mage-ai 0.9.70__py3-none-any.whl → 0.9.71__py3-none-any.whl

mage_ai/data_preparation/models/pipeline.py

@@ -11,12 +11,13 @@ from typing import Any, Callable, Dict, List, Optional, Tuple, Union
 
 import aiofiles
 import pytz
-import yaml
 from jinja2 import Template
 
 from mage_ai.authentication.permissions.constants import EntityName
 from mage_ai.cache.block import BlockCache
 from mage_ai.cache.pipeline import PipelineCache
+from mage_ai.data.constants import InputDataType
+from mage_ai.data.tabular.models import BatchSettings
 from mage_ai.data_preparation.models.block import Block, run_blocks, run_blocks_sync
 from mage_ai.data_preparation.models.block.block_factory import BlockFactory
 from mage_ai.data_preparation.models.block.data_integration.utils import (
@@ -27,6 +28,10 @@ from mage_ai.data_preparation.models.block.dynamic.utils import (
     is_dynamic_block_child,
 )
 from mage_ai.data_preparation.models.block.errors import HasDownstreamDependencies
+from mage_ai.data_preparation.models.block.extension.utils import compare_extension
+from mage_ai.data_preparation.models.block.settings.variables.models import (
+    ChunkKeyTypeUnion,
+)
 from mage_ai.data_preparation.models.constants import (
     DATA_INTEGRATION_CATALOG_FILE,
     PIPELINE_CONFIG_FILE,
@@ -47,7 +52,10 @@ from mage_ai.data_preparation.models.file import File
 from mage_ai.data_preparation.models.pipelines.models import PipelineSettings
 from mage_ai.data_preparation.models.project import Project
 from mage_ai.data_preparation.models.project.constants import FeatureUUID
-from mage_ai.data_preparation.models.utils import is_yaml_serializable
+from mage_ai.data_preparation.models.utils import (
+    is_yaml_serializable,
+    warn_for_repo_path,
+)
 from mage_ai.data_preparation.models.variable import Variable
 from mage_ai.data_preparation.repo_manager import (
     RepoConfig,
@@ -73,6 +81,7 @@ from mage_ai.shared.io import safe_write, safe_write_async
 from mage_ai.shared.path_fixer import remove_base_repo_path
 from mage_ai.shared.strings import format_enum
 from mage_ai.shared.utils import clean_name
+from mage_ai.shared.yaml import load_yaml, yaml
 
 CYCLE_DETECTION_ERR_MESSAGE = 'A cycle was detected in this pipeline'
 
@@ -81,10 +90,11 @@ class Pipeline:
     def __init__(
         self,
         uuid,
-        repo_path=None,
+        repo_path: str = None,
         config=None,
         repo_config=None,
         catalog=None,
+        context_data: Dict = None,
         use_repo_path: bool = False,
         description: str = None,
         tags: List[str] = None,
@@ -103,6 +113,14 @@ class Pipeline:
         self.extensions = {}
         self.name = None
         self.notification_config = dict()
+
+        # For multi project
+        warn_for_repo_path(repo_path)
+        self.context_data = context_data
+        if self.context_data is None:
+            self.context_data = dict()
+        self._project = None
+
         self.repo_path = repo_path or get_repo_path()
         self.retry_config = {}
         self.run_pipeline_in_one_process = False
@@ -130,17 +148,28 @@ class Pipeline:
         # Used for showing the operation history. For example: recently viewed pipelines.
         self.history = []
 
+        # Path of the pipeline metadata.yaml file
+        self._config_path = None
         if config is None:
             self.load_config_from_yaml()
         else:
             self.load_config(config, catalog=catalog)
 
-    @classmethod
-    def build_config_path(self, uuid: str, repo_path: str, use_repo_path: bool = False) -> str:
+    def build_config_path(
+        self,
+        uuid: str,
+        repo_path: str,
+        context_data: Dict = None,
+        use_repo_path: bool = False,
+    ) -> str:
         if project_platform_activated() and not use_repo_path:
             from mage_ai.settings.platform.utils import get_pipeline_config_path
 
-            config_path, _repo_path = get_pipeline_config_path(uuid)
+            config_path, _repo_path = get_pipeline_config_path(
+                uuid,
+                context_data=context_data,
+                repo_path=repo_path,
+            )
             if config_path:
                 return config_path
 
@@ -153,14 +182,25 @@ class Pipeline:
 
     @property
     def config_path(self):
-        return self.build_config_path(self.uuid, self.repo_path, use_repo_path=self.use_repo_path)
+        if not self._config_path:
+            self._config_path = self.build_config_path(
+                self.uuid,
+                self.repo_path,
+                context_data=self.context_data,
+                use_repo_path=self.use_repo_path,
+            )
+        return self._config_path
 
     @property
     def catalog_config_path(self):
         if project_platform_activated() and not self.use_repo_path:
             from mage_ai.settings.platform.utils import get_pipeline_config_path
 
-            config_path, _repo_path = get_pipeline_config_path(self.uuid)
+            config_path, _repo_path = get_pipeline_config_path(
+                self.uuid,
+                context_data=self.context_data,
+                repo_path=self.repo_path,
+            )
             if config_path:
                 return os.path.join(os.path.dirname(config_path), DATA_INTEGRATION_CATALOG_FILE)
 
