Asking your advice

[MarcoRianiUNIPR]

Function getYahoo has many options.

% Optional input arguments:
%
% LastPeriod : Period to analyze. Character or string scalar.
% Possible values are:
% '1d' '5d' '1mo' '3mo' '6mo' '1y' '2y' '5y' '10y' 'ytd' 'max'
% Default is '1y'.
% Example - 'LastPeriod','6mo'
% Data Types - char | string
%
% interval : Sampling interval requested from Yahoo Finance.
% Character or string scalar.
% Possible values are:
% '1m','2m','5m','15m','30m','60m','90m','1h','1d','5d','1wk','1mo','3mo'
% Default is '1m'. If the LastPeriod/interval pair is not
% admissible, the interval is replaced automatically if
% autoFixInterval=true.
% Example - 'interval','1d'
% Data Types - char | string
%
% widthFactor : Candle width scaling factor. Scalar.
% Used only when topPanelMode='candle'.
% Default is 1.5.
% Example - 'widthFactor',1.2
% Data Types - double
%
% layoutHeights : Relative heights of the three panels. Numeric vector.
% Vector with three positive elements controlling the
% relative heights of:
% 1) top panel = price / candles / moving averages;
% 2) middle panel= volume;
% 3) bottom panel= technical indicator.
% The actual heights are obtained by normalizing the vector
% so that its elements sum to 1.
% Default is [1 1 1], which gives the same height to the
% three panels.
% For example:
% - [2 1 1] makes the top panel occupy about half of the
% plotting area and the other two panels about one quarter
% each.
% - [3 1 1] gives even more emphasis to the top panel.
% - [1.5 1 0.7] gives moderate emphasis to the top panel and
% reduces the space of the bottom panel.
% Example - 'layoutHeights',[2 1 1]
% Data Types - double
%
% removeGaps : Remove night or weekend gaps from the x-axis. Boolean.
% If true, the x-axis is based on the progressive index of the
% time series.
% If false, the x-axis is based on actual datetime.
% Default is true.
% Example - 'removeGaps',false
% Data Types - logical
%
% breakAtSession : Add separators between sessions. Boolean.
% Used only when removeGaps=true.
% Default is true.
% Example - 'breakAtSession',false
% Data Types - logical
%
% nTicks : Number of ticks shown on the x-axis when removeGaps=true.
% Positive integer scalar.
% Default is 8.
% Example - 'nTicks',10
% Data Types - double
%
% topPanelMode : Type of plot in the first panel. String or char.
% Possible values are:
% 'candle' = candlestick chart
% 'line' = close price only
% 'ma' = close price and moving averages
% Default is 'candle'.
% Example - 'topPanelMode','ma'
% Data Types - char | string
%
% maFastLen : Fast moving average length. Positive scalar integer.
% Used only when topPanelMode='ma'.
% Default is 10.
% Example - 'maFastLen',5
% Data Types - double
%
% maMidLen : Medium moving average length. Positive scalar integer.
% Used only when topPanelMode='ma'.
% Default is 30.
% Example - 'maMidLen',20
% Data Types - double
%
% maSlowLen : Slow moving average length. Positive scalar integer.
% Used only when topPanelMode='ma'.
% Default is 60. Note that maSlowLen > maMidLen > maFastLen.
% Example - 'maSlowLen',100
% Data Types - double
%
% showMACrossovers : Show moving average crossover markers. Boolean.
% Used only when topPanelMode='ma'.
% Default is true.
% Example - 'showMACrossovers',false
% Data Types - logical
%
% upColor : Color for up candles (top panel) and bars (mid panel). RGB row vector.
% Default is [0 0.7 0].
% Example - 'upColor',[0 0.5 0]
% Data Types - double
%
% downColor : Color for down candles (top panel) and bars (mid panel). RGB row vector.
% Default is [0.85 0 0].
% Example - 'downColor',[0.7 0 0]
% Data Types - double
%
% bottomPanelMode : Type of indicator shown in the third panel. String or char.
% Possible values are:
% 'rsi' = Relative Strength Index
% 'stoch' = Stochastic oscillator
% 'macd' = MACD
% 'williamsr' = Williams %R
% 'roc' = Rate of Change
% Default is 'stoch'.
% Example - 'bottomPanelMode','macd'
% Data Types - char | string
%
% topRSI : Upper RSI threshold. Scalar in the interval [50 100].
% Default is 80.
% Example - 'topRSI',70
% Data Types - double

% and many others

For example, the options

% maFastLen : Fast moving average length. Positive scalar integer.
% Used only when topPanelMode='ma'.
% Default is 10.
% Example - 'maFastLen',5
% Data Types - double
%
% maMidLen : Medium moving average length. Positive scalar integer.
% Used only when topPanelMode='ma'.
% Default is 30.
% Example - 'maMidLen',20
% Data Types - double
%
% maSlowLen : Slow moving average length. Positive scalar integer.
% Used only when topPanelMode='ma'.
% Default is 60. Note that maSlowLen > maMidLen > maFastLen.
% Example - 'maSlowLen',100
% Data Types - double

work only if you choose

'topPanelMode','ma'

as in

getYahoo('G.MI','topPanelMode','ma','maSlowLen',40)

but they have no effect if, for example, you choose

'topPanelMode','candle'

In other words, the option 'maSlowLen',40 has no effect in the call

getYahoo('G.MI','topPanelMode','candle','maSlowLen',40)

Now there are two possible approaches.

