`tsdat.io`¶

The tsdat.io package provides the classes that the data pipeline uses to manage I/O for the pipeline. Specifically, it includes:

The FileHandler infrastructure used to read/write to/from specific file formats, and

The Storage infrastructure used to store/access processed data files

We warmly welcome community contribututions to increase the list of supported FileHandlers and Storage destinations.

Subpackages¶

tsdat.io.filehandlers

Submodules¶

Classes¶

`AbstractFileHandler`	Abstract class to define methods required by all FileHandlers. Classes
`AwsStorage`	DatastreamStorage subclass for an AWS S3-based filesystem.
`CsvHandler`	FileHandler to read from and write to CSV files. Takes a number of
`DatastreamStorage`	DatastreamStorage is the base class for providing
`DisposableLocalTempFile`	DisposableLocalTempFile is a context manager wrapper class for a temp file on
`DisposableLocalTempFileList`	Provides a context manager wrapper class for a list of
`DisposableStorageTempFileList`	Provides is a context manager wrapper class for a list of
`FileHandler`	Class to provide methods to read and write files with a variety of
`FilesystemStorage`	Datastreamstorage subclass for a local Linux-based filesystem.
`NetCdfHandler`	FileHandler to read from and write to netCDF files. Takes a number of
`S3Path`	This class wraps a ‘special’ path string that lets us include the
`SplitNetCdfHandler`	FileHandler to read from and write to netCDF files. Takes a number of
`TemporaryStorage`	Each DatastreamStorage should contain a corresponding

class tsdat.io.AbstractFileHandler(parameters: Union[Dict, None] = None)¶

Abstract class to define methods required by all FileHandlers. Classes derived from AbstractFileHandler should implement one or more of the following methods:

write(ds: xr.Dataset, filename: str, config: Config, **kwargs)

read(filename: str, **kwargs) -> xr.Dataset

Parameters: parameters (Dict, optional) – Parameters that were passed to the FileHandler when it was registered in the storage config file, defaults to {}.

Class Methods

`read`	Reads in the given file and converts it into an Xarray dataset for
`write`	Saves the given dataset to a file.

Method Descriptions

read(self, filename: str, **kwargs) → xarray.Dataset¶

Reads in the given file and converts it into an Xarray dataset for use in the pipeline.

Parameters: filename (str) – The path to the file to read in.
Returns: A xr.Dataset object.
Return type: xr.Dataset

write(self, ds: xarray.Dataset, filename: str, config: tsdat.config.Config = None, **kwargs) → None¶

Saves the given dataset to a file.

Parameters

ds (xr.Dataset) – The dataset to save.
filename (str) – The path to where the file should be written to.
config (Config, optional) – Optional Config object, defaults to None

class tsdat.io.AwsStorage(parameters: Union[Dict, None] = None)[source]¶

Bases: tsdat.io.DatastreamStorage

DatastreamStorage subclass for an AWS S3-based filesystem.

Parameters

parameters (dict, optional) –

Dictionary of parameters that should be set automatically from the storage config file when this class is intantiated via the DatstreamStorage.from-config() method. Defaults to {}

Key parameters that should be set in the config file include

retain_input_files: Whether the input files should be cleaned up after they are done processing
root_dir: The bucket ‘key’ to use to prepend to all processed files created in the persistent store. Defaults to ‘root’
temp_dir: The bucket ‘key’ to use to prepend to all temp files created in the S3 bucket. Defaults to ‘temp’
bucket_name: The name of the S3 bucket to store to

Class Methods

`delete`	Deletes datastream data in the datastream store in between the
`exists`	Checks if any data exists in the datastream store for the provided
`fetch`	Fetches files from the datastream store using the datastream_name,
`find`	Finds all files of the given type from the datastream store with the
`root`
`s3_client`
`s3_resource`
`save_local_path`	Given a path to a local file, save that file to the storage.
`temp_path`
`tmp`	Each subclass should define the tmp property, which provides

Method Descriptions

delete(self, datastream_name: str, start_time: str, end_time: str, filetype: int = None) → None¶

Deletes datastream data in the datastream store in between the specified time range.