@@ -198,6 +238,16 @@ class Pipeline:
             self.uuid,
         )
 
+    @property
+    def project(self):
+        if self._project is not None:
+            return self._project
+
+        from mage_ai.data_preparation.models.project import Project
+
+        self._project = Project(context_data=self.context_data, repo_config=self.repo_config)
+        return self._project
+
     @property
     def remote_variables_dir(self):
         remote_variables_dir = self.repo_config.remote_variables_dir
@@ -216,18 +266,20 @@ class Pipeline:
 
     @property
     def all_block_configs(self) -> List[Dict]:
-        return self.block_configs + \
-            self.conditional_configs + \
-            self.callback_configs + \
-            self.widget_configs
+        return (
+            self.block_configs
+            + self.conditional_configs
+            + self.callback_configs
+            + self.widget_configs
+        )
 
     @classmethod
     def create(
         self,
         name: str,
+        repo_path: str = None,
         description: str = None,
         pipeline_type: PipelineType = PipelineType.PYTHON,
-        repo_path: str = None,
         tags: List[str] = None,
     ):
         """
@@ -235,6 +287,8 @@ class Pipeline:
         2. Create a new yaml file to store pipeline config
         3. Create other files: requirements.txt, __init__.py
         """
+        warn_for_repo_path(repo_path)
+
         uuid = clean_name(name)
         pipeline_path = os.path.join(repo_path, PIPELINES_FOLDER, uuid)
         if os.path.exists(pipeline_path):
@@ -243,14 +297,17 @@ class Pipeline:
         copy_template_directory('pipeline', pipeline_path)
         # Update metadata.yaml with pipeline config
         with open(os.path.join(pipeline_path, PIPELINE_CONFIG_FILE), 'w') as fp:
-            yaml.dump(dict(
-                created_at=str(datetime.now(tz=pytz.UTC)),
-                description=description,
-                name=name,
-                tags=tags or [],
-                uuid=uuid,
-                type=format_enum(pipeline_type or PipelineType.PYTHON),
-            ), fp)
+            yaml.dump(
+                dict(
+                    created_at=str(datetime.now(tz=pytz.UTC)),
+                    description=description,
+                    name=name,
+                    tags=tags or [],
+                    uuid=uuid,
+                    type=format_enum(pipeline_type or PipelineType.PYTHON),
+                ),
+                fp,
+            )
 
         pipeline = Pipeline(
             uuid,
@@ -269,10 +326,13 @@ class Pipeline:
                 zip_size = sum(e.file_size for e in zipf.infolist())  # calc zip size in bytes
                 if zip_size / 1000 > PIPELINE_MAX_FILE_SIZE:  # prevention against zip-bombs
                     raise PipelineZipTooLargeError(
-                        f'Pipeline zip exceeds size limit {PIPELINE_MAX_FILE_SIZE/1000}Kb')
+                        f'Pipeline zip exceeds size limit {PIPELINE_MAX_FILE_SIZE / 1000}Kb'
+                    )
 
                 # Ignore `__MACOSX` for zips created on macOS systems
-                zip_contents = [path for path in zipf.namelist() if not path.startswith('__MACOSX')]
+                zip_contents = [
+                    path for path in zipf.namelist() if not path.startswith('__MACOSX')
+                ]
                 # Verify if zip contents are part of a root folder
                 prefix = os.path.commonpath(zip_contents)
 
@@ -302,7 +362,9 @@ class Pipeline:
                     raise FileWriteError(f'Failed to write pipeline file to {destination}.')
 
             # return the pipeline configuration file
-            config_destination_path = pipeline_files[0][1]  # First item is the pipeline config path
+            config_destination_path = pipeline_files[0][
+                1
+            ]  # First item is the pipeline config path
             ret_file = File.from_path(config_destination_path)
             ret_file.filename = new_pipeline_uuid
 
@@ -332,8 +394,10 @@ class Pipeline:
             repo_path=source_pipeline.repo_path,
         )
 
-        if source_pipeline.type == PipelineType.INTEGRATION and \
-                source_pipeline.data_integration is not None:
+        if (
+            source_pipeline.type == PipelineType.INTEGRATION
+            and source_pipeline.data_integration is not None
+        ):
             with open(duplicate_pipeline.catalog_config_path, 'w') as fp:
                 json.dump(source_pipeline.data_integration, fp)
 
@@ -358,7 +422,7 @@ class Pipeline:
 
             block_cache = await BlockCache.initialize_cache()
             for block in blocks:
-                block_cache.add_pipeline(block, duplicate_pipeline)
+                block_cache.add_pipeline(block, duplicate_pipeline, duplicate_pipeline.repo_path)
 
         return cls.get(
             duplicate_pipeline_uuid,
@@ -377,30 +441,46 @@ class Pipeline:
 
     @classmethod
     def get(
-        self,
+        cls,
         uuid,
         repo_path: str = None,
+        repo_config=None,
         check_if_exists: bool = False,
         all_projects: bool = False,
+        context_data: Dict = None,
         use_repo_path: bool = False,
     ):
-        from mage_ai.data_preparation.models.pipelines.integration_pipeline import (
-            IntegrationPipeline,
-        )
+        warn_for_repo_path(repo_path)
 
