If ax.disable_x_index() is set up then the chart is disappeared after update_gfx()

[fmvin]

• [x] I realize finplot is not a web lib. (Hint: it's native!)
• [x] I've read the snippets and not found what I'm looking for.
• [x] I've searched for any related issues and avoided creating a duplicate issue.
• [x] I've updated finplot (pip install -U finplot).
• [x] I've supplied the required data to run my code below.

Code to reproduce

def updates():
    # data is updated on each call
    if first_call_flag:
          plots={}
          plots["data"] =fplt.plot(data, width=3, color='magenta', legend='data', style='-',ax=ax)
          fplt.refresh()
    else:
          plots["data"].update_data(data, gfx=False)
          for k in plots:
              plots[k].update_gfx()
    first_call_flag = False
#
first_call_flag = True
plots={}
ax = fplt.create_plot('Test', rows=1, init_zoom_periods=300)
ax.disable_x_index()
fplt.timer_callback(updates, 2)
fplt.show()

Describe the bug

I borrowed chart update logic from the complicated.py example. The X axis is not a time series. Correct plot is shown first time if first_call_flag is True. But after calling update_gfx() the plot becomes empty. Also any realtime plot's updates produce the same result if ax.disable_x_index() is used.

The data looks like this:

             vol
price             
83000        33.18
83500        31.14
84000        29.13
...

price is a index and X axis values, vol is Y axis values

Expected behavior

To see updated data on the plot.

Reproducible in:

OS: Windows 2019 Server finplot version: 1.9.3 pyqtgraph version: 0.13.3 pyqt version: 6.6.1

[highfestiva]

I attempted setting ax.disable_x_index() in the complicated.py example, and that still worked. Please supply a complete, minimal example (containing the required data, if any).

[fmvin]

@highfestiva, for example, I changed line 35 in overlay-correlate.py to:

fp=fplt.plot(dfc, style='o', color=1, ax=ax2)
fp.update_data(dfc,gfx=True)

Dots on subplot 2 have disappeared.

[fmvin]

@highfestiva , this is complete, minimal example. Please comment line 13 ax.disable_x_index() to see a difference.

import pandas as pd
import numpy as np
import finplot as fplt
#
smile = None
def update_data():
    global smile
    smile=smile**1.04
    plots["one"].update_data(smile, gfx=False)
    plots["one"].update_gfx()
#
ax = fplt.create_plot('Smiles', rows=1, init_zoom_periods=300)
ax.disable_x_index()
plots={}
x=np.arange(-20, 20, 1, dtype=int)
y= x**2*1e-2
x=83000+x*500
smile = pd.DataFrame({'price': x, 'vol':y})
smile = smile.set_index('price')
plots['one']=fplt.plot(smile, width=3, ax=ax)
fplt.timer_callback(update_data, 2)
fplt.show()

[fmvin]

Also I noticed that on X axis some labels are missed.

[highfestiva]

Ah, check if 2df5abda6d076222a356b737f5149eb904274dc2 does the trick.

[fmvin]

Good. Charts are shown after update with ax.disable_x_index(). But still some X-axis labels are missed and zoom works strange. I need a time to formulate zoom problem if it's not related to my code.

[highfestiva]

It's caused by pyqtgraph.AxisItem.tickValues() default implementation. As you see in finplot's __init__.py:118, I check if it isn't x-indexed, and if so, I use pyqtgraph's default implementation. If you like, you can inherit the finplot.EpochAxisItem class and override the def tickValues(self, minVal, maxVal, size) function. Then you only need to set the finplot._create_axis method so your override is used instead of the default finplot.EpochAxisItem. GL!

[fmvin]

@highfestiva, please help me with the zoom problem I mentioned before. In the example above the one line was removed to prevent data modifications.

smile=smile**1.04

The problem is if zoom in is made by right mouse then on next update the zoom is reset automatically. How to make the zoom remain in a state I selected?

[highfestiva]

Perhaps you're looking for finplot.y_pad is defaults to 0.03? I.e. 3% margin at the top and the bottom of the charts?

[fmvin]

finplot.y_pad=0.03 by default in my environment. I've tested different values for y_pad right now with no success. Zoom returns to default state back on each update_data() invoke.

[fmvin]

"A picture is worth a thousand words". This gif demonstrates the problem.The zoomed area resets after each data update.

[highfestiva]

This zoom comes from pyqtgraph, and is not removed by finplot. But also no particular support has been added to it. To disable auto-y-zooming, you could temporarily set ax.vb.update_y_zoom to a dummy function, so no zooming takes place after you manually change the zoom area in both X and Y. You would then have to reset it whenever you'd want to use the auto-y-zoom feature again.