store

Store functionality for the CLI

Notes

A store is an abstraction for a folder in the user’s local data directory which pertains to a specific dataset that comprises of RNG output. The store can subsequently store test results and report markup for said results.

exception coinflip.store.DataParsingError

Base class for parsing-related errors

coinflip.store.parse_data(data_file, dtype_str=None) → pandas.core.series.Series

Reads file containing data into a pandas Series

Reads from file containing RNG output and produces a representitive pandas Series. The appropiate dtype is inferred from the data itself, or optionally from the supplied dtype_str.

Parameters:

• data_file (file-like object) – File containing RNG output
• dtype_str (str, optional) – String representation of desired dtype. If not supplied, it is inferred from the data.

Returns:

Series – A pandas Series which represents the data

Raises:

• TypeNotRecognizedError – If supplied dtype_str does not recognise a dtype
• MultipleColumnsError – If inputted data contains multiple values per line
• NonBinarySequenceError – If sequence does not contain only 2 values

See also

pandas.read_csv()

The pandas method for reading data_file

store_data()

Calls this method, and handles subsequent storage of data

exception coinflip.store.StoreError(store_name)

Base class for store-related errors

coinflip.store.init_store(name=None, overwrite=False)

Creates store in local data

A name supplied or generated is used to initialise a store. If supplied, the name is sanitised to remove invalid characters for filepaths. If generated, the name will be a timestamp of initialisation.

Parameters:

• name (str, optional) – Desired name of the store, which will be sanitised. If not supplied, a name is generated automatically.
• overwrite (boolean, default False) – If a name conflicts with an existing store, this decides whether to overwrite it.

Returns:

• store_name (str) – Internal name of the initialised store
• store_path (Path) – Path of the initialised store

Raises:

• NameConflictError – If attempts at generating a unique name fails
• StoreExistsError – If a store of the same name exists already (and overwrite is set to False)
• NonBinarySequenceError – If sequence does not contain only 2 values

See also

store_data()

Parses data and calls this method, to then save data in store

coinflip.store.store_data(data_file, name=None, dtype_str=None, overwrite=False)

Load and parse RNG output, serialised to a local data directory

Reads from file containing RNG output and produces a representitive pandas Series. The appropiate dtype is inferred from the data itself, or optionally from the supplied dtype_str.

A name supplied or generated is used to initialise a store. If supplied, the name is sanitised to remove invalid characters for filepaths. If generated, the name will be a timestamp of initialisation.

The representive Series is serialised using Python’s pickle module, saved in the initialised store.

The store’s name is also written to a file in the user data directory, to be accessed later when identifying the last initialised store.

Parameters:

• data_file (file-like object) – File containing RNG output
• name (str, optional) – Desired name of the store, which will be sanitised. If not supplied, a name is generated automatically.
• dtype_str (str, optional) – String representation of desired dtype. If not supplied, it is inferred from the data.
• overwrite (bool, default False) – If a name conflicts with an existing store, this decides whether to overwrite it.

Raises:

• TypeNotRecognizedError – If supplied dtype_str does not recognise a dtype
• MultipleColumnsError – If inputted data contains multiple values per line
• NameConflictError – If attempts at generating a unique name fails
• StoreExistsError – If a store of the same name exists already (and overwrite is set to False)

See also

parse_data()

Loads and parses data_file

init_store()

Initialises the store

find_latest_store()

Accesses the name of the last initialised store

exception coinflip.store.NoLatestStoreRecordedError

Error for when latest store cannot be identified

coinflip.store.find_latest_store() → str

Find out the last initialised store

A file is kept in the root user data directory to record the last initialised store’s name, which this method reads to identify the store.

Returns:store_name (str) – Name of the last initialised store Raises:NoLatestStoreRecordedError – When no last initialised store is found

coinflip.store.get_data(store_name) → pandas.core.series.Series

Access data of a store

Parameters:

store_name (str) – Name of the store

Returns:

Series – A pandas Series which represents the data

Raises:

• StoreNotFoundError – If requested store does not exist
• DataNotFoundError – If requested store has no data

coinflip.store.drop(store_name)

Remove store from local data

Parameters:store_name (str) – Name of store to remove

coinflip.store.list_stores()

List all stores in local data

coinflip.store.store_result(store_name, randtest_name, result: coinflip.randtests._result.TestResult)

Store result of a statistical test

Parameters:

• store_name (str) – Name of store to save result in
• randtest_name (str) – Name of statistical test the result came from
• result (TestResult) – Result of the statistical test

See also

store_results()

Store multiple results from multiple statistical tests

coinflip.store.store_results(store_name, results_dict: Dict[str, coinflip.randtests._result.TestResult])

Store results of multiple statistical tests

Parameters:

• store_name (str) – Name of store to save result in
• results_dict (Dict[str, TestResult]) – Mapping of statistical tests to their respective results

See also

store_result()

Store a single results from a single statistical test

coinflip.store.open_results(store_name)

Context manager to read/write results of a store

Parameters:store_name (str) – Name of store to access results in Yields:results (Dict[str, TestResult]) – Previously stored results of statistical tests Raises:StoreNotFoundError – If requested store does not exist