pocolog2msgpack package

Submodules

pocolog2msgpack.BufferedFileDecompressor module

class pocolog2msgpack.BufferedFileDecompressor.BufferedFileDecompressor(path)

Bases: object

CHUNKSIZE = 32768

close()

open()

read(n)

reset()

seek(n)

pocolog2msgpack.FifoFileBuffer module

class pocolog2msgpack.FifoFileBuffer.FifoFileBuffer

Bases: object

read(size=None)

Reads size bytes from buffer

write(data)

Appends data to buffer

pocolog2msgpack.object2relational module

pocolog2msgpack.object2relational.object2relational(input_filename, output_filename, whitelist=())

Convert a MsgPack logfile from object-oriented to relational storage.

The log data produced by pocolog2msgpack can usually be accessed with log[port_name][sample_idx][field_a]…[field_b]. This is an object-oriented view of the data because you can easily access a whole object from the log. This is not convenient if you want to use the data for, e.g., machine learning, where you typically need the whole dataset in a 2D array, i.e. a relational view on the data, in which you can access data in the form log[port_name][feature][sample_idx].

Parameters

input_filenamestr

Name of the original logfile

output_filenamestr

Name of the converted logfile

whitelistlist or tuple

Usually arrays and vectors (represented as lists in Python) are handled as basic types and are put in a single column because they can have dynamic sizes. This is a list of fields that will be scanned recursively and interpreted as arrays with a fixed length. Note that you only have to give the name of the field, not the port name. An example would be [“elements”, “names”] if you want fully unravel a JointState object.

pocolog2msgpack.readhelpers module

pocolog2msgpack.readhelpers.get_fields(filename, decompress='auto')

pocolog2msgpack.readhelpers.read_field(filename, srcKey, dataToNumpyArray=True, decompress='auto')

pocolog2msgpack.readhelpers.read_fields(filename, srcKeys=None, renameMap=None, dataToNumpyArray=True, allowDuplicates=False, decompress='auto')

Read the list of fields in srcKeys from compress msgpack file determined by filename.

Parameters:

• filename – the path of the input file (e.g. a *.msg.zz file)

• srcKeys – the list of fields (to level keys) to read or None to read all or can be the same as renameMap

• renameMap – a dict listing how the srcKeys should be called in the output

• dataToNumpyArray – if true, each field is converted to numpy array

• allowDuplicates – if true, values of output dict become lists

Returns:

a dict with the field names or the values of renameMap as keys, and the data as values

Raises:

IndexError – raises an exception a key cannot be found or is not unique

Module contents

Top-level package for pocolog2msgpack.