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108” to search for data ending before January 8th, 2021.
filetype (str, optional) – A file type from the DatastreamStorage.file_filters keys. If no type is specified, all files will be deleted. Defaults to None.

exists(self, datastream_name: str, start_time: str, end_time: str, filetype: int = None) → bool¶

Checks if any data exists in the datastream store for the provided datastream and time range.

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108” to search for data ending before January 8th, 2021.
filetype (str, optional) – A file type from the DatastreamStorage.file_filters keys. If none specified, all files will be checked. Defaults to None.

Returns

True if data exists, False otherwise.

Return type

bool

fetch(self, datastream_name: str, start_time: str, end_time: str, local_path: str = None, filetype: int = None) → tsdat.io.DisposableLocalTempFileList ¶

Fetches files from the datastream store using the datastream_name, start_time, and end_time to specify the file(s) to retrieve. If the local path is not specified, it is up to the subclass to determine where to put the retrieved file(s).

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108” to search for data ending before January 8th, 2021.
local_path (str, optional) – The path to the directory where the data should be stored. Defaults to None.
filetype (int, optional) – A file type from the DatastreamStorage.file_filters keys If no type is specified, all files will be returned. Defaults to None.

Returns

A list of paths where the retrieved files were stored in local storage. This is a context manager class, so it this method should be called via the ‘with’ statement and all files referenced by the list will be cleaned up when it goes out of scope.

Return type

DisposableLocalTempFileList:

find(self, datastream_name: str, start_time: str, end_time: str, filetype: str = None) → List[S3Path]¶

Finds all files of the given type from the datastream store with the given datastream_name and timestamps from start_time (inclusive) up to end_time (exclusive). Returns a list of paths to files that match the criteria.

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106.000000” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108.000000” to search for data ending before January 8th, 2021.
filetype (str, optional) – A file type from the DatastreamStorage.file_filters keys If no type is specified, all files will be returned. Defaults to None.

Returns

A list of paths in datastream storage in ascending order

Return type

List[str]

property root(self)¶

property s3_client(self)¶

property s3_resource(self)¶

save_local_path(self, local_path: str, new_filename: str = None)¶

Given a path to a local file, save that file to the storage.

Parameters

local_path (str) – Local path to the file to save. The file should be named according to ME Data Standards naming conventions so that this method can automatically parse the datastream, date, and time from the file name.
new_filename (str, optional) – If provided, the new filename to save as. This parameter should ONLY be provided if using a local path for dataset_or_path. Must also follow ME Data Standards naming conventions. Defaults to None.

Returns

The path where this file was stored in storage. Path type is dependent upon the specific storage subclass.

Return type

Any

property temp_path(self)¶

property tmp(self)¶

Each subclass should define the tmp property, which provides access to a TemporaryStorage object that is used to efficiently handle reading/writing temporary files used during the processing pipeline, or to perform fileystem actions on files other than processed datastream files that reside in the same filesystem as the DatastreamStorage. Is is not intended to be used outside of the pipeline.

Raises: NotImplementedError – [description]

class tsdat.io.CsvHandler(parameters: Union[Dict, None] = None)¶

Bases: tsdat.io.filehandlers.file_handlers.AbstractFileHandler

FileHandler to read from and write to CSV files. Takes a number of parameters that are passed in from the storage config file. Parameters specified in the config file should follow the following example:

parameters:
  write:
    to_dataframe:
      # Parameters here will be passed to xr.Dataset.to_dataframe()
    to_csv:
      # Parameters here will be passed to pd.DataFrame.to_csv()
  read:
    read_csv:
      # Parameters here will be passed to pd.read_csv()
    to_xarray:
      # Parameters here will be passed to pd.DataFrame.to_xarray()

Parameters: parameters (Dict, optional) – Parameters that were passed to the FileHandler when it was registered in the storage config file, defaults to {}.

Class Methods

`read`	Reads in the given file and converts it into an Xarray dataset for
`write`	Saves the given dataset to a csv file.

Method Descriptions

read(self, filename: str, **kwargs) → xarray.Dataset¶

Reads in the given file and converts it into an Xarray dataset for use in the pipeline.

Parameters: filename (str) – The path to the file to read in.
Returns: A xr.Dataset object.
Return type: xr.Dataset

write(self, ds: xarray.Dataset, filename: str, config: tsdat.config.Config = None, **kwargs) → None¶

Saves the given dataset to a csv file.

Parameters

ds (xr.Dataset) – The dataset to save.
filename (str) – The path to where the file should be written to.
config (Config, optional) – Optional Config object, defaults to None

class tsdat.io.DatastreamStorage(parameters: Union[Dict, None] = None)[source]¶

Bases: abc.ABC

DatastreamStorage is the base class for providing access to processed data files in a persistent archive. DatastreamStorage provides shortcut methods to find files based upon date, datastream name, file type, etc. This is the class that should be used to save and retrieve processed data files. Use the DatastreamStorage.from_config() method to construct the appropriate subclass instance based upon a storage config file.

Class Methods

`delete`	Deletes datastream data in the datastream store in between the
`exists`	Checks if any data exists in the datastream store for the provided
`fetch`	Fetches files from the datastream store using the datastream_name,
`find`	Finds all files of the given type from the datastream store with the
`from_config`	Load a yaml config file which provides the storage constructor
`save`	Saves a local file to the datastream store.
`save_local_path`	Given a path to a local file, save that file to the storage.
`tmp`	Each subclass should define the tmp property, which provides

Method Descriptions

abstract delete(self, datastream_name: str, start_time: str, end_time: str, filetype: str = None) → None¶

Deletes datastream data in the datastream store in between the specified time range.

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108” to search for data ending before January 8th, 2021.
filetype (str, optional) – A file type from the DatastreamStorage.file_filters keys. If no type is specified, all files will be deleted. Defaults to None.

abstract exists(self, datastream_name: str, start_time: str, end_time: str, filetype: str = None) → bool¶

Checks if any data exists in the datastream store for the provided datastream and time range.

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108” to search for data ending before January 8th, 2021.
filetype (str, optional) – A file type from the DatastreamStorage.file_filters keys. If none specified, all files will be checked. Defaults to None.

Returns

True if data exists, False otherwise.

Return type

bool

abstract fetch(self, datastream_name: str, start_time: str, end_time: str, local_path: str = None, filetype: int = None)¶

Fetches files from the datastream store using the datastream_name, start_time, and end_time to specify the file(s) to retrieve. If the local path is not specified, it is up to the subclass to determine where to put the retrieved file(s).

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108” to search for data ending before January 8th, 2021.
local_path (str, optional) – The path to the directory where the data should be stored. Defaults to None.
filetype (int, optional) – A file type from the DatastreamStorage.file_filters keys If no type is specified, all files will be returned. Defaults to None.

Returns

A list of paths where the retrieved files were stored in local storage. This is a context manager class, so it this method should be called via the ‘with’ statement and all files referenced by the list will be cleaned up when it goes out of scope.

Return type

DisposableLocalTempFileList:

abstract find(self, datastream_name: str, start_time: str, end_time: str, filetype: str = None) → List[str]¶

Finds all files of the given type from the datastream store with the given datastream_name and timestamps from start_time (inclusive) up to end_time (exclusive). Returns a list of paths to files that match the criteria.

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106.000000” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108.000000” to search for data ending before January 8th, 2021.
filetype (str, optional) – A file type from the DatastreamStorage.file_filters keys If no type is specified, all files will be returned. Defaults to None.