-        config_path, repo_path = self._get_config_path(
+        config_path, repo_path = cls._get_config_path(
             uuid,
             repo_path=repo_path,
             all_projects=all_projects,
+            context_data=context_data,
             use_repo_path=use_repo_path,
         )
 
-        if check_if_exists and not os.path.exists(config_path):
+        if check_if_exists and (not config_path or not os.path.exists(config_path)):
             return None
 
-        pipeline = self(uuid, repo_path=repo_path, use_repo_path=use_repo_path)
+        pipeline = cls(
+            uuid,
+            repo_path=repo_path,
+            repo_config=repo_config,
+            context_data=context_data,
+            use_repo_path=use_repo_path,
+        )
         if PipelineType.INTEGRATION == pipeline.type:
-            pipeline = IntegrationPipeline(uuid, repo_path=repo_path)
+            from mage_ai.data_preparation.models.pipelines.integration_pipeline import (
+                IntegrationPipeline,
+            )
+
+            pipeline = IntegrationPipeline(
+                uuid,
+                repo_path=repo_path,
+                repo_config=repo_config,
+                context_data=context_data,
+            )
 
         return pipeline
 
@@ -425,7 +505,7 @@ class Pipeline:
             return None
 
         with open(metadata_path) as fp:
-            config = yaml.full_load(fp) or {}
+            config = load_yaml(fp) or {}
         return config
 
     @classmethod
@@ -434,12 +514,17 @@ class Pipeline:
         uuid,
         repo_path: str = None,
         all_projects: bool = False,
+        context_data: Dict = None,
         use_repo_path: bool = False,
     ) -> Tuple[str, str]:
         if all_projects and not use_repo_path and project_platform_activated():
             from mage_ai.settings.platform.utils import get_pipeline_config_path
 
-            config_path, repo_path = get_pipeline_config_path(uuid)
+            config_path, repo_path = get_pipeline_config_path(
+                uuid,
+                context_data=context_data,
+                repo_path=repo_path,
+            )
         else:
             repo_path = repo_path or get_repo_path()
             config_path = os.path.join(
@@ -479,7 +564,7 @@ class Pipeline:
 
             config = None
             async with aiofiles.open(config_path, mode='r') as f:
-                config = yaml.safe_load(await f.read()) or {}
+                config = load_yaml(await f.read()) or {}
         except Exception as e:
             if raise_exception:
                 raise e
@@ -496,16 +581,19 @@ class Pipeline:
         uuid,
         repo_path: str = None,
         all_projects: bool = False,
+        context_data: Dict = None,
         use_repo_path: bool = False,
     ):
-        from mage_ai.data_preparation.models.pipelines.integration_pipeline import (
-            IntegrationPipeline,
-        )
+        warn_for_repo_path(repo_path)
 
         if all_projects and not use_repo_path and project_platform_activated():
             from mage_ai.settings.platform.utils import get_pipeline_config_path
 
-            config_path, repo_path = get_pipeline_config_path(uuid)
+            config_path, repo_path = get_pipeline_config_path(
+                uuid,
+                context_data=context_data,
+                repo_path=repo_path,
+            )
         else:
             repo_path = repo_path or get_repo_path()
             config_path = os.path.join(
@@ -515,12 +603,16 @@ class Pipeline:
                 PIPELINE_CONFIG_FILE,
             )
 
-        if not os.path.exists(config_path):
+        if not config_path or not os.path.exists(config_path):
             raise Exception(f'Pipeline {uuid} does not exist.')
         async with aiofiles.open(config_path, mode='r', encoding='utf-8') as f:
-            config = yaml.safe_load(await f.read()) or {}
+            config = load_yaml(await f.read()) or {}
 
         if PipelineType.INTEGRATION == config.get('type'):
+            from mage_ai.data_preparation.models.pipelines.integration_pipeline import (
+                IntegrationPipeline,
+            )
+
             catalog = None
             catalog_config_path = os.path.join(
                 repo_path,
@@ -554,9 +646,15 @@ class Pipeline:
         **kwargs,
     ) -> Union[List[str], List[Tuple[str, str]]]:
         if project_platform_activated():
-            repo_paths = [d.get(
-                'full_path',
-            ) for d in build_repo_path_for_all_projects(mage_projects_only=True).values()]
+            repo_paths = [
+                d.get(
+                    'full_path',
+                )
+                for d in build_repo_path_for_all_projects(
+                    context_data=kwargs.get('context_data'),
+                    mage_projects_only=True
+                ).values()
+            ]
 
             return Pipeline.get_all_pipelines(
                 *args,
@@ -572,6 +670,7 @@ class Pipeline:
         repo_paths: List[str] = None,
         disable_pipelines_folder_creation: bool = False,
         include_repo_path: bool = False,
+        **kwargs,
     ) -> Union[List[str], List[Tuple[str, str]]]:
         arr = []
 
@@ -581,7 +680,7 @@ class Pipeline:
         if repo_paths:
             paths.extend(repo_paths)
 
