multiple line(style,color,width) in make_addplot()

[vinaychuri]

"https://github.com/matplotlib/mplfinance/blob/master/examples/addplot.ipynb" mentions about adding multiple columns to make_addplot() via a dataframe.
Ex: adding multiple lines for the (bbands,ADX,macd)

But the make_addplot() does not accept list of values for the linecolor,style,width etc.
Interestingly, the label keyword accepts list of lables

Since the function does accept multiple datasets, is it possible to color/style the lines in the same function call?
Ex:
mpl.make_addplot(
plot_df[["A","B" ,"C"]],
linestyle=["dashdot","dashed","dotted"],
width=[0.5,1,1.5],
panel=1,
color=["blue","green","black"],
label=["A","_" ,"C"],
)

[DanielGoldfarb]

You should call make_addplot() three times and put the results in a list.

Something like this:

add_plots = [
    mpf.make_addplot(plot_df[["A"]],linestyle="dashdot",width=0.5,panel=1,color="blue",label="A") ,
    mpf.make_addplot(plot_df[["B"]],linestyle="dashed",width=1,panel=1,color="green",label="_") ,
    mpf.make_addplot(plot_df[["C"]],linestyle="dotted",width=1.5,panel=1,color="black",label="C") ,
]

mpf.plot(data, addplot=add_plots)

[vinaychuri]

Hi,
Reason for the change as a python newbie:

Adding multiple correlated line charts ex: Bollinger bands etc . This can be done by providing list of series to mpf.make_addplot
But it does not allow to tweak the linestyle or color or widths for each of the series at the same time in the current version
The existing method was as per your example provided above.

Learnt about the git process of forking and created my fork "mplfinance_vc" that allows single string or list of values for [linestyle or color or width] within mpf.make_addplot function

1. I have tried to keep the existing logic used for the "label" argument that allows list of values in current version, most of the warnings.warn are reused, existing Validators are reused
2. BUT, If len() list of values for [linestyle or color or width] is < len() of data series provided , the code silently uses the first element of list. Need your guidance here as I am unsure how to fetch the "Default" values at this location

master...vinaychuri:mplfinance_vc:linestyleAsList

[DanielGoldfarb]

BUT, If len() list of values for [linestyle or color or width] is < len() of data series provided , the code silently uses the first element of list. Need your guidance here as I am unsure how to fetch the "Default" values at this location

My recommendation is is to raise a ValueError. The caller is trying to do something very specific. Falling back to a default value when the user intended multiple values, even with a warning, doesn't seem right. Let the caller know specifically that they messed up and how they can fix it.

---

In addition, here are a couple of concerns that copilot raised, that I partially agree with: These are not critical, but something to consider:

• Inconsistent validation: color and width validators are not updated to accept lists/tuples like linestyle was — passing a list for those will silently fail validation or error unexpectedly.
• Mismatch warning only for linestyle: Similar count-mismatch warnings were added for linestyle but not for color or width — inconsistent behavior.

[vinaychuri]

Thank you for the update. I checked the code and was able to fix them but few questions on the correct route to take

Option1

Current code change is happening within plot() which checks the [linestyle or color or width] with len(data.columns) and raises ValueError accordingly , and provides detailed error message on "why the error"
BUT the candlestick is still displayed without any additional plots but has allowed incorrect kwargs to reach plot()

Option 2

Make changes to the make_addplot() to ensure it checks the [linestyle or color or width] with len(data.columns) and raises ValueError.
But note that currently, make_addplot() does not have any variables to detect the len(data.columns) nor the type(data). This will have to be calculated and then sent to _valid_addplot_kwargs() for Validator to incorporate.

Validator then just prints "TypeError: kwarg "linestyle" validator returned False for value: "['solid', 'dashdot']"

1. This test had 3 columns to print with 2 linestyle and hence error'd out
2. Given the current setup of Validator function it does not mention anything about "why is the error"
3. But this does not allows incorrect kwargs to reach plot()
4. It is always a TypeError

Code:

ap0 = [ mpf.make_addplot(
                df[["bbl_20_2_0", "bbm_20_2_0", "bbu_20_2_0"]], 
                linestyle=["solid","dashdot"],
                # color=["blue","blue","red"],
                color="blue",
                width=[1,3,4],
                panel=0,
            )]
mpf.plot(df,type='candle',volume=True,addplot=ap0)

[DanielGoldfarb]

@vinaychuri Good Question! Sorry for the delayed reply; swamped with other projects right now. I want to look more closely at the code before I answer. I started looking, and noticed that even for the existing labels if nlabels < ncolumns: code we are just giving a warning, not raising (so not following my advice above). Give me a few days; hopefully next week I will have the time to look through the code more closely so as not to make a hasty decision.