Returns

A list of paths in datastream storage in ascending order

Return type

List[str]

static from_config(storage_config_file: str)¶

Load a yaml config file which provides the storage constructor parameters.

Parameters: storage_config_file (str) – The path to the config file to load
Returns: A subclass instance created from the config file.
Return type: DatastreamStorage

save(self, dataset_or_path: Union[str, xarray.Dataset], new_filename: str = None) → List[Any]¶

Saves a local file to the datastream store.

Parameters

dataset_or_path (Union[str, xr.Dataset]) – The dataset or local path to the file to save. The file should be named according to ME Data Standards naming conventions so that this method can automatically parse the datastream, date, and time from the file name.
new_filename (str, optional) – If provided, the new filename to save as. This parameter should ONLY be provided if using a local path for dataset_or_path. Must also follow ME Data Standards naming conventions. Defaults to None.

Returns

A list of paths where the saved files were stored in storage. Path type is dependent upon the specific storage subclass.

Return type

List[Any]

abstract save_local_path(self, local_path: str, new_filename: str = None) → Any¶

Given a path to a local file, save that file to the storage.

Parameters

local_path (str) – Local path to the file to save. The file should be named according to ME Data Standards naming conventions so that this method can automatically parse the datastream, date, and time from the file name.
new_filename (str, optional) – If provided, the new filename to save as. This parameter should ONLY be provided if using a local path for dataset_or_path. Must also follow ME Data Standards naming conventions. Defaults to None.

Returns

The path where this file was stored in storage. Path type is dependent upon the specific storage subclass.

Return type

Any

property tmp(self)¶

Each subclass should define the tmp property, which provides access to a TemporaryStorage object that is used to efficiently handle reading/writing temporary files used during the processing pipeline, or to perform fileystem actions on files other than processed datastream files that reside in the same filesystem as the DatastreamStorage. Is is not intended to be used outside of the pipeline.

Raises: NotImplementedError – [description]

class tsdat.io.DisposableLocalTempFile(filepath: str, disposable=True)[source]¶

DisposableLocalTempFile is a context manager wrapper class for a temp file on the LOCAL FILESYSTEM. It will ensure that the file is deleted when it goes out of scope.

Parameters

filepath (str) – Path to a local temp file that could be deleted when it goes out of scope.
disposable (bool, optional) – True if this file should be automatically deleted when it goes out of scope. Defaults to True.

Class Methods

`__enter__`
`__exit__`

Method Descriptions

__enter__(self)¶

__exit__(self, type, value, traceback)¶

class tsdat.io.DisposableLocalTempFileList(filepath_list: List[str], delete_on_exception=False, disposable=True)[source]¶

Bases: list

Provides a context manager wrapper class for a list of temp files on the LOCAL FILESYSTEM. It ensures that if specified, the files will be auto-deleted when the list goes out of scope.

Parameters

filepath_list (List[str]) – A list of local temp files
delete_on_exception (bool, optional) – Should the local temp files be deleted if an error was thrown when processing. Defaults to False.
disposable (bool, optional) – Should the local temp files be auto-deleted when they go out of scope. Defaults to True.

Initialize self. See help(type(self)) for accurate signature.

Class Methods

`__enter__`
`__exit__`

Method Descriptions

__enter__(self)¶

__exit__(self, type, value, traceback)¶

class tsdat.io.DisposableStorageTempFileList(filepath_list: List[str], storage, disposable_files: Union[List, None] = None)[source]¶

Bases: list

Provides is a context manager wrapper class for a list of temp files on the STORAGE FILESYSTEM. It will ensure that the specified files are deleted when the list goes out of scope.

Parameters

filepath_list (List[str]) – A list of files in temporary storage area
storage (TemporaryStorage) – The temporary storage service used to clean up temporary files.
disposable_files (list, optional) – Which of the files from the filepath_list should be auto-deleted when the list goes out of scope. Defaults to []

Initialize self. See help(type(self)) for accurate signature.

