **kwargs):
+        """
+        Perform a GET request and optionally convert response to a pydantic model.
+
+        Args:
+            path (str): URL path to request.
+            url_conversion (dict, optional): Configuration for URL to path conversion.
+            pydantic_model (Type, optional): Pydantic model to convert the response to.
+            **kwargs: Additional keyword arguments to pass to the request.
+
+        Returns:
+            The response data, optionally converted to a pydantic model.
+        """
         response = self._request('get', path, **kwargs)
+
         if url_conversion:
             if url_conversion['is_list']:
                 files_url_to_path_from_objs(response['results'], **url_conversion, is_async=True)
             else:
                 files_url_to_path_from_objs(response, **url_conversion)
+
+        if pydantic_model:
+            return self._validate_response_with_pydantic_model(response, pydantic_model)
+
         return response
 
-    def _post(self, path, **kwargs):
-        return self._request('post', path, **kwargs)
+    def _post(self, path, pydantic_model=None, **kwargs):
+        response = self._request('post', path, **kwargs)
+        if pydantic_model:
+            return self._validate_response_with_pydantic_model(response, pydantic_model)
+        else:
+            return response
 
-    def _put(self, path, **kwargs):
-        return self._request('put', path, **kwargs)
+    def _put(self, path, pydantic_model=None, **kwargs):
+        response = self._request('put', path, **kwargs)
+        if pydantic_model:
+            return self._validate_response_with_pydantic_model(response, pydantic_model)
+        else:
+            return response
 
-    def _patch(self, path, **kwargs):
-        return self._request('patch', path, **kwargs)
+    def _patch(self, path, pydantic_model=None, **kwargs):
+        response = self._request('patch', path, **kwargs)
+        if pydantic_model:
+            return self._validate_response_with_pydantic_model(response, pydantic_model)
+        else:
+            return response
 
-    def _delete(self, path, **kwargs):
-        return self._request('delete', path, **kwargs)
+    def _delete(self, path, pydantic_model=None, **kwargs):
+        response = self._request('delete', path, **kwargs)
+        if pydantic_model:
+            return self._validate_response_with_pydantic_model(response, pydantic_model)
+        else:
+            return response
 
     def _list(self, path, url_conversion=None, list_all=False, **kwargs):
         response = self._get(path, **kwargs)
@@ -121,3 +154,17 @@ class BaseClient:
 
     def exists(self, api, *args, **kwargs):
         return getattr(self, api)(*args, **kwargs)['count'] > 0
+
+    def _validate_response_with_pydantic_model(self, response, pydantic_model):
+        """Validate a response with a pydantic model."""
+        # Check if model is a pydantic model (has the __pydantic_model__ attribute)
+        if (
+            hasattr(pydantic_model, '__pydantic_model__')
+            or hasattr(pydantic_model, 'model_validate')
+            or hasattr(pydantic_model, 'parse_obj')
+        ):
+            pydantic_model.model_validate(response)
+            return response
+        else:
+            # Not a pydantic model
+            raise TypeError('The provided model is not a pydantic model')

