-
Notifications
You must be signed in to change notification settings - Fork 639
Archive
Daniel Goldfarb edited this page Apr 3, 2020
·
1 revision
I was using wiki to track work-in-progress, and work done, on mplfinance
. But then I decided it's a lot simpler to do that using the list of "Issues".
Now, so as not to loose anything in case there is any info there I want to refer back to at a later date, just gonna archive that page here:
- Research into External Axes mode to be able to create unlimited subplots:
(Perhaps through different API namempf.axplot()
, perhaps justmpf.plot(data,ax=axes,...)
External Axes mode REQUIRES:- If volume is not None, then volume must be an axes for volume.
- All
addplot()
calls must have also specified their own ax=Axes- is secondary_y used just to position the y-ticks/labels, or ignored.
- certainly we don't
twinx()
??
- Initial experiments indicate that style is not applied if axis is created first, therefore maybe we provide
mpf.subplot()
to create subplots with the style already applied, and then continue as noted above passing in the axes that were created.
- ax.set_title('title') via
mpf.addplot(title=...)
if/when subplots become available - third panel subplot (not this time around)
- In README.md CHANGE
readme_files/readme_29_1.png
TOhttps://raw.githubusercontent.com/matplotlib/mplfinance/master/readme_files/readme_29_1.png
(and re-publish to Pypi) - minor wording changes to
addplot.ipynb
- "Notice also that we pass an alternative ..." (just above cell [16])
- "This is done by passing information [get rid of word
necessary
] (just below cell [2]) - cell [17] indicate "# this calculation has no financial meaning whatsoever; just generating some data to modify the order of magnitude and range to be able to demonstrate
secondary_y
on the main panel"
- Customizing the Appearance of Plots (colors, date format, style, etc)
- style (colors, etc)
-
secondary_y=[True|False|auto]
(default isauto
) for both panels, and for each call toaddplot()
- README.md : provide a link for
examples
folder, at the end of "Basic Usage" just before "Release Notes" -
no_xgaps
deprecated in favor ofshow_nontrading
-
block=False
for non-blocking call tompf.plot()
- Customizations:
- ability to modify plot aspect ratio (not just scale size)
- customize Title and Axis Labels
- Add a "deprecation" warning for
mpl-finance
APIs. (Can these be coded to write out only once after import?)
- Technical Studies
- which ones? integrate with tulipy?
- Prioritize Studies:
- Bollinger
- RSI
- Ichimoku
- Candle Patterns
- Renko
-
noshow=True
and return fig and [axes].
or, alternatively, acceptaxis=ax
on which to plot OHLC or candlestick data. - Fix and/or Replace old matplotlib finance documentation:
https://matplotlib.org/api/finance_api.html - Show trading signals on plot
- Mouse-over displays plotted values
- (maybe) Automatically determine if
addplot
values require a secondary_y (twinx). - try in
_check_and_prepare_data(data)
around accessing DataFrame keys: if exception, check for whitespace in column names and raise key-error exception with more specific error message (unexpected white space in column names for 'Open', 'High', 'Low', 'Close' and/or 'Volume'). - Clean up axis labels
- Title for Figure.
- Axis Labels
- Legend (?)
- Volume Profile: https://stackoverflow.com/questions/58517234/using-a-charting-library-to-overlay-volume-profile-on-a-candlestick-chart-in-pyt
- README: use absolute (vs relative) links in "Contents" section, so links will work on PyPI.
- keep in mind this will mean links on PyPI will send user back to analogous place in README on GitHub, due to the limitations described here: https://github.com/pypa/warehouse/issues/4064
- README: near bottom, link to examples folder (not to a specific example notebook).
Better yet remove this section and create an Examples section (with its own internal link) near the top of the README.- clean up examples folder (create
scratch
subfolder).
- clean up examples folder (create
- Release Notes: fix typo in date 2019-13-20 -> 2019-12-20
- Ability to Save plot to File (png, etc)