Class Methods

`__enter__`
`__exit__`

Method Descriptions

__enter__(self)¶

__exit__(self, type, value, traceback)¶

class tsdat.io.FileHandler¶

Class to provide methods to read and write files with a variety of extensions.

FILEREADERS :Dict[str, AbstractFileHandler]¶

FILEWRITERS :Dict[str, AbstractFileHandler]¶

Class Methods

`read`	Reads in the given file and converts it into an xarray dataset object using the
`register_file_handler`	Method to register a FileHandler for reading from or writing to files matching one or
`write`	Calls the appropriate FileHandler to write the dataset to the provided filename.

Method Descriptions

read(self, filename: str, **kwargs) → xarray.Dataset¶

Reads in the given file and converts it into an xarray dataset object using the registered FileHandler for the provided filepath.

Parameters: filename (str) – The path to the file to read in.
Returns: The raw file as an Xarray.Dataset object.
Return type: xr.Dataset

register_file_handler(self, method: Literal[read, write], patterns: Union[str, List[str]], handler: AbstractFileHandler)¶

Method to register a FileHandler for reading from or writing to files matching one or more provided file patterns.

Parameters

method ("Literal") – The method the FileHandler should call if the pattern is
Must be one of (matched.) – “read”, “write”.
patterns (Union[str, List[str]]) – The file pattern(s) that determine if this
should be run on a given filepath. (FileHandler) –
handler (AbstractFileHandler) – The AbstractFileHandler to register.

write(self, ds: xarray.Dataset, filename: str, config: tsdat.config.Config = None, **kwargs) → None¶

Calls the appropriate FileHandler to write the dataset to the provided filename.

Parameters

ds (xr.Dataset) – The dataset to save.
filename (str) – The path to the file where the dataset should be written.
config (Config, optional) – Optional Config object. Defaults to None.

class tsdat.io.FilesystemStorage(parameters: Union[Dict, None] = None)[source]¶

Bases: tsdat.io.DatastreamStorage

Datastreamstorage subclass for a local Linux-based filesystem.

TODO: rename to LocalStorage as this is more intuitive.

Parameters

parameters (dict, optional) –

Dictionary of parameters that should be set automatically from the storage config file when this class is intantiated via the DatstreamStorage.from-config() method. Defaults to {}

Key parameters that should be set in the config file include

retain_input_files: Whether the input files should be cleaned up after they are done processing
root_dir: The root path under which processed files will e stored.

Class Methods

`delete`	Deletes datastream data in the datastream store in between the
`exists`	Checks if any data exists in the datastream store for the provided
`fetch`	Fetches files from the datastream store using the datastream_name,
`find`	Finds all files of the given type from the datastream store with the
`save_local_path`	Given a path to a local file, save that file to the storage.
`tmp`	Each subclass should define the tmp property, which provides

Method Descriptions

delete(self, datastream_name: str, start_time: str, end_time: str, filetype: int = None) → None¶

Deletes datastream data in the datastream store in between the specified time range.

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108” to search for data ending before January 8th, 2021.
filetype (str, optional) – A file type from the DatastreamStorage.file_filters keys. If no type is specified, all files will be deleted. Defaults to None.

exists(self, datastream_name: str, start_time: str, end_time: str, filetype: int = None) → bool¶

Checks if any data exists in the datastream store for the provided datastream and time range.

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108” to search for data ending before January 8th, 2021.
filetype (str, optional) – A file type from the DatastreamStorage.file_filters keys. If none specified, all files will be checked. Defaults to None.

Returns

True if data exists, False otherwise.

Return type

bool

fetch(self, datastream_name: str, start_time: str, end_time: str, local_path: str = None, filetype: int = None) → tsdat.io.DisposableLocalTempFileList ¶

Fetches files from the datastream store using the datastream_name, start_time, and end_time to specify the file(s) to retrieve. If the local path is not specified, it is up to the subclass to determine where to put the retrieved file(s).

