Plot for scientific validation#
We recommend executing this notebook with kbatch_papermill.
Given a result repository result_root, a tag name tag_name and the associated tag repository tag_root, the following cells compute a plot containing:
The temperature and pressure timeseries of the tag.
The different reconstructed trajectories of the fish.
# install pangeo-fish
!pip install pangeo-fish
!pip install --upgrade "dask<2025" "distributed<2025"
!pip install dask[dataframe]
!pip list | grep -e dask -e distributed
# import libraries
import xarray as xr
import pangeo_fish
import warnings
import holoviews as hv
import pandas as pd
import movingpandas as mpd
import geopandas as gpd
from holoviews.element.chart import Curve
from holoviews.core.overlay import Overlay
from pangeo_fish.io import open_tag, read_trajectories, save_html_hvplot
from pangeo_fish.tags import to_time_slice
pangeo_fish.__version__
# this will be updated by `papermill` (or do it yourself if you execute the notebook manually!)
tag_root = "s3://gfts-ifremer/bar_taos/formatted/"
tag_name = "PE_A18891"
result_root = "s3://gfts-ifremer/bar_taos/run/username/test"
storage_options = {
"anon": False,
"profile": "gfts",
"client_kwargs": {
"endpoint_url": "https://s3.gra.perf.cloud.ovh.net",
"region_name": "gra",
},
}
track_modes = ["mean", "mode"]
target_root = f"{result_root}/{tag_name}"
target_root
# load trajectories
trajectories = read_trajectories(
track_modes, target_root, storage_options, format="parquet"
)
# retrieve the pd.DataFrame objects
mean_df = trajectories.trajectories[0].df
mode_df = trajectories.trajectories[1].df
Plotting Functions#
opts_kwargs = {
"color": "red",
"line_width": 1,
"alpha": 0.5,
"line_dash": "dashed",
}
def create_vertical_lines(da: xr.DataArray, opts={}):
# creates vertical lines styled w.r.t ``opts``
vlines = hv.VLines(da.time.values).opts(**opts)
return vlines
def add_vertical_lines(plot, da: xr.DataArray, opts={}, margin_factor=0.1):
# adds vertical lines to a Curve object while preserving the initial vertical vertical limits
if isinstance(plot, Curve):
dim_name = plot.dimensions()[-1]
elif isinstance(plot, Overlay):
dim_name = plot.get_dimension(plot.ddims[-1])
else:
raise Exception(f"unknown type for plot: {type(plot)}")
y_range = plot.range(dim_name)
y_min, y_max = y_range
padding = (y_max - y_min) * margin_factor
y_min -= padding
y_max += padding
vlines = create_vertical_lines(da, opts)
return (plot * vlines).opts(ylim=(y_min, y_max))
def plot_ts(mean_df: pd.DataFrame, mode_df: pd.DataFrame, tag: xr.DataTree):
time_slice = to_time_slice(tag["tagging_events/time"])
tag_log = tag["dst"].ds.sel(time=time_slice)
# Creating pandas series for xarrray dataset
mean_lon_ = pd.Series(mean_df.geometry.x, name="longitude")
mean_lat_ = pd.Series(mean_df.geometry.y, name="latitude")
mode_lon_ = pd.Series(mode_df.geometry.x, name="longitude")
mode_lat_ = pd.Series(mode_df.geometry.y, name="latitude")
# Creating xarray datasets
mean_coords = xr.Dataset(pd.concat([mean_lon_, mean_lat_], axis=1))
mode_coords = xr.Dataset(pd.concat([mode_lon_, mode_lat_], axis=1))
# Assigning dataarrays to variables
mean_lon = mean_coords["longitude"]
mean_lat = mean_coords["latitude"]
mode_lon = mode_coords["longitude"]
mode_lat = mode_coords["latitude"]
width = 500
height = 250
tag_log["depth"] = tag_log["pressure"]
temp_plot = tag_log["temperature"].hvplot(
color="Red", title="Temperature (°C)", grid=True
)
temp_plot = add_vertical_lines(
temp_plot, da=tag["acoustic"], opts=opts_kwargs, margin_factor=0.1
).opts(height=height, width=width)
depth_plot = (-tag_log["depth"]).hvplot(color="Blue", title="Depth (m)", grid=True)
depth_plot = add_vertical_lines(