synapse_sdk-1.0.0a33/synapse_sdk/plugins/categories/export/actions/export.py

@@ -0,0 +1,222 @@
+from abc import ABC, abstractmethod
+from typing import Any, Literal
+
+from pydantic import BaseModel, field_validator
+from pydantic_core import PydanticCustomError
+
+from synapse_sdk.clients.exceptions import ClientError
+from synapse_sdk.i18n import gettext as _
+from synapse_sdk.plugins.categories.base import Action
+from synapse_sdk.plugins.categories.decorators import register_action
+from synapse_sdk.plugins.enums import PluginCategory, RunMethod
+from synapse_sdk.utils.storage import get_pathlib
+
+
+class ExportTargetHandler(ABC):
+    """
+    Abstract base class for handling export targets.
+
+    This class defines the blueprint for export target handlers, requiring the implementation
+    of methods to validate filters, retrieve results, and process collections of results.
+    """
+
+    @abstractmethod
+    def validate_filter(self, value: dict, client: Any):
+        """
+        Validate filter query params to request original data from api.
+
+        Args:
+            value (dict): The filter criteria to validate.
+            client (Any): The client used to validate the filter.
+
+        Raises:
+            PydanticCustomError: If the filter criteria are invalid.
+
+        Returns:
+            dict: The validated filter criteria.
+        """
+        pass
+
+    @abstractmethod
+    def get_results(self, client: Any, filters: dict):
+        """
+        Retrieve original data from target sources.
+
+        Args:
+            client (Any): The client used to retrieve the results.
+            filters (dict): The filter criteria to apply.
+
+        Returns:
+            tuple: A tuple containing the results and the total count of results.
+        """
+        pass
+
+    @abstractmethod
+    def get_export_item(self, results):
+        """
+        Providing elements to build export data.
+
+        Args:
+            results (list): The results to process.
+
+        Yields:
+            generator: A generator that yields processed data items.
+        """
+        pass
+
+
+class AssignmentExportTargetHandler(ExportTargetHandler):
+    def validate_filter(self, value: dict, client: Any):
+        if 'project' not in value:
+            raise PydanticCustomError('missing_field', _('Project is required for Assignment.'))
+        try:
+            client.list_assignments(params=value)
+        except ClientError:
+            raise PydanticCustomError('client_error', _('Unable to get Assignment.'))
+        return value
+
+    def get_results(self, client: Any, filters: dict):
+        return client.list_assignments(params=filters, list_all=True)
+
+    def get_export_item(self, results):
+        for result in results:
+            yield {
+                'data': result['data'],
+                'files': result['file'],
+                'id': result['id'],
+            }
+
+
+class GroundTruthExportTargetHandler(ExportTargetHandler):
+    def validate_filter(self, value: dict, client: Any):
+        if 'ground_truth_dataset_version' not in value:
+            raise PydanticCustomError('missing_field', _('Ground Truth dataset version is required.'))
+        try:
+            client.get_ground_truth_version(value['ground_truth_dataset_version'])
+        except ClientError:
+            raise PydanticCustomError('client_error', _('Unable to get Ground Truth dataset version.'))
+        return value
+
+    def get_results(self, client: Any, filters: dict):
+        filters['ground_truth_dataset_versions'] = filters.pop('ground_truth_dataset_version')
+        return client.list_ground_truth_events(params=filters, list_all=True)
+
+    def get_export_item(self, results):
+        for result in results:
+            files_key = next(iter(result['data_unit']['files']))
+            yield {
+                'data': result['data'],
+                'files': result['data_unit']['files'][files_key],
+                'id': result['ground_truth'],
+            }
+
+
+class TaskExportTargetHandler(ExportTargetHandler):
+    def validate_filter(self, value: dict, client: Any):
+        if 'project' not in value:
+            raise PydanticCustomError('missing_field', _('Project is required for Task.'))
+        try:
+            client.list_tasks(params=value)
+        except ClientError:
+            raise PydanticCustomError('client_error', _('Unable to get Task.'))
+        return value
+
+    def get_results(self, client: Any, filters: dict):
+        filters['expand'] = 'data_unit'
+        return client.list_tasks(params=filters, list_all=True)
+
+    def get_export_item(self, results):
+        for result in results:
+            files_key = next(iter(result['data_unit']['files']))
+            yield {
+                'data': result['data'],
+                'files': result['data_unit']['files'][files_key],
+                'id': result['id'],
+            }
+
+
+class TargetHandlerFactory:
+    @staticmethod
+    def get_handler(target: str) -> ExportTargetHandler:
+        if target == 'assignment':
+            return AssignmentExportTargetHandler()
+        elif target == 'ground_truth':
+            return GroundTruthExportTargetHandler()
+        elif target == 'task':
+            return TaskExportTargetHandler()
+        else:
+            raise ValueError(f'Unknown target: {target}')
+
+
+class ExportParams(BaseModel):
+    """
+    Parameters for the export action.
+
+    Attributes:
+        storage (int): The storage ID to save the exported data.
+        save_original_file (bool): Whether to save the original file.
+        path (str): The path to save the exported data.
+        target (str): The target source to export data from. (ex. ground_truth, assignment, task)
+        filter (dict): The filter criteria to apply.
+    """
+
+    storage: int
+    save_original_file: bool = True
+    path: str
+    target: Literal['assignment', 'ground_truth', 'task']
+    filter: dict
+
+    @field_validator('storage')
+    @staticmethod
+    def check_storage_exists(value, info):
+        action = info.context['action']
+        client = action.client
+        try:
+            client.get_storage(value)
+        except ClientError:
+            raise PydanticCustomError('client_error', _('Unable to get storage from Synapse backend.'))
+        return value
+
+    @field_validator('filter')
+    @staticmethod
+    def check_filter_by_target(value, info):
+        action = info.context['action']
+        client = action.client
+        target = action.params['target']
+        handler = TargetHandlerFactory.get_handler(target)
+        return handler.validate_filter(value, client)
+
+
+@register_action
+class ExportAction(Action):
+    name = 'export'
+    category = PluginCategory.EXPORT
+    method = RunMethod.JOB
+    params_model = ExportParams
+    progress_categories = {
+        'dataset_conversion': {
+            'proportion': 100,
+        }
+    }
+
+    def get_filtered_results(self, filters, handler):
+        """Get filtered target results."""
+        try:
+            result_list = handler.get_results(self.client, filters)
+            results = result_list[0]
+            count = result_list[1]
+        except ClientError:
+            raise PydanticCustomError('client_error', _('Unable to get Ground Truth dataset.'))
+        return results, count
+
+    def start(self):
+        filters = {'expand': 'data', **self.params['filter']}
+        target = self.params['target']
+        handler = TargetHandlerFactory.get_handler(target)
+
+        self.params['results'], self.params['count'] = self.get_filtered_results(filters, handler)
+        export_items = handler.get_export_item(self.params['results'])
+
+        storage = self.client.get_storage(self.params['storage'])
+        pathlib_cwd = get_pathlib(storage, self.params['path'])
+        return self.entrypoint(self.run, export_items, pathlib_cwd, **self.params)

