Closed fred4ets closed 2 years ago
Hello,
Any news about this issue? I'm stuck with it :/
Thanks.
We're sorry @fred4ets but we haven't been able to look into this yet and we're not sure when we will be able to. Have you been able to investigate further? We are definitely open to external contributions which might fix this issue
I can shed some light I think - time axes in Chaco are for absolute times (ie. the value is seconds since the Unix Epoch in UTC) which are then converted to the current computer timezone for display - I think that the "negative" values are coming from the timezone conversion where the Epoch occurred at 1am in your timezone. But there is a question of why the zero baseline isn't being mapped the same way so there may be a bug here (if you're on Windows and in the UK there could be additional issues because of daylight savings experimentation in the UK in 1970 which Windows doesn't correctly handle). You might be able to fix by providing a starting_value
data source which is all zeros.
More deeply, what I suspect is happening is that the data you have are really timedeltas, rather than absolute times. Chaco doesn't currently have a timedelta scale system, so you may need to roll your own. The main benefit of doing this would be aesthetics around having tick marks at values like 15, 30, 45 rather than 25, 50, 75 for example, and displaying values as HH:MM:SS (although you could do this with a custom formatter). This is likely a fair bit of work to implement.
Hello,
Thanks for the hint, setting starting_value to -3600 fixes the issue.
Thanks a lot!
Regards
Hello,
I would like to display data using BarPlot but with vertical axis in time.
Please see the CME below to reproduce the issue shown in attached snapshots.
If you set time_scale_flag to False, display is good:
But if you set it to True, display should look like the first, but this is not the case, it is bad:
Am I doing something wrong here, or missed something?
Thanks in advance.
Regards,
Debian x86_64, Python 3.9.2, ETS source from git