depth_plot, da=tag["acoustic"], opts=opts_kwargs, margin_factor=0.1
).opts(height=height, width=width)
lon_plot = (
mean_lat.hvplot(label="mean", clim=[mean_lat_.min(), mean_lat_.max()])
* mode_lat.hvplot(label="mode", clim=[mode_lat_.min(), mean_lat_.max()])
).opts(show_grid=True, title="Fish latitude over time")
lon_plot = add_vertical_lines(
lon_plot, da=tag["acoustic"], opts=opts_kwargs, margin_factor=0.1
).opts(height=height, width=width)
lat_plot = (
mean_lon.hvplot(label="mean", clim=[mean_lon_.min(), mean_lat_.max()])
* mode_lon.hvplot(label="mode", clim=[mode_lon_.min(), mean_lat_.max()])
).opts(show_grid=True, title="Fish longitude over time")
lat_plot = add_vertical_lines(
lat_plot, da=tag["acoustic"], opts=opts_kwargs, margin_factor=0.1
).opts(height=height, width=width)
return (temp_plot + depth_plot + lon_plot + lat_plot).cols(1)
def gather_points(tag: xr.DataTree):
def _aux_df(data: dict, **points_kwargs):
df = pd.DataFrame.from_dict(data)
gdf = gpd.GeoDataFrame(
df, geometry=gpd.points_from_xy(df.longitude, df.latitude), crs="EPSG:4326"
)
kwargs = {
"geo": True,
"tiles": "CartoLight",
"x": "longitude",
"y": "latitude",
"size": 50,
}
kwargs.update(points_kwargs)
return gdf.hvplot.points(**kwargs)
# adds the initial position
longitude = [tag["tagging_events"].isel(event_name=0).longitude.to_numpy().item()]
latitude = [tag["tagging_events"].isel(event_name=0).latitude.to_numpy().item()]
data = {"longitude": longitude, "latitude": latitude}
init_plot = _aux_df(
data, color="red", marker="o", label="Release", tiles=None, legend=True
)
# adds the detection positions
# the variable names might depend on the tags
lon_var_name, lat_var_name = None, None
for var_name in tag["acoustic"].data_vars:
if "longitude" in var_name:
lon_var_name = var_name
if "latitude" in var_name:
lat_var_name = var_name
if (lon_var_name is None) or (lat_var_name is None):
warnings.warn(
'Lon/lat variables in `tag["acoustic"]` could not be found.', RuntimeWarning
)
return init_plot
longitude = tag["acoustic"][lon_var_name].to_numpy()
latitude = tag["acoustic"][lat_var_name].to_numpy()
data = {"longitude": longitude, "latitude": latitude}
detections_plot = _aux_df(
data, color="black", marker="x", label="Detections", tiles=None, legend=True
)
return init_plot * detections_plot
def plot_track(mean_df: pd.DataFrame, mode_df: pd.DataFrame):
try:
sigma = pd.read_json(f"{target_root}/parameters.json").to_dict()[0]["sigma"]
except FileNotFoundError:
print(
'Optimisation result ("parameters.json") not found. Sigma won\'t be shown in the title.'
)
sigma = ""
mean_df["month"] = mean_df.index.strftime("%B")
mode_df["month"] = mode_df.index.strftime("%B")
# Converting back to trajectories (...)
mean_traj = mpd.Trajectory(
mean_df, traj_id=mean_df.traj_id.drop_duplicates().values[0]
)
mode_traj = mpd.Trajectory(
mode_df, traj_id=mode_df.traj_id.drop_duplicates().values[0]
)
trajectories = mpd.TrajectoryCollection([mean_traj, mode_traj])
plots = []
for i, traj in enumerate(trajectories.trajectories):
title = f"track mode: {traj.id}"
title += f", {tag_name}, {sigma:.5f}" if (i % 2 == 0) else ""
traj_plot = traj.hvplot(
c="month",
tiles="CartoLight",
cmap="rainbow",
title=title,
width=500,
height=400,
legend=True, # (i % 2 == 0)
)
# if (i % 2 == 0):
# traj_plot = traj_plot.opts(legend_position="bottom_right")
plots.append(traj_plot)
return hv.Layout(plots).cols(2)
Execution#
tag = open_tag(tag_root, tag_name)
tag
ts_plot = plot_ts(mean_df, mode_df, tag)
track_plot = plot_track(mean_df, mode_df) * gather_points(tag)
plot = (ts_plot + track_plot).cols(2)
plot
save_html_hvplot(plot, f"{target_root}/ts_track_plot.html", storage_options)