{synapse_sdk-1.0.0a32 → synapse_sdk-1.0.0a33}/synapse_sdk/plugins/categories/export/templates/plugin/export.py

@@ -4,12 +4,12 @@ from pathlib import Path
 import requests
 
 
-def export(run, input_dataset, path_root, **params):
+def export(run, export_items, path_root, **params):
     """Executes the export task.
 
     Args:
         run : Execution object
-        input_dataset (generator):
+        export_items (generator):
             - data (dict): dm_schema_data information.
             - files (dict): File information. Includes file URL, original file path, metadata, etc.
             - id (int): ground_truth ID
@@ -42,11 +42,11 @@ def export(run, input_dataset, path_root, **params):
     total = params['count']
     # progress init
     run.set_progress(0, total, category='dataset_conversion')
-    for no, input_data in enumerate(input_dataset, start=1):
+    for no, export_item in enumerate(export_items, start=1):
         run.set_progress(no, total, category='dataset_conversion')
         if no == 1:
             run.log_message('Converting dataset.')
-        preprocessed_data = before_convert(input_data)
+        preprocessed_data = before_convert(export_item)
         converted_data = convert_data(preprocessed_data)
         final_data = after_convert(converted_data)
 
@@ -97,7 +97,7 @@ def get_original_file_pathlib(files):
     Returns:
         pathlib.Path: The original file path extracted from the metadata.
     """
-    return Path(next(iter(files.values()))['meta']['path_original'])
+    return Path(files['meta']['path_original'])
 
 
 def save_original_file(result, base_path, error_file_list):
@@ -108,7 +108,7 @@ def save_original_file(result, base_path, error_file_list):
         base_path (Path): The directory where the file will be saved.
         error_file_list (list): A list to store error files.
     """
-    file_url = next(iter(result['files'].values()))['url']
+    file_url = result['files']['url']
     file_name = get_original_file_pathlib(result['files']).name
     response = requests.get(file_url)
     try:

{synapse_sdk-1.0.0a32 → synapse_sdk-1.0.0a33}/synapse_sdk/plugins/categories/smart_tool/templates/config.yaml

@@ -6,3 +6,5 @@ actions:
     entrypoint: plugin.auto_label.MyAutoLabel
     model:
       neural_net__plugin__code: sam2
+    ui_schema: |
+      Dumped FormKit Schema for auto_label action

{synapse_sdk-1.0.0a32 → synapse_sdk-1.0.0a33}/synapse_sdk/utils/storage/__init__.py

@@ -6,13 +6,21 @@ from synapse_sdk.utils.storage.registry import STORAGE_PROVIDERS
 
 
 def get_storage(connection_param: str | dict):
+    """Get storage class with connection param.
+
+    Args:
+        connection_param (str | dict): The connection param for the Storage provider.
+
+    Returns:
+        BaseStorage: The storage class object with connection param.
+    """
     storage_scheme = None
     if isinstance(connection_param, dict):
         storage_scheme = connection_param['provider']
     else:
         storage_scheme = urlparse(connection_param).scheme
 
-    assert storage_scheme in STORAGE_PROVIDERS.keys(), _('지원하지 않는 저장소입니다.')
+    assert storage_scheme in STORAGE_PROVIDERS.keys(), _('Storage provider not supported.')
     return STORAGE_PROVIDERS[storage_scheme](connection_param)
 
 

{synapse_sdk-1.0.0a32 → synapse_sdk-1.0.0a33/synapse_sdk.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: synapse-sdk
-Version: 1.0.0a32
+Version: 1.0.0a33
 Summary: synapse sdk
 Author-email: datamaker <developer@datamaker.io>
 License: MIT

