Hello gentlemen,

since I have discovered backtrader several weeks ago I am quite amazed by its capabilities and by its code quality. I still felt it would be nice to have better plotting integrated into backtrader.

In my opinion Bokeh (https://bokeh.pydata.org/en/latest/docs/gallery.html) is an awesome plotting library for Python. So I gave it a try to have backtrader make use of it. It turned out that it was not very hard thanks to backtrader's great flexibility and its modularized nature.
I also tried to integrade Plotly but it actually became pretty slow when using larger amounts of data.

I have implemented several of backtrader's internal plotting options but still alot is missing. Also it will need alot more optimizations to make it polished.

Demonstrations here:
https://verybadsoldier.github.io/backtrader_plotting/

Please find the code, installation instruction and demos here:
https://github.com/verybadsoldier/backtrader_plotting

I recently found time to work on this a bit again and finally reached something I call now version 1.0.0. I did alot of polishing and cleaning up the code base. All open issues on GitHub are fixed with this as far as I know.

Some additions include:

• being able to configure plotting completely independent from strategy code (which bothered me)
• configurable hover tooltip
• much easier optimization result browser
• etc.

It was not possible to edit the first post of this thread but the project is still available on GitHub :)
https://github.com/verybadsoldier/backtrader_plotting

I also tried my best to document the core components:
https://github.com/verybadsoldier/backtrader_plotting/wiki

I am very happy to hear your feedback and critics. Feel free to use also the issue tracker on GitHub:
https://github.com/verybadsoldier/backtrader_plotting/issues

I know the generated pages to not look very great as I am bad at designing. I would be happy if someone of you who is maybe good with css could give it a try spice it a visually. I would be happy to include an updated css scheme!

After alot of time I made a version 2.0.0 of this with some new features (e.g. live plotting) and many bug fixes. In case someone is interested, it is available on PyPi.

What would be the preferred way to package this? Currently the code is meant to include some plotting classes for backtrader. So currently I am thinking about two options:

• create a seperate package, e.g. backtrader_plotting. This would reflect the intent to be a package which provides general plotting extensions for backtrader. I am especially unsure if @backtrader would be fine with the name since it is very close to the original backtrader name and people could think it is an official package or in any way related to the official repository. I do not want to offend anyone.

• another option would be to try to get it into backtrader_contrib. But it seems the idea of the contrib repo did not really take off so far as it still has no commits. Also as I don't really know the background of the split, I am unsure if I understand its real intent. Also I am still hoping and begging that @backtrader will not abandon this great project :)

Updated version with support for optimization results.

Also updated project page on github with an ready-to-run example.

I agree the situation is not ideal. But I totally respect that people move on and don't enjoy maintaining projects forever. So yeah, the usual flow of things is that someone steps up willing to either get commit rights to the official repo or establish a fork which is considered as a new "official" repo (bt-next? bt-ng anyone?) :)

I have set up Travis CI on my fork to automatically run tests upon commit. Maybe it will help you:
https://travis-ci.com/github/verybadsoldier/backtrader

Both of the above has its own problems:

1. Most of the existing tests seems to be developed for manual run and inspection.

Hm, not sure. Which ones do you mean?
The ones that I am aware of can be run automatically, e.g.:
https://travis-ci.com/github/verybadsoldier/backtrader/jobs/295673688

2. Some parts of the code is hard to automatically test since interactions with external API could be required and there is no any mock interfaces or simulators available yet

Python's unittest package has quite good mocking features. So writing tests for external APIs shouldn't be hard. I think that is the standard way.

@kian-hong-Tan said in Bokeh Integration - Interactive Webbrowser Plotting:

@vbs tks for your reply. I tried commenting out the lines under metadata.py but still having the same error. Full stacktrace provided below:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
~\anaconda3\envs\env_backtrader\lib\site-packages\backtrader_plotting\html\metadata.py in _get_strategy(strategy)
     67 
     68         # md += 'Source Code:\n'
---> 69         # md += f'\n```\n{inspect.getsource(strategy.__class__)}\n```\n\n'
     70 
     71 

Well, according to your stacktrace the code gets execute even though commented out :) I have no good idea about that, sorry. I would suggest to debug that. That line should not be executed.

Hello gentlemen,

I would like to integrate Bokeh plotting for optimization results but I have a problem interpreting optimization results when using multiple strategies.

