Sorry to post here, but i couldn't find a better place... Github is really puzzling me !

Is there a way to NOT synchonize the Y axis ? I have very different Y axis scales and synchonizing makes the out of range curves disappear.

Thanks,

Add the feature of placing and displaying time labels according to UTC instead of local time. Provide a new option labelsDateUTC, and adapt the label and value date formatters, and the date tick computation accordingly:

• The option labelsDateUTC is coherent whith other options like labelsKMG.
• The date tick computation is simpler and homogeneous, and always places the ticks at nice instants.
• When using local time, ticks are properly generated on DST transitions (with some minor differences with regard to the original implementation).

This is a better implementation of some ideas in pull request #291, where only the labels are formatted according to UTC. In this version, the ticks are also placed according to UTC at nice calendar positions.

I have implemented multi-axis support which allows any number of independent axes to be defined. Only one axis can be shown at a time though, so it is by default coloured to that of the associated data series.

When the "showAxis" option is used, specifying a data series, separate independent axes are generated for each. This allows separate value ranges to be set for each data series as well as allowing some axes to be be defined with an integer y axis.

Please see tests/multi-y-axis.html for examples of it's use.

• Add "showAxis" option to display an independant Y axis.
• Add "colourShownAxis" option to colour a shownAxis the same as it's data series.
• add "integerYAxis" to only display integer y-axis labels.

Hey Dan and Robert,

this is the pull request for my proposal: https://docs.google.com/document/d/1IWmX4oDbQbVtUoRNzSRG3yMpoBQ7LseVCQhGnuOZz_A/edit?pli=1#

Lint runs without any issues and so do all of the tests. (Including additional ones I added)

As you can see I have done some performance tweaks and this solution is now is faster than the current version in every aspect. This implementation may still be extended and changed here and there but it is fully functional and introduces the new structure which fixes some bugs and is a great benefit for future feature development, at least in my point of view. I would really appreciate if we could think about pulling it in the near future as an internal change and only add external options later when you have the feeling that everything is the way it should be.

The problem for me in the moment is that I have to merge every changes you do on the master and this is quite a bit of work. I think my proposal is in the right stadium to be integrated and further developed on your code base.

What do you think?

Best Regards

David

Added new Features:

• The gridline options may now be set on a per-axis basis.
• Now both y and y2 can have separate gridlines.
• Added new gridline option "gridLinePattern" working the same as the strokePattern for the series.
• Added option description for the already present option "independentTicks"
• Added examples and tests to verify the new functionality.

This path fixes Issue 137 where once a graph it zoomed on the y-axis and the data updated, the zoom is lost.

It also has a fix for the zooCallback where the wrong y-axis data was being passed.

I have also added some flags: zoomed, zoomedX and zoomedY that simply show if a graph has been zoomed. I was previously able to detect a x-zoom programmatically but have found no way to do this for the y-zoom, so I added these flags which makes my code nice and simple.

I visualize central heating data which combines the on/off state of the heater with a temperature sensor. for the visualization of the on/off data I've written my own plotter, but because the source data is 0 or 1, while the temperature is around 20, the binary data should be excluded from ranging the y-axis

• Move tests to karma test runner and mocha framework, also adding chai assertions
• Add gulp tasks for testing, watching, concatting and compressing
• Update travis file to use new build system
• Remove jstestDriver stuff
• Reorganize a bit so that it's easier to see what's going on
• Create a dist folder to hold the output of builds (should probably look at having a bower branch/repository where these are held, instead of master)

I'm pretty sure we don't want to push this to master, but that's where this was diffed against.

This requires registering mousemove/mouseup on document, rather than the canvas. This is a bit scary! But the tests pass, and this is generally pretty well-tested code.

This gives a 200px buffer on any side of the chart.

Here's a demo.

Fixes #323.

From [email protected] on April 30, 2013 20:46:43

Hi there,

I'm using this wonderful library for plotting a fairly large amount of data (100,000 data points).

Dygraphs itself handles it very well, and it's all pretty fast and snappy --- until I enable the RangeSelector, which pretty much gets things grinding to a halt.

I know this is a very broad complaint -- you're welcome to say: well, then, fix it! -- but there seems to be a large discrepancy in terms of performance. Is this a know issue?

Thanks and best wishes,

Florian

Original issue: http://code.google.com/p/dygraphs/issues/detail?id=468

For stacked graphs, draw gaps for just the specific parts of series that are missing data, and use interpolation to help ensure that the overall graph shape doesn't get distorted for missing points.

This replaces https://github.com/danvk/dygraphs/pull/139 "Fix NaN handling for stacked graphs" which had aggressively propagated NaNs to ensure that graphs don't get drawn with incorrect stacking.