-        for path in paths:
+        for path in set(paths):
             pipelines_folder = os.path.join(path, PIPELINES_FOLDER)
             pipelines_folder_exists = os.path.exists(pipelines_folder)
             if not pipelines_folder_exists and not disable_pipelines_folder_creation:
@@ -599,7 +698,10 @@ class Pipeline:
         return arr
 
     @classmethod
-    def get_pipelines_by_block(self, block, repo_path=None, widget=False) -> List['Pipeline']:
+    def get_pipelines_by_block(
+        self, block, repo_path: str = None, widget=False
+    ) -> List['Pipeline']:
+        warn_for_repo_path(repo_path)
         repo_path = repo_path or get_repo_path()
         pipelines_folder = os.path.join(repo_path, PIPELINES_FOLDER)
         pipelines = []
@@ -690,6 +792,7 @@ class Pipeline:
             from mage_ai.data_preparation.executors.streaming_pipeline_executor import (
                 StreamingPipelineExecutor,
             )
+
             StreamingPipelineExecutor(self).execute(
                 build_block_output_stdout=build_block_output_stdout,
                 global_vars=global_vars,
@@ -721,7 +824,7 @@ class Pipeline:
         if not os.path.exists(self.config_path):
             raise Exception(f'Pipeline {self.uuid} does not exist in repo_path {self.repo_path}.')
         with open(self.config_path, encoding='utf-8') as fp:
-            config = yaml.full_load(fp) or {}
+            config = load_yaml(fp) or {}
         return config
 
     def get_catalog_from_json(self):
@@ -790,7 +893,9 @@ class Pipeline:
 
             language = c.get('language')
 
-            return BlockFactory.block_class_from_type(block_type, language=language, pipeline=self)(
+            return BlockFactory.block_class_from_type(
+                block_type, language=language, pipeline=self
+            )(
                 c.get('name'),
                 c.get('uuid'),
                 block_type,
@@ -839,17 +944,28 @@ class Pipeline:
 
         for extension_uuid, extension_config in config.get('extensions', {}).items():
             extension_configs = extension_config.get('blocks') or []
-            extension_blocks = [build_shared_args_kwargs(merge_dict(c, dict(
-                extension_uuid=extension_uuid,
-            ))) for c in extension_configs]
+            extension_blocks = [
+                build_shared_args_kwargs(
+                    merge_dict(
+                        c,
+                        dict(
+                            extension_uuid=extension_uuid,
+                        ),
+                    )
+                )
+                for c in extension_configs
+            ]
 
-            self.extensions[extension_uuid] = merge_dict(extension_config, dict(
-                blocks_by_uuid=self.__initialize_blocks_by_uuid(
-                    extension_configs,
-                    extension_blocks,
-                    all_blocks,
+            self.extensions[extension_uuid] = merge_dict(
+                extension_config,
+                dict(
+                    blocks_by_uuid=self.__initialize_blocks_by_uuid(
+                        extension_configs,
+                        extension_blocks,
+                        all_blocks,
+                    ),
                 ),
-            ))
+            )
 
         blocks_with_callbacks = {}
         for callback_block in self.callbacks_by_uuid.values():
@@ -883,11 +999,13 @@ class Pipeline:
         for b in configs:
             block = blocks_by_uuid[b['uuid']]
             block.downstream_blocks = [
-                all_blocks_by_uuid[uuid] for uuid in b.get('downstream_blocks', [])
+                all_blocks_by_uuid[uuid]
+                for uuid in b.get('downstream_blocks', [])
                 if uuid in all_blocks_by_uuid
             ]
             block.upstream_blocks = [
-                all_blocks_by_uuid[uuid] for uuid in b.get('upstream_blocks', [])
+                all_blocks_by_uuid[uuid]
+                for uuid in b.get('upstream_blocks', [])
                 if uuid in all_blocks_by_uuid
             ]
 
@@ -941,7 +1059,9 @@ class Pipeline:
 
         blocks_data = [b.to_dict(**shared_kwargs) for b in self.blocks_by_uuid.values()]
         callbacks_data = [b.to_dict(**shared_kwargs) for b in self.callbacks_by_uuid.values()]
-        conditionals_data = [b.to_dict(**shared_kwargs) for b in self.conditionals_by_uuid.values()]
+        conditionals_data = [
+            b.to_dict(**shared_kwargs) for b in self.conditionals_by_uuid.values()
+        ]
         widgets_data = [b.to_dict(**shared_kwargs) for b in self.widgets_by_uuid.values()]
 
