Figures, Traces & Layout

The problem, from first principles

Plotly draws in a browser, so a plotly figure is not a picture. It is a description: a plain structure, a list of data (the traces, each one a scatter, a bar, a heatmap) and a layout (titles, axes, legend, size), which the plotly.js library in the browser turns into pixels. That is why a figure can be saved as JSON, sent anywhere, and rebuilt, and why it is interactive for free: hover, zoom and toggling a trace in the legend are plotly.js features working from the description.

The consequence for code is pleasant: building a figure is building a structure, and every property is addressable by a path like fig.layout.xaxis.title.text. This question builds a figure from scratch with plotly.graph_objects, adds a second trace, reads the structure back, sets axis ranges, and round-trips the whole figure through JSON.

From theory to code

Implement line_figure(x, y, name, title, xlabel, ylabel), add_series(fig, x, y, name, color), trace_summaries(fig), set_axis_ranges(fig, xrange, yrange) and json_roundtrip(fig). The signatures and docstrings are already in the editor.

Constraints

• Use plotly.graph_objects (import plotly.graph_objects as go). Do not call fig.show().
• line_figure returns a new go.Figure with one Scatter trace through (x, y) drawn with mode="lines+markers" and the given name, with the layout title title, x axis title xlabel and y axis title ylabel.
• add_series(fig, x, y, name, color) adds one more Scatter trace (lines only, mode="lines") to fig with that name and line colour, and returns the same figure object.
• trace_summaries(fig) returns a list with one dict per trace, in order: {"name": ..., "type": ..., "n_points": ...}, where type is the trace type (such as "scatter") and n_points is the number of x values.
• set_axis_ranges(fig, xrange, yrange) fixes the visible range of the x axis and of the y axis (each a [low, high] pair) and returns the same figure.
• json_roundtrip(fig) serialises the figure to a JSON string and rebuilds a new figure from it with plotly.io, which it returns.

Hints