Leave everything as it is, and perhaps issue a warning for options that are unused.
Modify topPanelMode so that it can also accept a struct as input.

% topPanelMode : Type of plot in the first panel. String or char or struct.
% Possible values are:
% 'candle' = candlestick chart (with all default candle options)
% 'line' = close price only (with all default line options)
% 'ma' = close price and moving averages (with all default MA options)
%
% If, instead, topPanelMode is a struct, then it becomes
% possible to control the characteristics of the selected
% visualization.
%
% For example, if in the top panel you want to show moving
% averages with crossover signals, with fast moving average
% length = 15, medium moving average length = 25, and slow
% moving average length = 50, you can use:
%
% myoptions=struct;
% myoptions.Name='ma';
% myoptions.maFastLen=15;
% myoptions.maMidLen=25;
% myoptions.maSlowLen=50;
% myoptions.showMACrossovers=true;

and call getYahoo as follows

getYahoo('G.MI','topPanelMode',myoptions)

The same applies, for example, to the options

% upColor : Color for up candles (top panel) and bars (mid panel). RGB row vector.
% Default is [0 0.7 0].
% Example - 'upColor',[0 0.5 0]
% Data Types - double
%
% downColor : Color for down candles (top panel) and bars (mid panel). RGB row vector.
% Default is [0.85 0 0].
% Example - 'downColor',[0.7 0 0]
% Data Types - double

which apply only when the top panel mode is 'candle'.

What is your opinion?
Your feedback is highly appreciated.

[BoschiGiacomo]

This is a very interesting matter, and thanks for letting us join the discussion, here's what i think:

Modifying the function so that it accepts a struct as an input can be a very good choice, since a lot of other FSDA functions also accept options as a struct.

However, this doesn't fully remove the need for a warning to make sure not to surprise the users, since the function could still be called with the traditional name, value inputs, and still have the unused inputs problem.

So it depends on how much do you want to guide the users, if you want to guide them as much as possible, then the warning is the best choice, despite the struct input being a good addition.

[NourTerryS]

Thank you Prof.Marco for this discussion and the opportunity to contribute. I believe the struct approach is definitely worth the extra setup step because it eliminates the need for users to remember which options work with each mode. With proper documentation and clear examples in the help file, this becomes an intuitive and professional way to handle complex options.

[Samuele7F]

Thanks for the discussion. I support the struct approach and I agree with @NourTerryS .
To me, it’s a more natural way to organize options, making the function easier to extend in the future without adding confusion, and resulting in a cleaner and more flexible design.

[matteozanettii]

Dear Professor,

From the perspective of Digital Sciences and software scalability, I strongly support the struct approach . As the library grows, grouping mode-specific parameters into dedicated structures prevents the main function signature from becoming cluttered and unmanageable. It makes the code more modular and 'future-proof'.

However, to maintain the library’s robustness for all users, I believe we should implement a hybrid solution:

• The Struct as the new professional standard for advanced configurations.

• A Warning System for the traditional Name-Value calls. If a user provides an option that doesn't match the selected topPanelMode (e.g., maSlowLen with candle), a clear warning should explain why it is being ignored.

This way, we provide a 'fail-safe' environment for beginners while offering a more sophisticated, structured interface for advanced automation

[hiba-has]

I strongly support the Struct approach because it’s much more organized and clear. It removes the confusion of ghost parameters since options are only available when they actually apply to the chosen mode.
I don’t think keeping both options is a good idea, it just creates more confusion for the user and risks parameter conflicts. A better way for the input is to accept a simple string for defaults and a struct for customization.

[DomenicoPerrottaJRC]

I also fully support the structure option. And I have a new request. Imagine now we want to monitor "CL=F" (Crude Oil). The data in TT would come from the standard section of Yahoo finance:

Now, imagine that we rather want to analyse the data in the futures section, that is this one:

How to do it?

[MarcoRianiUNIPR]

@DomenicoPerrottaJRC the call

getYahoo('CLK26.NYM')

produces

Just use as a ticker the Contract Name and that's it!

Is this what you were looking for?

[MarcoRianiUNIPR]

@DomenicoPerrottaJRC your remark made me think that we could write a function such as getFutures which could be similar to getTickers but it is referred to the futures.

To all: please let us know if this function could be of interest

[DomenicoPerrottaJRC]

Looks OK, Marco. But I need to understand better the output now. I am not good in finance!