         data = dict(
@@ -962,13 +1082,17 @@ class Pipeline:
                             include_outputs=include_outputs,
                             include_outputs_spark=include_outputs_spark,
                             sample_count=sample_count,
-                          ) for b in extension['blocks_by_uuid'].values()
+                        )
+                        for b in extension['blocks_by_uuid'].values()
                     ]
                 extensions_data[extension_uuid] = merge_dict(
-                    ignore_keys(extension, [
-                        'blocks',
-                        'blocks_by_uuid',
-                    ]),
+                    ignore_keys(
+                        extension,
+                        [
+                            'blocks',
+                            'blocks_by_uuid',
+                        ],
+                    ),
                     dict(
                         blocks=blocks,
                     ),
@@ -992,7 +1116,10 @@ class Pipeline:
         include_extensions: bool = False,
         include_outputs: bool = False,
         include_outputs_spark: bool = False,
-        sample_count: int = None,
+        sample_count: Optional[int] = None,
+        disable_block_output_previews: bool = False,
+        exclude_blank_variable_uuids: bool = False,
+        max_results: Optional[int] = None,
     ):
         shared_kwargs = dict(
             check_if_file_exists=True,
@@ -1006,26 +1133,36 @@ class Pipeline:
         )
         if include_block_pipelines:
             shared_kwargs['block_cache'] = BlockCache()
-        blocks_data = await asyncio.gather(
-            *[b.to_dict_async(**merge_dict(shared_kwargs, dict(
-                include_block_catalog=include_block_catalog,
-                include_block_pipelines=include_block_pipelines,
-                include_outputs_spark=include_outputs_spark,
-            ))) for b in self.blocks_by_uuid.values()]
-        )
-        callbacks_data = await asyncio.gather(
-            *[b.to_dict_async(**shared_kwargs) for b in self.callbacks_by_uuid.values()]
-        )
-        conditionals_data = await asyncio.gather(
-            *[b.to_dict_async(**shared_kwargs) for b in self.conditionals_by_uuid.values()]
-        )
-        widgets_data = await asyncio.gather(
-            *[b.to_dict_async(
+        blocks_data = await asyncio.gather(*[
+            b.to_dict_async(
+                **merge_dict(
+                    shared_kwargs,
+                    dict(
+                        include_block_catalog=include_block_catalog,
+                        include_block_pipelines=include_block_pipelines,
+                        include_outputs_spark=include_outputs_spark,
+                        disable_output_preview=disable_block_output_previews,
+                        exclude_blank_variable_uuids=exclude_blank_variable_uuids,
+                        max_results=max_results,
+                    ),
+                )
+            )
+            for b in self.blocks_by_uuid.values()
+        ])
+        callbacks_data = await asyncio.gather(*[
+            b.to_dict_async(**shared_kwargs) for b in self.callbacks_by_uuid.values()
+        ])
+        conditionals_data = await asyncio.gather(*[
+            b.to_dict_async(**shared_kwargs) for b in self.conditionals_by_uuid.values()
+        ])
+        widgets_data = await asyncio.gather(*[
+            b.to_dict_async(
                 include_content=include_content,
                 include_outputs=include_outputs,
                 sample_count=sample_count,
-              ) for b in self.widgets_by_uuid.values()]
-        )
+            )
+            for b in self.widgets_by_uuid.values()
+        ])
         data = dict(
             blocks=blocks_data,
             callbacks=callbacks_data,
@@ -1038,18 +1175,22 @@ class Pipeline:
             for extension_uuid, extension in self.extensions.items():
                 blocks = []
                 if 'blocks_by_uuid' in extension:
-                    blocks = await asyncio.gather(
-                        *[b.to_dict_async(
+                    blocks = await asyncio.gather(*[
+                        b.to_dict_async(
                             include_content=include_content,
                             include_outputs=include_outputs,
                             sample_count=sample_count,
-                          ) for b in extension['blocks_by_uuid'].values()]
-                    )
+                        )
+                        for b in extension['blocks_by_uuid'].values()
+                    ])
                 extensions_data[extension_uuid] = merge_dict(
-                    ignore_keys(extension, [
-                        'blocks',
-                        'blocks_by_uuid',
-                    ]),
+                    ignore_keys(
+                        extension,
+                        [
+                            'blocks',
+                            'blocks_by_uuid',
+                        ],
+                    ),
                     dict(
                         blocks=blocks,
                     ),
@@ -1087,6 +1228,8 @@ class Pipeline:
             self.uuid = new_uuid
             new_pipeline_path = self.dir_path
             os.rename(old_pipeline_path, new_pipeline_path)
+            # Force updating the config path
+            self._config_path = None
             await self.save_async()
             transfer_related_models_for_pipeline(old_uuid, new_uuid)
 
@@ -1100,7 +1243,7 @@ class Pipeline:
             should_update_tag_cache = True
 
             cache = PipelineCache()
-            cache.move_model(dict(uuid=new_uuid), dict(uuid=old_uuid))
+            cache.move_model(dict(uuid=new_uuid), dict(uuid=old_uuid), repo_path=self.repo_path)
 
         should_save = False
 
@@ -1108,11 +1251,12 @@ class Pipeline:
             for extension_uuid, extension in data['extensions'].items():
                 if extension_uuid not in self.extensions:
                     self.extensions[extension_uuid] = {}
-                self.extensions[extension_uuid] = merge_dict(
-                    self.extensions[extension_uuid],
-                    extension,
-                )
-            should_save = True
+                if compare_extension(extension, self.extensions[extension_uuid]):
+                    self.extensions[extension_uuid] = merge_dict(
+                        self.extensions[extension_uuid],
+                        extension,
+                    )
+                    should_save = True
 
         if 'tags' in data:
             new_tags = data.get('tags', [])
@@ -1138,16 +1282,17 @@ class Pipeline:
 
         for key in [
             'cache_block_output_in_memory',
+            'concurrency_config',
             'data_integration',
             'executor_type',
             'retry_config',
             'run_pipeline_in_one_process',
         ]:
-            if key in data:
+            if key in data and data.get(key) != getattr(self, key):
                 setattr(self, key, data.get(key))
                 should_save = True
 
