Examples: query, "exact match", wildcard*, wild?ard, wild*rd
Fuzzy search: cake~ (finds cakes, bake)
Term boost: "red velvet"^4, chocolate^2
Field grouping: tags:(+work -"fun-stuff")
Escaping: Escape characters +-&|!(){}[]^"~*?:\ with \, e.g. \+
Range search: properties.timestamp:[1587729413488 TO *] (inclusive), properties.title:{A TO Z}(excluding A and Z)
Combinations: chocolate AND vanilla, chocolate OR vanilla, (chocolate OR vanilla) NOT "vanilla pudding"
Field search: properties.title:"The Title" AND text
Answered
Hey Clearml Community! Quick Question About Plots - We'Re Trying To Draw A Reliability/Calibration Plot, We Want To Make It Square As Seen In The First Picture Since It Makes The Visual Analysis Of It Much Easier, But Clearml 'Insists' On Squishing It Dow

Hey ClearML community! Quick question about plots - we're trying to draw a reliability/calibration plot, we want to make it square as seen in the first picture since it makes the visual analysis of it much easier, but ClearML 'insists' on squishing it down 😅 as seen in the second picture, regardless on how much we try to hardcode it to be this way - are we missing anything?

  
  
Posted 2 years ago
Votes Newest

Answers 9


AgitatedDove14 sorry if this was confusing, i try to clarify:

  1. i don't use matplotlib period, hence don;t use report_matplotlib_figure . i use report_plotly

  2. the separator doesn't do the trick when there's just 1 plot in a row. and 1 plot in a row seems to be a "deeply" default behavior - i.e. would require ugly workarounds on our side like adding empty plots

  3. i explicitly set the dimensions in the plotly via the code above, before passing it to report_plotly . theoretically since it's plotly -native from the get-go, the dimensions should be respected since they're contained in the fig.layout field. unless you strip it somewhere from https://github.com/allegroai/clearml/blob/1e680ae82f21b12997ed624d4c41d609fba57d4d/clearml/backend_interface/metrics/reporter.py#L349 or unless plotly/dash at the web ui frontend do something funny and disregard certain fields.

  
  
Posted 2 years ago

Hi GloriousPenguin2 , Sorry this is a bit confusing. Let me expand:
When converting into a plotly object (the default), you cannot really control the dimensions of the plot in the UI programatically, you can however drag the seperator and expand width / height If you pass to report_matplotlib_figure the argument " report_image=True, " it will create a static image from matplotlib figure (as rendered locally) and use that as the figure, this way you get exactly wysiwyg , but the downside it is not interactive.Make sense ?

  
  
Posted 2 years ago

GloriousPenguin2 hmm the UI might strip it?! I mean in most case it should not be there in the first place. Maybe we need to make sure that if provided the web UI will use the stored plotly definition, if this is the case we need to make sure that by default we do not store it, so in most cases the UI can use it to improve the layout. wdyt?

  
  
Posted 2 years ago

GloriousPenguin2 could you open a GitHub issue on it? Just making sure this will actually get fixed 🙂

  
  
Posted 2 years ago

Also something we are very much interested in (including the logger-based scatter plots etc)

  
  
Posted 2 years ago

AgitatedDove14 sounds great!

  
  
Posted 2 years ago

AgitatedDove14
Sure!

Btw on "some layout properties should not be there". if you use px API of Plotly , create some figure and check fig.layout you'll see it's quite populated. It'll be empty if the low-level go API of Plotly is used.

  
  
Posted 2 years ago

Hi SuccessfulKoala55
Author of the figures above here

The problem is, that it doesn't help. Left one - IPython render (what a dev sees in the notebook). Right one - what gets rendered in the web UI. Also it's built in Plotly , not Matplotlib

the code that is used for axes:
... fig.update_layout(height=900, width=600) fig.update_xaxes( scaleanchor="y", scaleratio=1, row=1, col=1, ) fig.update_xaxes(range=[0.0, 1.0]) fig.update_yaxes( range=[0.0, 1.0], row=1, col=1, ... )

  
  
Posted 2 years ago

Hi ShortCoral15 !
Well, you can resize the plots either manually in the WebApp (hove over the separator) or programmatically using Matplotlib as a picture - see https://github.com/allegroai/clearml/blob/2ecf0b55f90f74a399270ae9f856b8950335f26b/examples/reporting/matplotlib_manual_reporting.py#L35

  
  
Posted 2 years ago
569 Views
9 Answers
2 years ago
one year ago
Tags