Fixes issue 446 - Improve handling of NaNs in stacked graphs.

Detailed changes:

• Remove layouts_.datasets, instead create point object arrays directly from gatherDatasets_(). This was necessary since the pre-stacked datasets aren't able to store the additional data needed for handling gaps.
• For stacked graphs, track yval_stacked/y_stacked separately from yval as point properties.
• Remove unstackPointAtIndex which is now no longer necessary, since the points keep the original yval. This helps ensure that the returned values keep their original NaN value and don't expose the interpolated stacked values used for drawing gaps.
• Remove evaluateWithError, evaluateLineCharts and Dygraph.seriesToPoints_ now handle error bars directly.
• In gatherDatasets_, remove the unconditional copy of rolledSeries, the refactoring appears to have fixed the bug related to zooming with error bars on. (I could reproduce it in the original code by zooming dygraphs/tests/custom-bars.html.) For date windows (horizontal zoom), use .splice() instead of manually coping.

What needs to be changed to make tests.js not contain dygraphs itself? (Other than adding the extra <script…></script> tag to also load dygraphs.js to the two HTML files.)

This separation would allow nice things, like running the tests against the minified version as well.

What was expected: g.updateOptions({ interactionModel: ... }) changes g's interaction model. What happened: Interaction model stays unchanged.

A comment in dygraph-interaction-model.js:625-633 suggests that using g.updateOptions() to change the current interactionModel is possible:

/**
 * Default interation model for dygraphs. You can refer to specific elements of
 * this when constructing your own interaction model, e.g.:
 * g.updateOptions( {
 *   interactionModel: {
 *     mousedown: DygraphInteraction.defaultInteractionModel.mousedown
 *   }
 * } );
 */

However, dygraphs.js' structure does not allow for the dynamic change of interaction model, as it is set up in createDragInterface_(), which is only ever called once, in __init__(), and not when options are updated.

Furthermore, createDragInterface_() calls into addAndTrackEvent(), and there is no way of removing a previously added event callback aside from removing them all with removeTrackedEvents_(); thus calling createDragInterface_() again is unsuitable. That said the event it adds are wrapped (to bind this), so could be extended to act as proxies to the current, uncached interaction model.

Dygraphs automatically detects if a CSV is the data source by means of checking the file name for a line break. If a line break isn't present, then it decides it's a CSV and looks for the file on the system.

Instead of having this automatic detection, it would make more sense to specify the data source.

This would make it possible to use JavaScript strings containing CSV data as a data source. Currently these variables are treated as files rather than strings and are searched for on the system. (returning 404)

I ask for the solvation of my problem to always show y2 axis.

Here is the article at stackoverflow without a result.

https://stackoverflow.com/questions/71143168/dygraph-y2-axis-was-never-shown

If I need different data series filled with data sets to show the right axis, then drawAxis is obsolete at the moment. But this shrink the functionality of dygraph to generate the same view to more than one dygraph in a site.

Next issue is related to the actual situation. If the y2 axis isn't shown, the dygraph automatically fill the space and disrespect the formerly set width of the canvas. The space of the second axis is not subtracted from the single axis dygraph.

These two problems in once steel me two weeks. Please fix it.

Hi. I saw there is an axisLabelFormatter option here to format the tick values that appear along an axis. It works just fine, however, I am wondering if there is a way to reset that option (go back to use the default formatter of Dygraphs) once I set it previously. I tried to set axisLabelFormatter to undefined or null but got this error in the console.

Uncaught TypeError: Cannot read properties of undefined (reading 'call')
    at getDateAxis (vendor.js:419283)
    at dateTicker (vendor.js:418918)
    at Dygraph.push.10681.Dygraph.addXTicks_ (vendor.js:422821)
    at Dygraph.push.10681.Dygraph.drawGraph_ (vendor.js:423184)
    at Dygraph.push.10681.Dygraph.predraw_ (vendor.js:422901)
    at Dygraph.push.10681.Dygraph.resize (vendor.js:424139)
    at resizeHandler_ (vendor.js:421653)
    at ZoneDelegate.invokeTask (polyfills.js:668)
    at Zone.runTask (polyfills.js:356)
    at ZoneTask.invokeTask [as invoke] (polyfills.js:820)

Any suggestion? Thank you.

https://dygraphs.com/tests/synchronize.html When deselecting the checkbox to unsync the Y axis; the graphs all reset when a new zoom is made. The Y-axis also zooms out on every selection in another graph. I would expect the Y-axis to stay the same when another graph is zoomed but the graphs just reset and redraw.

Shouldn't the behavoir be: when a graph is zoomed in the y-direction and another graph is zoomed in the x-direction, only the x-direction is updated in all of the sync graphs and leave the y-zoom alone?