def test_from_df():
df = traja.generate(n=599581)
df = df.filter(items=['x', 'y'])
save_path = 'temp/test'
warmup_steps = 50
# Run 1
timeseries_method = 'last_timestep'
model = traja.models.LSTM(input_size=2, hidden_size=2, num_layers=3, output_size=2, dropout=0, bidirectional=False)
writer = delve.writers.CSVandPlottingWriter(save_path, fontsize=16, primary_metric='test_accuracy')
saturation = delve.CheckLayerSat(save_path,
[writer], model,
stats=['embed'],
timeseries_method=timeseries_method)
sequence_length = 2000
train_fraction = .25
batch_size = 50
shift = 2
train_loader, test_loader = traja.models.get_timeseries_data_loaders(df, sequence_length,
train_fraction, batch_size, shift)
trainer = traja.models.Trainer(model, train_loader, test_loader, epochs=6, optimizer="adam",
warmup_steps=warmup_steps)
trainer.train()
pass
def test_color_dark():
df = traja.generate(n=10)
index = pd.DatetimeIndex(range(10))
df.index = index
traja.plot(df, interactive=False)
ax = plt.gca()
traja.color_dark(df.x, ax)
def test_plot_clustermap():
trjs = [traja.generate(seed=i) for i in range(20)]
# Calculate displacement
displacements = [trj.traja.calc_displacement() for trj in trjs]
traja.plot_clustermap(displacements)
def test_plot_actogram():
df = traja.generate(n=10)
index = pd.DatetimeIndex(range(10))
df.index = index
activity = traja.calc_displacement(df)
activity.name = "activity"
traja.plotting.plot_actogram(df.x, interactive=False)
def test_plot_actogram():
df = traja.generate(n=1000)
index = pd.date_range("2018-01-01", periods=1000, freq="T")
df.index = index
activity = traja.calc_displacement(df)
activity.name = "activity"
traja.plotting.plot_actogram(df.x, interactive=False)
def test_color_dark():
df = traja.generate(n=10)
index = pd.DatetimeIndex(range(10))
df.index = index
ax = plt.gca()
try:
traja.color_dark(df.x, ax)
except ValueError as e:
# catch unexplained datetime value error in travis
if "view limit minimum" in str(e):
pass
""" Average direction for each grid cell ==================================== See the flow between grid cells. """ import traja df = traja.generate(seed=0) ############################################################################### # Average Flow (3D) # ----------------- # Flow can be plotted by specifying the `kind` parameter of :func:`traja.plotting.plot_flow` # or by calling the respective functions. import traja traja.plotting.plot_surface(df, bins=32) ############################################################################### # Quiver # ------ # Quiver plot # Additional arguments can be specified as a dictionary to `quiverplot_kws`. traja.plotting.plot_quiver(df, bins=32) ############################################################################### # Contour # ------- # Parameters `filled` and `quiver` are both enabled by default and can be
import os
import numpy as np
import pandas as pd
import traja
df = traja.generate(n=20)
def test_from_df():
df = pd.DataFrame({"x": [1, 2, 3], "y": [2, 3, 4]})
trj = traja.parsers.from_df(df)
np.testing.assert_allclose(df, trj)
assert isinstance(trj, traja.TrajaDataFrame)
def test_read_file():
datapath = os.path.join(traja.__path__[0], "tests", "data", "3527.csv")
trj = traja.parsers.read_file(datapath)
assert isinstance(trj, traja.TrajaDataFrame)
assert "Frame" in trj
assert "Time" in trj
assert "TrackId" in trj
assert "x" in trj
assert "y" in trj
assert "ValueChanged" in trj
df = traja.TrajaDataFrame({'x': [0, 1, 2, 3, 4], 'y': [1, 3, 2, 4, 5]})
###############################################################################
# Plotting with Traja
# =====================
#
# We start out by plotting a basic sime series trajectory using the ``traja``
# accessor and ``.plot()`` method.
df.traja.plot()
###############################################################################
# Generate Random Walks
# =====================
#
# Also, random walks can be generated using ``generate``.
df = traja.generate(n=1000, random=True, fps=30)
df.traja.plot()
###############################################################################
# Traja can re-scale data with any units
df.traja.scale(100)
df.spatial_units = 'cm'
df.traja.plot()
###############################################################################
# Rediscretize step lengths
# =========================
#
# ``rediscretize`` method allows resampling the trajectory into an arbitrary step
# length ``R``.
"""
Comparing
-----------------------------------
traja allows comparing trajectories using various methods.
"""
import traja
df = traja.generate(angular_error_sd=0.5)
###############################################################################
# Fast Dynamic Time Warping of Trajectories
# =========================================
#
# Fast dynamic time warping can be performed using the optional package
# ``fastdtw``. It can be installed with ``pip install fastdtw``.
# Source article: `link <https://cs.fit.edu/~pkc/papers/tdm04.pdf>`_.
import numpy as np
rotated = traja.utils.rotate(df, angle=np.pi / 10)
rotated.traja.plot()
###############################################################################
# Compare trajectories point-wise
# =========================================
#
dist = traja.utils.distance(df.traja.xy, rotated.traja.xy)
print(f"Distance between the two trajectories is {dist}")
""" Comparing --------- traja allows comparing trajectories using various methods. """ import traja df = traja.generate(seed=0) df.traja.plot() ############################################################################### # Fast Dynamic Time Warping of Trajectories # ========================================= # # Fast dynamic time warping can be performed using ``fastdtw``. # Source article: `link <https://cs.fit.edu/~pkc/papers/tdm04.pdf>`_. import numpy as np rotated = traja.rotate(df, angle=np.pi / 10) rotated.traja.plot() ############################################################################### # Compare trajectories hierarchically # =================================== # Hierarchical agglomerative clustering allows comparing trajectories as actograms # and finding nearest neighbors. This is useful for comparing circadian rhythms, # for example. # Generate random trajectories trjs = [traja.generate(seed=i) for i in range(20)]
"""
Plotting Multiple Trajectories
------------------------------
Plotting multiple trajectories is easy with :meth:`~traja.frame.TrajaCollection.plot`.