-        if 'settings' in data:
+        if 'settings' in data and data.get('settings') != self.settings.to_dict():
             self.settings = PipelineSettings.load(**(data.get('settings') or {}))
             should_save = True
 
@@ -1225,7 +1370,7 @@ class Pipeline:
                         )
 
                         if hooks:
-                            for hook in (hooks or []):
+                            for hook in hooks or []:
                                 output = hook.output
                                 if not output:
                                     continue
@@ -1243,25 +1388,31 @@ class Pipeline:
                         old_block_content = await block.content_async()
                         if block_data['content'] != old_block_content:
                             if cache_block_action_object is None:
-                                cache_block_action_object = \
+                                cache_block_action_object = (
                                     await BlockActionObjectCache.initialize_cache()
+                                )
 
                             await block.update_content_async(block_data['content'], widget=widget)
 
                             cache_block_action_object.update_block(block)
 
-                    if 'callback_content' in block_data \
-                            and block.callback_block:
+                    if 'callback_content' in block_data and block.callback_block:
                         await block.callback_block.update_content_async(
                             block_data['callback_content'],
                             widget=widget,
                         )
-                    if 'outputs' in block_data:
-                        if not is_dynamic_block(block) and not is_dynamic_block_child(block):
-                            await block.save_outputs_async(
-                                block_data['outputs'],
-                                override=True,
-                            )
+                    if (
+                        'outputs' in block_data
+                        and not widget
+                        and not is_dynamic_block(block)
+                        and not is_dynamic_block_child(block)
+                    ):
+                        if BlockType.SCRATCHPAD == block.type:
+                            block.delete_variables()
+                        await block.save_outputs_async(
+                            block_data['outputs'],
+                            override=True,
+                        )
 
                     name = block_data.get('name')
 
@@ -1273,7 +1424,7 @@ class Pipeline:
                         block.update(extract(block_data, ['color']))
 
                     configuration = block_data.get('configuration')
-                    if configuration:
+                    if configuration and configuration != block.configuration:
                         block.configuration = configuration
                         should_save_async = should_save_async or True
 
@@ -1283,42 +1434,50 @@ class Pipeline:
                         if name and name != block.name:
                             keys_to_update.append('name')
 
-                        if block_data.get('upstream_blocks'):
+                        upstream_blocks = block_data.get('upstream_blocks') or []
+                        if upstream_blocks != block.upstream_block_uuids:
                             keys_to_update.append('upstream_blocks')
                             block_data['upstream_blocks'] = [
-                                block_uuid_mapping.get(b, b)
-                                for b in block_data['upstream_blocks']
+                                block_uuid_mapping.get(b, b) for b in block_data['upstream_blocks']
                             ]
-
                         if len(keys_to_update) >= 1:
                             block.update(extract(block_data, keys_to_update))
+                            should_save_async = should_save_async or True
 
-                        should_save_async = should_save_async or True
                     elif name and name != block.name:
                         from mage_ai.cache.block_action_object import (
                             BlockActionObjectCache,
                         )
 
                         if cache_block_action_object is None:
-                            cache_block_action_object = \
+                            cache_block_action_object = (
                                 await BlockActionObjectCache.initialize_cache()
+                            )
                         cache_block_action_object.update_block(block, remove=True)
 
+                        new_block_uuid = clean_name(name)
+                        (
+                            old_file_path,
+                            old_file_path_relative,
+                        ) = block.build_file_path_directory()
+                        (
+                            new_file_path,
+                            new_file_path_relative,
+                        ) = block.build_file_path_directory(
+                            block_uuid=new_block_uuid,
+                        )
+
                         block_update_payload = extract(block_data, ['name'])
-                        configuration = copy.deepcopy(block_data).get('configuration', {})
-                        file_path = (configuration.get('file_source') or {}).get('path')
-                        if file_path:
-                            # Check for block name with period to avoid replacing a directory name
-                            new_file_path = file_path.replace(
-                                f'{clean_name(block.name)}.', f'{clean_name(name)}.'
-                            )
-                            configuration['file_source']['path'] = new_file_path
-                            block_update_payload['configuration'] = configuration
+                        configuration = copy.deepcopy(block_data).get('configuration', {}) or {}
+                        configuration['file_path'] = new_file_path_relative
+
+                        if not configuration.get('file_source'):
+                            configuration['file_source'] = {}
+                        configuration['file_source']['path'] = new_file_path_relative
+                        block_update_payload['configuration'] = configuration
+
                         blocks_to_remove_from_cache.append(block.to_dict())
-                        block.update(
-                            block_update_payload,
-                            detach=block_data.get('detach', False)
-                        )
+                        block.update(block_update_payload, detach=block_data.get('detach', False))
 
                         block_uuids_to_add_to_cache.append(block.uuid)
                         cache_block_action_object.update_block(block)
@@ -1336,9 +1495,9 @@ class Pipeline:
         # we need to update mage_sources.yml
         if any(
             (
-                BlockType.DBT != block.type and
-                block.language in [BlockLanguage.SQL, BlockLanguage.PYTHON, BlockLanguage.R] and
-                any(
+                BlockType.DBT != block.type
+                and block.language in [BlockLanguage.SQL, BlockLanguage.PYTHON, BlockLanguage.R]
+                and any(
                     BlockType.DBT == downstream_block.type
                     for downstream_block in block.downstream_blocks
                 )
@@ -1371,7 +1530,7 @@ class Pipeline:
                         old_uuid,
                         self.repo_path,
                     )
-                cache.update_pipeline(block.to_dict(), self)
+                cache.update_pipeline(block.to_dict(), self, self.repo_path)
 
         if should_update_tag_cache:
             from mage_ai.cache.tag import TagCache
@@ -1453,8 +1612,7 @@ class Pipeline:
 
         files_to_be_written = []
         with open(config_zip_path, 'r') as pipeline_config:
-
-            config = yaml.safe_load(pipeline_config)
+            config = load_yaml(pipeline_config)
 
             # check if pipeline exists with same uuid and generate new one if necessary
             if not overwrite:
@@ -1564,8 +1722,10 @@ class Pipeline:
 
             if block_uuid in self.blocks_by_uuid:
                 block = self.blocks_by_uuid[block_uuid]
-                if upstream_blocks_reordered is not None and \
-                        upstream_blocks != upstream_blocks_reordered:
+                if (
+                    upstream_blocks_reordered is not None
+                    and upstream_blocks != upstream_blocks_reordered
+                ):
                     block.update(
                         dict(upstream_blocks=upstream_blocks_reordered),
                         check_upstream_block_order=True,
@@ -1613,7 +1773,8 @@ class Pipeline:
         all_block_uuids = {b.get('uuid') for b in self.all_block_configs}
         if block.uuid in all_block_uuids:
             raise InvalidPipelineError(
-                f'Block with uuid {block.uuid} already exists in pipeline {self.uuid}')
+                f'Block with uuid {block.uuid} already exists in pipeline {self.uuid}'
+            )
 
         if widget:
             self.widgets_by_uuid = self.__add_block_to_mapping(
@@ -1666,9 +1827,8 @@ class Pipeline:
 
                 self.update_block(
                     downstream_block,
-                    upstream_block_uuids=(
-                        downstream_block.upstream_block_uuids or []
-                    ) + [block.uuid],
+                    upstream_block_uuids=(downstream_block.upstream_block_uuids or [])
+                    + [block.uuid],
                 )
 
         self.validate('A cycle was formed while adding a block')
@@ -1728,9 +1888,13 @@ class Pipeline:
         sample_count: int = None,
         dynamic_block_index: int = None,
         dynamic_block_uuid: str = None,
+        input_data_types: Optional[List[InputDataType]] = None,
+        read_batch_settings: Optional[BatchSettings] = None,
+        read_chunks: Optional[List[ChunkKeyTypeUnion]] = None,
+        write_batch_settings: Optional[BatchSettings] = None,
+        write_chunks: Optional[List[ChunkKeyTypeUnion]] = None,
     ):
         block = self.get_block(block_uuid)
-
         data_integration_settings = block.get_data_integration_settings(
             from_notebook=from_notebook,
             global_vars=global_vars,
@@ -1758,6 +1922,11 @@ class Pipeline:
             variable_uuid=variable_name,
             dynamic_block_index=dynamic_block_index,
             dynamic_block_uuid=dynamic_block_uuid,
+            input_data_types=input_data_types,
+            read_batch_settings=read_batch_settings,
+            read_chunks=read_chunks,
+            write_batch_settings=write_batch_settings,
+            write_chunks=write_chunks,
         )
 
         return variable
@@ -1779,10 +1948,12 @@ class Pipeline:
 
     def has_block(self, block_uuid: str, block_type: str = None, extension_uuid: str = None):
         if extension_uuid:
-            return self.extensions and \
-                extension_uuid in self.extensions and \
-                'blocks_by_uuid' in self.extensions[extension_uuid] and \
-                block_uuid in self.extensions[extension_uuid]['blocks_by_uuid']
+            return (
+                self.extensions
+                and extension_uuid in self.extensions
+                and 'blocks_by_uuid' in self.extensions[extension_uuid]
+                and block_uuid in self.extensions[extension_uuid]['blocks_by_uuid']
+            )
         elif BlockType.CALLBACK == block_type:
             return block_uuid in self.callbacks_by_uuid
         elif BlockType.CONDITIONAL == block_type:
@@ -1810,9 +1981,9 @@ class Pipeline:
         if upstream_block_uuids is not None:
             curr_upstream_block_uuids = set(block.upstream_block_uuids)
             new_upstream_block_uuids = set(upstream_block_uuids)
-            if curr_upstream_block_uuids != new_upstream_block_uuids or \
-                (check_upstream_block_order and
-                    block.upstream_block_uuids != upstream_block_uuids):
+            if curr_upstream_block_uuids != new_upstream_block_uuids or (
+                check_upstream_block_order and block.upstream_block_uuids != upstream_block_uuids
+            ):
                 # Only set upstream block’s downstream to the current block if current block
                 # is not an extension block and not a callback/conditional block
                 if not is_extension and not is_callback and not is_conditional:
@@ -1871,9 +2042,9 @@ class Pipeline:
             # conditional_blocks field.
             block.update_conditional_blocks(conditional_blocks)
         elif downstream_block_uuids is not None:
-            block_uuids_to_remove = \
-                [uuid for uuid in block.downstream_block_uuids
-                    if uuid not in downstream_block_uuids]
+            block_uuids_to_remove = [
+                uuid for uuid in block.downstream_block_uuids if uuid not in downstream_block_uuids
+            ]
 