{synapse_sdk-1.0.0a32 → synapse_sdk-1.0.0a33}/synapse_sdk.egg-info/SOURCES.txt

@@ -49,8 +49,10 @@ synapse_sdk/clients/backend/__init__.py
 synapse_sdk/clients/backend/annotation.py
 synapse_sdk/clients/backend/core.py
 synapse_sdk/clients/backend/dataset.py
+synapse_sdk/clients/backend/hitl.py
 synapse_sdk/clients/backend/integration.py
 synapse_sdk/clients/backend/ml.py
+synapse_sdk/clients/backend/models.py
 synapse_sdk/clients/ray/__init__.py
 synapse_sdk/clients/ray/core.py
 synapse_sdk/clients/ray/serve.py
@@ -143,5 +145,8 @@ synapse_sdk/utils/storage/providers/gcp.py
 synapse_sdk/utils/storage/providers/s3.py
 synapse_sdk/utils/storage/providers/sftp.py
 tests/__init__.py
-tests/utils/__init__.py
+tests/clients/__init__.py
+tests/clients/test_backend_models.py
+tests/clients/test_base_client.py
+tests/utils/__init__,py
 tests/utils/test_debug.py

synapse_sdk-1.0.0a33/tests/clients/test_backend_models.py

@@ -0,0 +1,74 @@
+import pytest
+from pydantic import ValidationError
+
+from synapse_sdk.clients.backend.models import Storage, StorageCategory, StorageProvider
+
+
+def test_storage_model_validation_success():
+    storage_response = {
+        'id': 1,
+        'name': 'test_storage',
+        'category': 'internal',
+        'provider': 'file_system',
+        'configuration': {},
+        'is_default': True,
+    }
+    storage = Storage(**storage_response)
+    assert storage.id == 1
+
+
+def test_storage_model_validation_failed_with_invalid_category():
+    storage_response = {
+        'id': 1,
+        'name': 'test_storage',
+        'category': 'invalid_data',
+        'provider': 'file_system',
+        'configuration': {},
+        'is_default': True,
+    }
+    with pytest.raises(ValidationError) as exc_info:
+        Storage(**storage_response)
+    assert 'category' in str(exc_info.value)
+
+
+def test_storage_model_validation_failed_with_invalid_provider():
+    storage_response = {
+        'id': 1,
+        'name': 'test_storage',
+        'category': 'internal',
+        'provider': 'invalid_provider',
+        'configuration': {},
+        'is_default': True,
+    }
+    with pytest.raises(ValidationError) as exc_info:
+        Storage(**storage_response)
+    assert 'provider' in str(exc_info.value)
+
+
+def test_storage_model_validation_failed_with_missing_field():
+    storage_response = {
+        'id': 1,
+        'name': 'test_storage',
+        'category': 'internal',
+        'provider': 'file_system',
+        'is_default': True,
+        # Missing configuration field
+    }
+    with pytest.raises(ValidationError):
+        Storage(**storage_response)
+
+
+def test_storage_model_enum_values():
+    # Test that valid enum values work correctly
+    for category in StorageCategory:
+        for provider in StorageProvider:
+            storage = Storage(
+                id=1,
+                name='test_storage',
+                category=category,
+                provider=provider,
+                configuration={},
+                is_default=True,
+            )
+            assert storage.category == category
+            assert storage.provider == provider

synapse_sdk-1.0.0a33/tests/clients/test_base_client.py

@@ -0,0 +1,46 @@
+import pytest
+from pydantic import ValidationError
+
+from synapse_sdk.clients.backend.models import Storage
+from synapse_sdk.clients.base import BaseClient
+
+
+@pytest.fixture
+def base_client():
+    return BaseClient('http://fake_url')
+
+
+@pytest.fixture
+def valid_storage_response():
+    return {
+        'id': 1,
+        'name': 'test_storage',
+        'category': 'internal',
+        'provider': 'file_system',
+        'configuration': {},
+        'is_default': True,
+    }
+
+
+@pytest.fixture
+def invalid_storage_response(valid_storage_response):
+    response = valid_storage_response.copy()
+    response['provider'] = 'invalid_provider'
+    return response
+
+
+def test_validate_response_with_pydantic_model_success(base_client, valid_storage_response):
+    validated_response = base_client._validate_response_with_pydantic_model(valid_storage_response, Storage)
+    assert validated_response['id'] == valid_storage_response['id']
+
+
+def test_validate_response_with_pydantic_model_invalid_data(base_client, invalid_storage_response):
+    with pytest.raises(ValidationError) as exc_info:
+        base_client._validate_response_with_pydantic_model(invalid_storage_response, Storage)
+    assert '1 validation error' in str(exc_info.value)
+
+
+def test_validate_response_with_pydantic_model_not_pydantic_model(base_client, valid_storage_response):
+    with pytest.raises(TypeError) as exc_info:
+        base_client._validate_response_with_pydantic_model(valid_storage_response, {})
+    assert 'The provided model is not a pydantic model' in str(exc_info.value)