Let's say I have this test code:

cerebro.optstrategy(TestStrategy, buydate=range(14, 16, 1))
cerebro.optstrategy(TestStrategy2, buydate=range(11, 13, 1))
cerebro.addanalyzer(bt.analyzers.TradeAnalyzer)
data = ...
cerebro.adddata(data)
res = cerebro.run()

Now I have a list of 4 entries in the variable res (all possible strategy combinations), which is fine for me. Each of these entities contain a list 2 OptReturn objects. One for each strategy involved.
But how can I know which of these OptReturn objects belongs to which of the two strategies? I mean: I have 2 OptReturn objects and I have 2 strategies. They only contain the set of optimization parameters and a list of analyzers.

Thank you!

Ok thank you, that's great. I have put it here for now:
https://github.com/verybadsoldier/backtrader_plotting/tree/develop

It is really work in progress currently. Anything is questionable :) But feel free to test it and I am happy about feedback. As I have nearly no experience with actually using backtrader yet, I am especially interested in feedback about what is missing or should be changed to make this really useful in an everyday productive environment.

I have put some demos here which show the two currently available plotting modes (Single and Tabs) and also the two built-in schemes (Blackly and Tradimo, basically dark and bright):
https://verybadsoldier.github.io/backtrader_plotting/

After alot of time I made a version 2.0.0 of this with some new features (e.g. live plotting) and many bug fixes. In case someone is interested, it is available on PyPi.

@lclassen
As far as I can remember a cerebro might have problems getting pickled for 2 or 3 reasons. You need to implement the manually pickle trigger functions exclude those attributes from pickling.

Ok thanks. I wanted to add and update information in the first post as lots of it is outdated.

A general question: Does someone know how to edit posts on this forums? Quite sure a while ago I managed to do that. Did something change or ist it generally not possible anymore?

If you think there is a bug you can always post minimal runnable code that reproduces the problem and I will look into it.

I am sorry, since I am not using notebooks I can not really help here. Interesting is that someone not using notebooks was having the same problem about the missing toolbar:
https://community.backtrader.com/post/13521

I think that highly depends on how you implemented it but in general I would say: no.

Well, that is interesting. The point is that we are talking about a static HTML-file. So I am wondering if the HTML is correct but you still don't see the toolbar if you the HTML you generate differs from mine.

Do you see the toolbar when you navigate your browser here?
https://verybadsoldier.github.io/backtrader_plotting/demos/blackly_single.html

Could you please run the demo "blackly_single" and send me the output HTML?
https://github.com/verybadsoldier/backtrader_plotting/blob/master/demos/blackly_single.py

@Yang-Fu said in Bokeh Integration - Interactive Webbrowser Plotting:

@vbs Thank you for the quick reply! I did look on the right side of the plot, but no toolbar there. Also it doesn't plot in tabs even if I use the "multi" option. feel like something weird is going on with my setup. I am using windows10 + bokeh 2.0.2 + backtrader_plotting 1.1.0.

I do see in backtrader_plotting requirements says bokeh ~= 2.0.0. Wondering if anything in 2.0.2 breaks it?

Requirement already satisfied: bokeh~=2.0.0 in c:\users\yang fu\appdata\roaming\python\python37\site-packages (from backtrader_plotting) (2.0.2)

b = Bokeh(style='bar', plot_mode='multi', scheme=Tradimo())
cerebro.plot(b)

I dont't think that Bokeh 2.0.2 breaks it. Cannot say whats wrong there. I would suggest you start with a new/clean Python environment to test with.

@kian-hong-Tan said in Bokeh Integration - Interactive Webbrowser Plotting:

Hello @vbs , sorry to bother you again. Would like to ask you another question. The Interactive Plot is now shown in the cell (jupyter notebook) instead of a new browser. Is this the expected behavior? Below is my code. I have tried all output mode - show, memory, save, but I also couldn't see any file being saved.

Sorry I have never used notebooks and cannot help here.

@Yang-Fu said in Bokeh Integration - Interactive Webbrowser Plotting:

Hi @vbs, Excellent work! Really grateful for this nice tool. I do have a question about zoom, how can I do that? I see in the earlier replies you said there will be a toolbar on the left, but I don't see one in my plots.

Take a look at one of the demos please:
https://verybadsoldier.github.io/backtrader_plotting/demos/blackly_single.html

You will find the toolbar on the right (a bit cropped sadly):