`pandas_bridge`

The evo.tools.pandas_bridge module converts:

evo.core.trajectory.PosePath3D
evo.core.trajectory.PoseTrajectory3D
evo.core.result.Result

to Pandas dataframes.

Load some required modules for this demo:



In [ ]:

    
# magic plot configuration
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline
%matplotlib notebook

from IPython.display import display

Load Pandas and pandas_bridge:



In [ ]:

    
import pandas as pd
from evo.tools import pandas_bridge
from evo.tools import file_interface

Trajectories



In [ ]:

    
# load a trajectory from a ROS bag...
try:
    import rosbag
    bag_handle = rosbag.Bag("../test/data/ROS_example.bag")
    traj = file_interface.read_bag_trajectory(bag_handle, "S-PTAM")
except ImportError as e:
    print(e)  # ROS not found



In [ ]:

    
# ...or from a KITTI file...
traj = file_interface.read_kitti_poses_file("../test/data/KITTI_00_gt.txt")



In [ ]:

    
# ...or from a TUM file...
traj = file_interface.read_tum_trajectory_file("../test/data/fr2_desk_ORB.txt")

Convert a trajectory to a dataframe:



In [ ]:

    
traj_df = pandas_bridge.trajectory_to_df(traj)

Some examples for what you can do with it:



In [ ]:

    
print("First entries of the dataframe:")
display(traj_df.head())

print("Some statistics of the dataframe:")
display(traj_df.describe())

print("A plot:")
traj_df[["x", "y", "z"]].plot(kind="line", subplots=True)

Results



In [ ]:

    
# generate some result files
import subprocess as sp
cmd_1 = "evo_ape kitti ../test/data/KITTI_00_gt.txt ../test/data/KITTI_00_ORB.txt --save_results ../test/data/res1.zip --no_warnings"
cmd_2 = "evo_ape kitti ../test/data/KITTI_00_gt.txt ../test/data/KITTI_00_SPTAM.txt --save_results ../test/data/res2.zip --no_warnings"
sp.call(cmd_1.split(" "))
sp.call(cmd_2.split(" "))

Load results:



In [ ]:

    
result_1 = file_interface.load_res_file("../test/data/res1.zip")
result_2 = file_interface.load_res_file("../test/data/res2.zip")

Convert results into individual dataframes and concatenate them:



In [ ]:

    
result_df_1 = pandas_bridge.result_to_df(result_1)
result_df_2 = pandas_bridge.result_to_df(result_2)

result_df = pd.concat([result_df_1, result_df_2], axis="columns")
display(result_df)

Some examples for what you can do with it:



In [ ]:

    
display(result_df.loc["stats"])

exclude = result_df.loc["stats"].index.isin(["sse"])  # don't plot sse
result_df.loc["stats"][~exclude].plot(kind="bar")



In [ ]: