Python bindings for rock tools-pocolog_cpp
This repository provides Python bindings for the pocolog tools https://github.com/rock-core/tools-pocolog_cpp. As it relies on the mentioned tools-pocolog_cpp component, these Python bindings require a full rock installation. This repository needs to be cloned into ~/rock/tools/pocolog_pybind.
Assuming the home folder corresponds to /home/user, we can install the Python bindings:
python3 -m pip install /home/user/rock/tools/pocolog_pybind
Below you can find a minimal example to use the Python bindings to read-in a pocolog log:
import pocolog_pybind
# create file index. Its possible to specify multiple logfiles
multi_file_index = pocolog_pybind.pocolog.MultiFileIndex()
multi_file_index.create_index(["/home/user/rock/bundles/hdpr/logs/20210126-0820/ga_slam.0.log"])
# we gather all available streams and put them into a dictionary using their names as keys
streams = self.multi_file_index.get_all_streams()
# we loop through the /ga_slam.localElevationMapMean stream
stream = streams["/ga_slam.localElevationMapMean"]
print("type", stream.get_type_name())
print("size", stream.get_size())
print("first_sample_time", stream.get_first_sample_time())
print("last_sample_time", stream.get_last_sample_time())
for t in range(stream.get_size())):
value = stream.get_sample(t)
"""
print the structure of the Typelib::Value
The Typelib::Category is most likely a Container at the top-level of the stream thus we can print the structure of the container
"""
value.get_type().display(indent="")
"""
We cast the Typelib::Value object recursively into python types
Typelib::Compound --> Python dict
Typelib::Array --> Python list
Typelib::Container --> Python string or array
Typelib::Numeric --> corresponding numeric type in Python
If we set recursively=False, only the first level is casted
"""
py_value = value.cast(recursive=True)
# we need to free the memory in the heap of the Typelib::Value object:
value.destroy()
# now we can print the keys of the Python dictionary
print(py_value.keys())
# and access our local DEM
local_dem = np.array(py_value["data"])
local_dem = occ_dem.reshape((py_value["height"], py_value["width"]), order="F")