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108” to search for data ending before January 8th, 2021.
local_path (str, optional) – The path to the directory where the data should be stored. Defaults to None.
filetype (int, optional) – A file type from the DatastreamStorage.file_filters keys If no type is specified, all files will be returned. Defaults to None.

Returns

A list of paths where the retrieved files were stored in local storage. This is a context manager class, so it this method should be called via the ‘with’ statement and all files referenced by the list will be cleaned up when it goes out of scope.

Return type

DisposableLocalTempFileList:

find(self, datastream_name: str, start_time: str, end_time: str, filetype: str = None) → List[str]¶

Finds all files of the given type from the datastream store with the given datastream_name and timestamps from start_time (inclusive) up to end_time (exclusive). Returns a list of paths to files that match the criteria.

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106.000000” to search for data beginning on or after January 6th, 2021.
end_time (str) – The end time or date to stop searching for data (exclusive). Should be like “20210108.000000” to search for data ending before January 8th, 2021.
filetype (str, optional) – A file type from the DatastreamStorage.file_filters keys If no type is specified, all files will be returned. Defaults to None.

Returns

A list of paths in datastream storage in ascending order

Return type

List[str]

save_local_path(self, local_path: str, new_filename: str = None) → Any¶

Given a path to a local file, save that file to the storage.

Parameters

local_path (str) – Local path to the file to save. The file should be named according to ME Data Standards naming conventions so that this method can automatically parse the datastream, date, and time from the file name.
new_filename (str, optional) – If provided, the new filename to save as. This parameter should ONLY be provided if using a local path for dataset_or_path. Must also follow ME Data Standards naming conventions. Defaults to None.

Returns

The path where this file was stored in storage. Path type is dependent upon the specific storage subclass.

Return type

Any

property tmp(self)¶

Each subclass should define the tmp property, which provides access to a TemporaryStorage object that is used to efficiently handle reading/writing temporary files used during the processing pipeline, or to perform fileystem actions on files other than processed datastream files that reside in the same filesystem as the DatastreamStorage. Is is not intended to be used outside of the pipeline.

Raises: NotImplementedError – [description]

class tsdat.io.NetCdfHandler(parameters: Union[Dict, None] = None)¶

Bases: tsdat.io.filehandlers.file_handlers.AbstractFileHandler

FileHandler to read from and write to netCDF files. Takes a number of parameters that are passed in from the storage config file. Parameters specified in the config file should follow the following example:

parameters:
  write:
    to_netcdf:
      # Parameters here will be passed to xr.Dataset.to_netcdf()
  read:
    load_dataset:
      # Parameters here will be passed to xr.load_dataset()

Parameters: parameters (Dict, optional) – Parameters that were passed to the FileHandler when it was registered in the storage config file, defaults to {}.

Class Methods

`read`	Reads in the given file and converts it into an Xarray dataset for
`write`	Saves the given dataset to a netCDF file.

Method Descriptions

read(self, filename: str, **kwargs) → xarray.Dataset¶

Reads in the given file and converts it into an Xarray dataset for use in the pipeline.

Parameters: filename (str) – The path to the file to read in.
Returns: A xr.Dataset object.
Return type: xr.Dataset

write(self, ds: xarray.Dataset, filename: str, config: tsdat.config.Config = None, **kwargs) → None¶

Saves the given dataset to a netCDF file.

Parameters

ds (xr.Dataset) – The dataset to save.
filename (str) – The path to where the file should be written to.
config (Config, optional) – Optional Config object, defaults to None

class tsdat.io.S3Path(bucket_name: str, bucket_path: str = '', region_name: str = None)[source]¶

Bases: str

This class wraps a ‘special’ path string that lets us include the bucket name and region in the path, so that we can use it seamlessly in boto3 APIs. We are creating our own string to hold the region, bucket & key (i.e., path), since boto3 needs all three in order to access a file.

