Hacktoberfest 2026: le issue che i maintainer hanno segnato per ottobre, aperte e adatte ai principianti. Sfoglia le issue Hacktoberfest

fit_annuli: plots=None renders all panels (can't disable via None) and leaks Matplotlib figures in loops

Aperta
#35 0 commenti 0 reazioni 0 assegnatari Vedi su GitHub

Nessuno ha ancora preso questa issue.

Valutazione

Difficoltà
4/5
Tempo stimato
3-5 giorni
Idoneità per principianti
48/100
Tipo di issue
Bug
Chiarezza
Abbastanza chiara
Stato di attività
Tranquilla
Stack tecnologico
matplotlib, numpy, python

Direzione di ricerca

Start in eddy/rotationmap.py at fit_annuli around line 534, then inspect plot_velocity_profiles, plot_model, and plot_model_residual around the referenced plotting code. Run the 30-call reproduction and compare plots=None, plots=[], and the documented options. Done means the plots argument behavior, figure lifecycle, and docstring agree without accumulating open figures.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Descrizione

Summary

In rotationmap.fit_annuli, the plots argument defaults to None, but None is
interpreted as "plot all panels" rather than "plot nothing". To suppress plotting you
have to pass plots=[], which is neither documented nor obvious. Because the plotting
helpers open figures without closing them, calling fit_annuli in a loop (e.g. a
Monte-Carlo over many maps) leaks a figure per call — triggering matplotlib's
More than 20 figures have been opened warning and growing memory unbounded.

Version

eddy 3.1.0, eddy/rotationmap.py.

Where

# eddy/rotationmap.py, fit_annuli, ~L534
plots = ['profiles', 'model', 'residual'] if plots is None else plots
plots = np.atleast_1d(plots)
if 'profiles' in plots:
    self.plot_velocity_profiles(rpnts=rpnts, velo=velo, dvelo=dvelo)
if 'model' in plots:
    self.plot_model(model=model)
if 'residual' in plots:
    self.plot_model_residual(model=model)

Each of plot_velocity_profiles / plot_model / plot_model_residual does
fig, ax = plt.subplots() (e.g. rotationmap.py:2321) and never closes the figure.

Reproduce

import matplotlib.pyplot as plt
# rmap = some rotationmap / momentmap with a fittable map
for i in range(30):
    rmap.fit_annuli(x0=0, y0=0, inc=inc, PA=PA, rbins=rbins,
                    fit_vrad=True, fix_vlsr=vlsr, returns='residual')  # plots=None
print(len(plt.get_fignums()))   # ~90 open figures (3 per call); RuntimeWarning fires at 20

Expected vs. actual

  • Expected: plots=None (the default) produces no plots — the conventional meaning of
    an optional plots/plot kwarg — and repeated calls do not accumulate figures.
  • Actual: plots=None renders all three panels every call, and the figures are never
    closed, so a loop leaks figures and eventually warns / risks OOM.

Secondary: docstring mismatch

The docstring says:

plots (Optional[list]): Plots to generate after the fitting. Can be either of 'model'
and 'residual'. Default is both.

But the code default is three panels including 'profiles', and 'profiles' is a
valid (undocumented) option.

Suggested fix (your call)

  • Treat a falsy plots (None / [] / False) as no plots; if plot-by-default is
    wanted, use an explicit default list instead of the None sentinel.
  • Close the figures after drawing (or return them) so loop callers don't leak — or gate
    plotting behind an explicit plot=True.
  • Update the docstring to list 'profiles' and state the actual default.

Workaround

Pass plots=[] to suppress plotting.

Lingua principale
Python
Stelle
23
Fork
10
Merge medio
28m
PR unite (30g)
1

Guida per i contributori

Apri la guida per i contributori

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di PlanetFormationLab/eddy

Tutte le issue di PlanetFormationLab/eddy

Issue simili

Altre issue su Python

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.