New example - custom axis tickformatter #3314
Conversation
By defining a custom TickFormatter model, we are able to remove the weekend/holiday gaps in the candlestick financial chart. Dates are formatted into an array of Strings, e.g. ['Sep 01', 'Sep 02', ...], which is passed to the custom ticker as a property and then rendered on the axis. The downside to this approach is that we lose the dynamic nature of Datetimei. This can be overcome by making DateGapTickFormatter receive more than one list of Strings, each one representing a different Date format, e.g. List 1 = ['Sep 01', 'Sep 02', ...]; List 2 = ['Sep', 'Sep', ... 'Oct', ...] and so on. This remains as a suggestion only and is not implemented by this commit.
|
@lbrito1 quote note, our automated image diff testing machinery assumes the output filename matches the python filename, e.g. |
|
@bryevdv Thanks, fixed that. |
| """ | ||
|
|
||
| df = pd.DataFrame(MSFT)[:50] | ||
| df = df.set_index([range(0,len(df))]) |
There was a problem hiding this comment.
You should wrap range(...) in list() for python 3, or maybe use six.
There was a problem hiding this comment.
Yup this is causing the one remaining test failure: https://travis-ci.org/bokeh/bokeh/jobs/96149634#L1032
|
One other suggestion I think some of the description here should go in either a comment at the top of the example or in comments spread throughout. |
Refactored date_labels definition using list comprehension instead of map + lambda. Removed df index setting. The pandas df imported from bokeh.sampledata.stocks.MSFT has integer indices, so we don't need to define them manually with df.set_index. This should be done only if the df doesn't have an integer index. Also added some comments spread out through the example.
|
Thanks for the suggestions @bryevdv . I realized the MSFT df used in the example actually has integer indices by default, so I also changed the lambda map to list comprehension and added some comments throughout the example. Not sure what the commenting policy is here (if any), so if the comments are subpar in any way let me know. |
|
LGTM @bokeh/core any final thoughts? I will merge tomorrow if not. |
| # Strings (e.g. ['Sep 01', 'Sep 02', ...]) passed to the date_labels property. | ||
| class DateGapTickFormatter(TickFormatter): | ||
| date_labels = List(item_type=String) | ||
| __implementation__ = """ |
There was a problem hiding this comment.
- Indentation is off, should be 4 spaces.
- For readability, one empty line before
__implementation__should be added. item_typeis unnecessary,List(String)is the proper style (which should resemble scala'sList[String]syntax).
|
I pointed out some minor stylistic complaints. Otherwise LGTM. |
Fixed indentation and improved coffeescript for loop using one-liner syntax
|
@mattpap Nice, I wasn't aware of Coffeescript's one-line loop syntax. I updated the branch with your suggestions. |
|
Thanks for the great PR @lbrito1 ! |
New example - custom axis tickformatter
damianavila
added
reso: completed
and removed
reso: completed
tag: component: examples
type: task
labels
issues: fixes #3193
By defining a custom TickFormatter model, we are able to remove the weekend/holiday gaps in the candlestick financial chart. Dates are formatted into an array of Strings, e.g. ['Sep 01', 'Sep 02', ...], which is passed to the custom ticker as a property and then rendered on the axis.
The downside to this approach is that we lose the dynamic nature of Datetimei. This can be overcome by making DateGapTickFormatter receive more than one list of Strings, each one representing a different Date format, e.g. List 1 = ['Sep 01', 'Sep 02', ...]; List 2 = ['Sep', 'Sep', ... 'Oct', ...] and so on. This remains as a suggestion only and is not implemented by this commit.