Example: .. code-block:: python

s3_client = boto3.client(‘s3’, region_name=’eu-central-1’) s3_client.download_file(bucket, key, download_path)

Parameters

bucket_name (str) – The S3 bucket name where this file is located
bucket_path (str, optional) – The key to access this file in the bucket
region_name (str, optional) – The AWS region where this file is located, defaults to None, which inherits the default configured region.

Initialize self. See help(type(self)) for accurate signature.

Class Methods

`__str__`	Return str(self).
`bucket_name`
`bucket_path`
`join`	Joins segments in an S3 path. This method behaves
`region_name`

Method Descriptions

__str__(self)¶: Return str(self).

property bucket_name(self)¶

property bucket_path(self)¶

join(self, *args)¶

Joins segments in an S3 path. This method behaves exactly like os.path.join.

Returns: A New S3Path with the additional segments added.
Return type: S3Path

property region_name(self)¶

class tsdat.io.SplitNetCdfHandler(parameters: Union[Dict, None] = None)¶

Bases: NetCdfHandler

FileHandler to read from and write to netCDF files. Takes a number of parameters that are passed in from the storage config file. Parameters specified in the config file should follow the following example:

parameters:
  write:
    to_netcdf:
      # Parameters here will be passed to xr.Dataset.to_netcdf()
  read:
    load_dataset:
      # Parameters here will be passed to xr.load_dataset()

Parameters: parameters (Dict, optional) – Parameters that were passed to the FileHandler when it was registered in the storage config file, defaults to {}.

Class Methods

`read`	Reads in the given file and converts it into an Xarray dataset for
`write`	Saves the given dataset to netCDF file(s) based on the ‘time_interval’

Method Descriptions

abstract read(self, filename: str, **kwargs)¶

Reads in the given file and converts it into an Xarray dataset for use in the pipeline.

Parameters: filename (str) – The path to the file to read in.
Returns: A xr.Dataset object.
Return type: xr.Dataset

write(self, ds: xarray.Dataset, filename: str, config: tsdat.config.Config = None, **kwargs) → None¶

Saves the given dataset to netCDF file(s) based on the ‘time_interval’ and ‘time_unit’ config parameters.

Parameters

ds (xr.Dataset) – The dataset to save.
filename (str) – The path to where the file should be written to.
config (Config, optional) – Optional Config object, defaults to None

class tsdat.io.TemporaryStorage(storage: DatastreamStorage)[source]¶

Bases: abc.ABC

Each DatastreamStorage should contain a corresponding TemporaryStorage class which provides access to a TemporaryStorage object that is used to efficiently handle reading/writing temporary files used during the processing pipeline, or to perform fileystem actions on files other than processed datastream files that reside in the same filesystem as the DatastreamStorage.

TemporaryStorage methods return a context manager so that the created temporary files can be automatically removed when they go out of scope.

TemporaryStorage is a helper class intended to be used in the internals of pipeline implementations only. It is not meant as an external API for interacting with files in DatastreamStorage.

TODO: rename to a more intuitive name…

Parameters: storage (DatastreamStorage) – A reference to the corresponding DatastreamStorage

Class Methods

`clean`	Clean any extraneous files from the temp working dirs. Temp files
`create_temp_dir`	Create a new, temporary directory under the local tmp area managed by
`delete`	Remove a file from storage temp area if the file exists. If the file
`extract_files`	If provided a path to an archive file, this function will extract the
`fetch`	Fetch a file from temp storage to a local temp folder. If
`fetch_previous_file`	Look in DatastreamStorage for the first processed file before the given date.
`get_temp_filepath`	Construct a filepath for a temporary file that will be located in the
`ignore_zip_check`	Return true if this file should be excluded from the zip file check.
`local_temp_folder`	Default method to get a local temporary folder for use when retrieving

Method Descriptions

clean(self)¶

Clean any extraneous files from the temp working dirs. Temp files could be in two places:

the local temp folder - used when fetching files from the store

the storage temp folder - used when extracting zip files in some stores (e.g., AWS)

This method removes the local temp folder. Child classes can extend this method to clean up their respective storage temp folders.

create_temp_dir(self) → str¶

Create a new, temporary directory under the local tmp area managed by TemporaryStorage.

Returns: Path to the local dir.
Return type: str

abstract delete(self, file_path: str)¶

Remove a file from storage temp area if the file exists. If the file does not exist, this method will NOT raise an exception.

Parameters: file_path (str) – The path of a file located in the same filesystem as the storage.

abstract extract_files(self, file_path: Union[str, List[str]]) → DisposableStorageTempFileList ¶

If provided a path to an archive file, this function will extract the archive into a temp directory IN THE SAME FILESYSTEM AS THE STORAGE. This means, for example that if storage was in an s3 bucket ,then the files would be extracted to a temp dir in that s3 bucket. This is to prevent local disk limitations when running via Lambda.

If the file is not an archive, then the same file will be returned.

This method supports zip, tar, and tar.g file formats.

Parameters: file_path (Union[str, List[str]]) – The path of a file or a list of files that should be processed together, located in the same filesystem as the storage.
Returns: A list of paths to the files that were extracted. Files will be located in the temp area of the storage filesystem.
Return type: DisposableStorageTempFileList

abstract fetch(self, file_path: str, local_dir=None, disposable=True) → Union[DisposableLocalTempFile, str]¶

Fetch a file from temp storage to a local temp folder. If disposable is True, then a DisposableLocalTempFile will be returned so that it can be used with a context manager.

Parameters

file_path (str) – The path of a file located in the same filesystem as the storage.
local_dir ([type], optional) – The destination folder for the file. If not specified, it will be created int the storage-approved local temp folder. defaults to None.
disposable (bool, optional) – True if this file should be auto-deleted when it goes out of scope. Defaults to True.

Returns

If disposable, return a DisposableLocalTempFile, otherwise return the path to the local file.

Return type

Union[DisposableLocalTempFile, str]

abstract fetch_previous_file(self, datastream_name: str, start_time: str) → DisposableLocalTempFile ¶

Look in DatastreamStorage for the first processed file before the given date.

Parameters

datastream_name (str) – The datastream_name as defined by ME Data Standards.
start_time (str) – The start time or date to start searching for data (inclusive). Should be like “20210106” to search for data beginning on or after January 6th, 2021.

Returns

If a previous file was found, return the local path to the fetched file. Otherwise return None. (Return value wrapped in DisposableLocalTempFile so it can be auto-deleted if needed.)

Return type

DisposableLocalTempFile

get_temp_filepath(self, filename: str = None, disposable: bool = True) → DisposableLocalTempFile ¶

Construct a filepath for a temporary file that will be located in the storage-approved local temp folder and will be deleted when it goes out of scope.

Parameters

filename (str, optional) – The filename to use for the temp file. If no filename is provided, one will be created. Defaults to None
disposable (bool, optional) – If true, then wrap in DisposableLocalTempfile so that the file will be removed when it goes out of scope. Defaults to True.

Returns

Path to the local file. The file will be automatically deleted when it goes out of scope.

Return type

DisposableLocalTempFile

ignore_zip_check(self, filepath: str) → bool¶

Return true if this file should be excluded from the zip file check. We need this for Office documents, since they are actually zip files under the hood, so we don’t want to try to unzip them.

Parameters: filepath (str) – the file we are potentially extracting
Returns: whether we should check if it is a zip or not
Return type: bool

property local_temp_folder(self) → str¶

Default method to get a local temporary folder for use when retrieving files from temporary storage. This method should work for all filesystems, but can be overridden if needed by subclasses.

Returns: Path to local temp folder
Return type: str

tsdat.io¶

Subpackages¶

Submodules¶

Classes¶

`tsdat.io`¶