perseuspy.io.perseus package

Submodules

perseuspy.io.perseus.matrix module

class perseuspy.io.perseus.matrix.PathOrFile(path_or_file, mode=None, reset=False)

Bases: object

Small context manager for file paths or file-like objects :param path_or_file: Path to a file or file-like object :param mode: Set reading/writing mode :param reset: Reset file-like to initial position. Has no effect on path.

perseuspy.io.perseus.matrix.create_column_index(annotations)

Create a pd.MultiIndex using the column names and any categorical rows. Note that also non-main columns will be assigned a default category ‘’.

perseuspy.io.perseus.matrix.read_annotations(path_or_file, separator, type_map={'C': 'category', 'T': <class 'str'>, 'N': <class 'float'>, 'M': <class 'str'>, 'E': <class 'float'>}, reset=True)

Read all annotations from the specified file.

>>> annotations = read_annotations(path_or_file, separator)
>>> colnames = annotations['Column Name']
>>> types = annotations['Type']
>>> annot_row = annotations['Annot. row name']

Parameters:

• path_or_file – Path or file-like object
• separator – Column separator
• type_map – Mapping Perseus types to numpy.dtype
• reset – Reset the file after reading. Useful for file-like, no-op for paths.

Returns:

Ordered dictionary of annotations.

perseuspy.io.perseus.matrix.read_perseus(path_or_file, type_map={'C': 'category', 'T': <class 'str'>, 'N': <class 'float'>, 'M': <class 'str'>, 'E': <class 'float'>}, **kwargs)

Read a Perseus-formatted matrix into a pd.DataFrame. Annotation rows will be converted into a multi-index.

By monkey-patching the returned pd.DataFrame a to_perseus method for exporting the pd.DataFrame is made available.

Parameters:

• path_or_file – File path or file-like object
• type_map – How to map Perseus types to numpy.dtype
• kwargs – Keyword arguments passed as-is to pandas.read_csv

Returns:

The parsed data frame

perseuspy.io.perseus.matrix.to_perseus(df, path_or_file, main_columns=None, separator='\t', type_map={category: 'C', dtype('<U'): 'T', dtype('O'): 'T', dtype('float64'): 'N', dtype('int64'): 'N'}, numerical_annotation_rows=set())

Save pd.DataFrame to Perseus text format.

Parameters:

• df – pd.DataFrame
• path_or_file – File name or file-like object
• main_columns – Main columns. Will be infered if set to None. All numeric columns up-until the first non-numeric column are considered main columns.
• separator – For separating fields, default ‘ ‘

perseuspy.io.perseus.network module

perseuspy.io.perseus.network.from_perseus(network_table, networks)

Create networkx graph from network tables

>>> from perseuspy import read_networks, nx
>>> network_table, networks = read_networks(folder)
>>> graphs = nx.from_perseus(network_table, networks)

perseuspy.io.perseus.network.read_networks(folder)

Read perseus network collection folder format

>>> network_table, networks = read_networks(folder)

Parameters:folder – Path to network collection Returns:Network table and dictionary with ‘name’, ‘edge_table’, and ‘node_table’ keys.

perseuspy.io.perseus.network.to_perseus(graphs)

perseuspy.io.perseus.network.write_networks(folder, network_table, networks)

Writing networkTable, nodes and edges to Perseus readable format.

Parameters:

• folder – Path to output directory.
• network_table – Network table.
• networks – Dictionary with node and edge tables, indexed by network guid.

Module contents