             for block_uuid in block_uuids_to_remove:
                 block_inner = self.get_block(block_uuid)
@@ -1881,10 +2052,12 @@ class Pipeline:
                     continue
                 block_inner.update(
                     dict(
-                        upstream_blocks=list(filter(
-                            lambda x, uuid=block.uuid: x != uuid,
-                            block_inner.upstream_block_uuids or [],
-                        )),
+                        upstream_blocks=list(
+                            filter(
+                                lambda x, uuid=block.uuid: x != uuid,
+                                block_inner.upstream_block_uuids or [],
+                            )
+                        ),
                     ),
                     check_upstream_block_order=check_upstream_block_order,
                 )
@@ -1992,7 +2165,8 @@ class Pipeline:
     def update_global_variable(self, key, value):
         if not is_yaml_serializable(key, value):
             raise SerializationError(
-                f'Failed to update variable {key} because the value is not serializable.')
+                f'Failed to update variable {key} because the value is not serializable.'
+            )
         if self.variables is None:
             self.variables = {}
         self.variables[key] = value
@@ -2055,13 +2229,20 @@ class Pipeline:
             ]
             if self.type == PipelineType.INTEGRATION or force:
                 for downstream_block in block.downstream_blocks:
-                    upstream_block_uuids = list(filter(
-                        lambda uuid: uuid != block.uuid,
-                        downstream_block.upstream_block_uuids
-                    ))
-                    downstream_block.update(dict(
-                        upstream_blocks=[*upstream_block_uuids, *block.upstream_block_uuids]
-                    ))
+                    upstream_block_uuids = list(
+                        filter(
+                            lambda uuid: uuid != block.uuid,
+                            downstream_block.upstream_block_uuids,
+                        )
+                    )
+                    downstream_block.update(
+                        dict(
+                            upstream_blocks=[
+                                *upstream_block_uuids,
+                                *block.upstream_block_uuids,
+                            ]
+                        )
+                    )
             elif len(downstream_block_uuids) > 0:
                 raise HasDownstreamDependencies(
                     f'Block(s) {downstream_block_uuids} are depending on block {block.uuid}'
@@ -2111,7 +2292,7 @@ class Pipeline:
         blocks_current = sorted([b.uuid for b in self.blocks_by_uuid.values()])
 
         if block_uuid is not None:
-            current_pipeline = Pipeline(self.uuid, self.repo_path)
+            current_pipeline = Pipeline(self.uuid, repo_path=self.repo_path)
             block = self.get_block(
                 block_uuid,
                 block_type=block_type,
@@ -2167,18 +2348,21 @@ class Pipeline:
     def should_save_trigger_in_code_automatically(self) -> bool:
         from mage_ai.data_preparation.models.project import Project
 
-        if self.settings and \
-                self.settings.triggers and \
-                self.settings.triggers.save_in_code_automatically is not None:
-
+        if (
+            self.settings
+            and self.settings.triggers
+            and self.settings.triggers.save_in_code_automatically is not None
+        ):
             return self.settings.triggers.save_in_code_automatically
 
-        project = Project(self.repo_config)
+        project = Project(repo_config=self.repo_config)
 
-        return project.pipelines and \
-            project.pipelines.settings and \
-            project.pipelines.settings.triggers and \
-            project.pipelines.settings.triggers.save_in_code_automatically
+        return (
+            project.pipelines
+            and project.pipelines.settings
+            and project.pipelines.settings.triggers
+            and project.pipelines.settings.triggers.save_in_code_automatically
+        )
 
     async def save_async(
         self,
@@ -2235,7 +2419,7 @@ class Pipeline:
             success = True
             with open(test_path, mode='r', encoding='utf-8') as fp:
                 try:
-                    yaml.full_load(fp)
+                    load_yaml(fp)
                 except yaml.scanner.ScannerError:
                     success = False
 
@@ -2285,7 +2469,7 @@ class Pipeline:
                 index += 1
 
             cycle = [frame.uuid for frame in virtual_stack[index:]]
-            return " --> ".join(cycle)
+            return ' --> '.join(cycle)
 
         def __check_cycle(block: Block):
             virtual_stack = [StackFrame(block)]
@@ -2316,7 +2500,8 @@ class Pipeline:
             uuid = config.get('uuid')
             if uuid in check_block_uuids:
                 raise InvalidPipelineError(
-                    f'Pipeline is invalid: duplicate blocks with uuid {uuid}')
+                    f'Pipeline is invalid: duplicate blocks with uuid {uuid}'
+                )
             check_block_uuids.add(uuid)