"""
import traja
from traja import TrajaCollection
# Create a dictionary of DataFrames, with 'id' as key.
dfs = {idx: traja.generate(idx, seed=idx) for idx in range(10, 15)}
# Create a TrajaCollection.
trjs = TrajaCollection(dfs)
# Note: A TrajaCollection can also be instantiated with a DataFrame, containing and id column,
# eg, TrajaCollection(df, id_col="id")
# 'colors' also allows substring matching, eg, {"car":"red", "person":"blue"}
lines = trjs.plot(
colors={10: "red", 11: "blue", 12: "blue", 13: "orange", 14: "purple"}
)
""" Animate trajectories ------------------------------- traja allows animating trajectories. """ import traja df = traja.generate(1000, seed=0) ############################################################################### # Plot a animation of trajectory # ============================== # An animation is generated using :func:`~traja.plotting.animate`. anim = traja.plotting.animate(df) # save=True saves to 'trajectory.mp4' #################################################################################### # .. raw:: html # # <iframe width="560" height="315" src="https://www.youtube.com/embed/-_-Ay9phB5k" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>
"""
3D Plotting with traja
----------------------
Plot trajectories with time in the vertical axis.
Note: Adjust matplotlib args ``dist``, ``labelpad``, ``aspect`` and ``adjustable```
as needed.
"""
import traja
df = traja.TrajaDataFrame({"x": [0, 1, 2, 3, 4], "y": [1, 3, 2, 4, 5]})
trj = traja.generate()
ax = trj.traja.plot_3d(dist=15, labelpad=32, title="Traja 3D Plot")
########
# Colors
# -------
#
# `Matplotlib cmaps<https://matplotlib.org/examples/color/colormaps_reference.html>`_ are available
trj.traja.plot_3d(dist=15, labelpad=32, title="Traja 3D Plot", cmap="jet")
def test_generate():
df = traja.generate(n=20)
actual = df.traja.xy[:3]
expected = np.array([[0.0, 0.0], [1.162_605_74, 1.412_179_34],
[1.861_836_8, 2.727_243_73]])
npt.assert_allclose(actual, expected, rtol=1e-1)
from traja.dataset import dataset
from traja.dataset.example import jaguar
from traja.models.generative_models.vae import MultiModelVAE
from traja.models.train import HybridTrainer
from traja.plotting import plot_prediction
matplotlib.use("Agg")
import matplotlib.pyplot as plt
import pandas as pd
import traja
warnings.filterwarnings("ignore", category=UserWarning, module="matplotlib")
df = traja.generate(n=10)
def test_stylize_axes():
collection = traja.plot(df, interactive=False)
traja.plotting.stylize_axes(collection.axes)
def test_sans_serif():
traja.plotting.sans_serif()
def test_plot_3d():
traja.plot_3d(df, interactive=False)
df = traja.TrajaDataFrame({'x':[0,1,2,3,4],'y':[1,3,2,4,5]})
###############################################################################
# Plotting with Traja
# =====================
#
# We start out by plotting a basic sime series trajectory using the ``traja``
# accessor and ``.plot()`` method.
df.traja.plot()
###############################################################################
# Generate Random Walks
# =====================
#
# Also, random walks can be generated using ``generate``.
df = traja.generate(n=1000, fps=30)
df.traja.plot()
###############################################################################
# Traja can re-scale data with any units
df.traja.scale(100)
df.spatial_units='cm'
df.traja.plot()
###############################################################################
# Rediscretize step lengths
# =========================
#
# ``rediscretize`` method allows resampling the trajectory into an arbitrary step
# length ``R``.
""" Average direction for each grid cell ==================================== See the flow between grid cells. """ import traja df = traja.generate() ############################################################################### # Average Flow (3D) # ----------------- # Flow can be plotted by specifying the `kind` parameter of :func:`traja.plotting.plot_flow` # or by calling the respective functions. import traja traja.plotting.plot_surface(df, bins=32) ############################################################################### # Quiver # ------ # Quiver plot # Additional arguments can be specified as a dictionary to `quiverplot_kws`. traja.plotting.plot_quiver(df, bins=32) ############################################################################### # Contour # ------- # Parameters `filled` and `quiver` are both enabled by default and can be
import numpy as np
import numpy.testing as npt
import pytest
from pandas.util.testing import assert_series_equal
import traja
df = traja.generate(n=20, convex_hull=True)
def test_polar_to_z():
df_copy = df.copy()
polar = traja.cartesian_to_polar(df_copy.traja.xy)
z = traja.polar_to_z(*polar)
z_actual = z[:10]
z_expected = np.array([
0.0 + 0.0j,
1.162_605_74 + 1.412_179_34j,
1.861_836_8 + 2.727_243_73j,
1.860_393_36 + 4.857_966_96j,
-0.096_486_29 + 5.802_456_77j,
-1.735_291_68 + 4.940_704_34j,
-4.189_217_4 + 4.951_826_17j,
-5.712_624_22 + 4.177_006j,
-7.567_193_14 + 3.404_176_98j,
-9.415_289_13 + 2.743_725_89j,
])
npt.assert_allclose(z_actual, z_expected)
