.. so available for current ride
.. and refreshed() when notifyDataChanged() is called
.. but not many classes call it!
.. we are moving to having the filecache read from the
RideCache rather than straight from disk so we have an
in-memory and persisted version.
... create folder structure for existing and new athletes
... upgrade existing athlete to new structure when opening / incl.
logging and error handling
... convert any "Downloaded" file to .JSON directly - store source file
in /downloads
... convert any "Imported" file to .JSON directly - store source file in
/imports
... introduce new "home" Folder structure object representing the
previous flat-folder
... adjust relevant QDir home - access to use the sub-folder method of
the new object
Note: the new structure is not yet active ! - everything is written to
the main folder still
To Do:
... add upgrade procedure for existing folder content to merge to the
new folders
.. RideFile - add rounding approach for gear values (different roundings
depending on value) - leading to discrete GearRatio values
.. AllPlot - exclude Gear Ratio from Smoothing (to keep the discrete
value) - and change curve type to "Steps"
.. ScatterPlot - specific handling for GearRatio since values between
0.01 and 1 are relevant for GearRatio and must not be filtered out
.. RideFileChache - add Distribution Data for Gear Ratio and fix
Distribution Cache for DataSeries with > 0 decimalsFor
.. Histogram - add GearRatio to Histograms (both Rides and Trends)
.. added the code to aggregate power data when deriving
model parameter history over time.
.. need to get a class that applies the various models to
generate the parameters we need (so we can reuse) in
part 2.
.. then need to store and turn into a metric of some kind
in part 3.
.. this is largely a checkpoint commit to enable cross platform
testing with Mac and Windows.
.. to extract just the meanmax power data from a ridefile cache
so we can aggregate and derive model parameters for new metrics
to track W', CP, FTP, max-Power.
.. by checking the ride file CRC before recomputing
either the metrics or the ridefilecache.
.. this means that users that routinely copy or backup
or use dropbox to keep things in sync won't see any
unneccessary metric computations.
.. to indicate which power-durations are getting worked hardest
.. Calculate a 'heat' score for the power duration curve.
A score that counts how many rides in the aggregate
have a power-duration best that is within 10% of the
best value calculated.
.. next step is to add it to the CP curve (when showing
power) to provide some kind of shade/heat indicator
for the area of the curve.
.. simplified acceleration calculation to convert to m/s before calculation
.. removed distribution for deltas as they were HUGE and introduced a
terrible performance degradation where aggregation took >30s for a
ride set that previously took <5s.
.. Following on from the recent update to add acceleration
this update adds other derived data series based upon the
rate of change.
.. Added to the ride plot and the CP plot.
Not sure of the overall utility of these updates but bear in
mind that they are targetting sprinting and track users and
analysis.
As well as the positive side of this (development of power
cadence etc) we also want to think about and collect data
on fatigue rate (possibly only power and torque)
- fatigue over time
- fatigue over pedal stroke
.. now can show in zones but using the polarised zones
rather than user defined zones.
.. we may need to revisit this since it uses zone 2 rather
than estimate LT1 from CP.
NOTE: The compare mode needs updating to support this option.
When an interval is added to the compare pane we now
compute the RideFileCache so it can be re-used in the
Histogram and CP charts without computing it each time.
Peak aPower durations now computed and cached;
* updated RideFileCache to work with aPower
* CP curve now allows you to plot aPower
* LTM plot allows you to plot aPower peaks
* DataFilter allows 'best(apower, duration)' function
Updated RideFileCache and LTMWindow/Plot to
plot the user defined durations for peak power.
Actually we support all the different meanmax
measures available in the CP chart, so you can
now track peaks for user defined durations of;
* Power
* Watts per Kg
* xPower
* NP
* Heartrate
* Speed
* Cadence
* Torque
* VAM
I have not spent much time optimising for performance
but reusing the same 'seek' approach used by the recent
update for best() and tiz() functions in the datafilter.
Fixes#322.
.. to make sure the v2 LTMSettings
uses the bestSymbol.
Just pushing this (WIP) to make sure
we don't get half baked LTM settings
from people building against the previous
commit.
.. forgot that the values are converted to
zero decimals when stored.
This fixes strange results when filtering by
wpk or speed, or any series that has decimal
places.
Added static functions to ridefilecache
to get best and tiz values for series and duration
or zones and glued them in to the datafilter code
recently added.
It uses seek (via QDataStream::skipRawData) and is
quite fast on my Linux development PC.
We now need to update the LTM settings dialogs to
the new mockups so we can plot custom durations for
peak power / NP etc too.
NOTE: Not tried on Windows / Mac, so may not be quite
so fast there, will need to check.
To try this out, in the search filter box type:
best(power, 3600) > 250
to filter only those rides where the best hour
power is greater than 250 watts. It should
return almost instantly.
Decoupled classes from MainWindow to reference Context
and Athlete (and introduced a couple of new headers).
We no longer pass around a MainWindow pointer to children
but pass a context instead.
There are still a few pieces left in MainWindow that need
to move to a better place;
* Setting/clearing filter selection
* Working with Intervals
* Adding/Deleting Rides
* Save on Exit
As mentioned previously there are lots of other parts to
this refactor left to do;
* break MainWindow Gui elements into Toolbar and Views
* migrate from RideItem and Ridelist to ActivityCollection
and Activity classes that are not tied into gui elements.
* introduce Application Context and AthleteCollection
Breaking the MainWindow 'god object' into
separate classes for Athlete and Context.
Further updates will need to;
- break MainWindow Gui elements into Toolbar and Views
- migrate from RideItem and Ridelist to ActivityCollection
and Activity classes that are not tied into gui elements.
- introduce Application Context and AthleteCollection
Once these are done we will be in a position to decouple
most classes from mainwindow and also introduce tabbed
athletes.
When the user free text searches or applies a data filter in
the top right toolbar search box it gets applied to any charts
that plot data from many rides, including;
- CP curve
- Calendar
- Histogram
- LTM
- TreeMap
- Summary
I need to update on a Mac to no longer use a Mac text search box
but now use a search filter box on the top right.
This reverts commit 6fdc55528f.
The performance decrease is considerable. Using mallocs across
my personal ride collection takes 5:19 to compute all the caches.
Using the QVector equivalent took 22:40.
Need to consider looking at QVector across the GC codebase.
If an incore cache of an aggregated RideFileCache is invalidated
by a ride in the same period being edited it does not clear
until the cache array is cleared (25 of them so likely to be
never) or upon a restart.
This fixes them being invalidated by refreshCache.
