Thanks. I wanted to finalize it as it took me already much longer than I expected
AgitatedDove14 no, I mean the exemple worked locally and on the server, but m, plot is shown only locally
AgitatedDove14 unfortunately I still have issues with the plot. After removing the first row I get a wierd empty remote plot where the axis is a counter instead of a date. Seems not to be clearml related and I need to get more in touch with plotly to analyze it.
AgitatedDove14 I'm just before finalizing the article and have issues now in getting a clone processed by an agent:(
The problem is plotly and something called orca. First it was not installed, than it was not able to start a browser as I run the agent as different user in an su - ... shell
I will try to use the plotly optionauto_open=False
Hi WickedGoat98
I try to write an article on medium about ClearML and face some a problem with plotly figures.
This is awesome !
I ran the plotly_reporting.py example locally and the uploaded plot was ok.
So are you saying the same example code from the repository worked okay on your server but showed nothing on the hosted server ?
Hey WickedGoat98
I found the bug, it is due to the fact the numpy (passed to plotly) contains both datetime and nan, and plotly.js does not like it. I'll make sure this is fixed, in the meantime you can just remove the first row (it contains the nan):df = pd.concat([tickerDf.Close, tickerDf_Change.Close_pcent], axis=1) df = df[1:]
Okay, I was able to reproduce it (this is odd) let me check ...
WickedGoat98 what's the clearml version you are using?
AgitatedDove14 the first plotly plot is fine and I added a second one which fails again 😞 I checked that the indes is of type str but the remote plot is again having integer as x-axis
The code is in github if you would like to check. I will proceed converting the index back and forth to find a way getting it running
WickedGoat98 until the next RC release (should not take long) this will solve it:df = pd.concat([tickerDf.Close, tickerDf_Change.Close_pcent], axis=1) df = df[1:] df.index = df.index.astype(str) setattr(df, 'ticker', args.symbol)
Basically removing the nan and converting the datetime to string representation (so plotly.js likes it)
WickedGoat98 Nice!!!
BTW: The fix should solve both (i.e. no need to manually cast), I'll make sure the fix is on GitHub so you'll be able to verify 🙂
WickedGoat98 is this related to plotly opening a web page when you call show()
method ?
You can do:if not Task.running_locally() fig.show()
Sounds good :) I'm currently trying to run an orca instance ... but without success
AgitatedDove14 I fixed it. The index looked like a string but wasn't ...
WickedGoat98 give me a minute, I'm not sure it is not ClearML related
AgitatedDove14 the index astype(str) did the magic 🙂 thanks
AgitatedDove14 I would like to publish additional 2 articles handling the use with Docker and Kubernetes. Docker I managed, but my Kubernetes knowledge is quite low. I managed to set-up K3S cluster which might be also worth an article, but I still habe not realy the understanding to add workers with agents to it...
It might take some time till I will write the Kubernetes stuff. Once I'm doing I will let you know
WickedGoat98 Same for me, let me ask the UI guys, I think this is a UI bug.
Also maybe before you post the article we could release a fix to both, what do you think?
EDIT:
Never mind 🙂 i just saw the medium link, very cool!!!
WickedGoat98 this is awesome! Let me know how I could help 🙂
BTW: I checked regrading the plot comparison, this is a BE issue due to the size of the plot, I was told a fix will be deployed in a day or two.
AgitatedDove14 it seems a comparison of plots the orig. & cloned experiment is not possible. Is this a bug on the server?
withif task.running_locally(): fig.show()
it works 🙂
thanks you for the support