Important
This project depends on uom-plugin-0.3.0.0 that was constrained to use
ghc <= 8.2. A new version, uom-plugin-0.4.0.0, has been released that
works with ghc >= 9 so this project could now be upgraded.
Flare Timing is a reference implementation of GAP scoring for cross country hang gliding and paragliding racing.
Its command line console apps, one for each step in scoring, write down their
workings along with their outputs. With these workings we can trace how
a pilot's score has been calculated. There's no visual competition editor. The
one file that defines a competition can be generated from an FS database, the
*.fsdb file. It includes a web app that can be hosted locally for visual
checks and comparisons with expected or official results. It is possible to
publish the data alongside this web app standalone as done at flaretiming, the
web site.
Download the source, build and install the command line apps with stack or with ghc and cabal that can be installed with ghcup:
> git clone https://github.com/BlockScope/flare-timing.git --recursive
> cd flare-timing/lang-haskell
# with stack
> stack install
# with cabal supplying options --installdir and --overwrite-policy
> cabal v2-install all:exes
There's more in the building guide. There's a guide for testing too.
Let's get the inputs and outputs from FS, do the scoring and host the comp locally.
Start by preparing the FS comp file, the *.fsdb. This will often contain
sensitive personal information such as birthdays, phone numbers and notes that
we'll want to avoid publishing. We'll also want some output data for making
comparisons between flare-timing and FS.
- Clean out the sensitive pilot personal stuff we don't need with
fs-clean. - Trim away data we don't need with
fs-trim. - Grab the optimal route around the tasks found by FS with
fs-route. - Grab the arrival times and positions from FS with
fs-arrival. - Grab the landouts from FS with
fs-effort. - Grab the scores from FS with
fs-score.
That's the *.fsdb file done with. From here on, flare-timing deals with the
trimmed XML, the *.trim-fsdb.xml. If we have the track logs we can score the
tasks:
- Extract the inputs with
ft-extract-input. - Trace the shortest path to fly a task with
ft-task-length. - Find flying times with
ft-fly-time. - Find pairs of fixes crossing over zones with
ft-cross-zone. - Interpolate between crossing fixes for the time and place where a track tags
a zone with
ft-tag-zone. - Unpack the
*.igcor*.kmltracklogs withft-unpack-track. - Peg the timing window to a reference frame with
ft-peg-frame. - Index fixes from the time of first crossing with
ft-align-time. - Discard fixes that get further from goal and note leading area with
ft-discard-further. - Draw out leading areas
ft-lead-area. - Mask a task over its tracklogs with the following, run in any order:
- Group and count land outs with
ft-land-out. - Group and count far outs with
ft-far-out. - Score the competition with
ft-gap-point.
To get the backend server for hosting the comp data running locally:
Start the server with
ft-comp-serve.
To host the frontend web app for the comp locally:
- Change directory:
> cd lang-haskell - Open a try-reflex shell with:
> reflex-platform/try-reflex - Build the frontend and start its webpack dev server with:
> ../stack-shake-build.sh view-start-ghcjs - Open a browser at the hosted URL, usually http://localhost:9000/app.html.
Documentation is available online at flare-timing.readthedocs.io and there's a worked example too.
The GAP rules have changed over the years. Here are the features that flare-timing includes or not.
- Scoring Method
- GAP
- GAP2000
- GAP2002
- OzGAP2005
- GAP2007
- GAP2008
- GAP2011 tested with:
Forbes 2012 , Forbes 2014 , Forbes 2015 - GAP2013 tested with:
Forbes 2017 - GAP2014
- GAP2015 tested with:
Big Spring 2016 , Green Swamp 2016 , Green Swamp Sport 2016 , Forbes 2016 , Quest 2016 - GAP2016 tested with:
Dalmatian 2018 , Dalby 2017 , Forbes 2018 , Forbes 2019 - GAP2018 tested with:
Dalmatian 2019 , Italy 2019 - GAP2020
- PWC (GAP variant)
- PWC2007
- PWC2008
- PWC2009
- PWC2011
- PWC2012
- PWC2013
- PWC2014
- PWC2015
- PWC2016
- PWC2017
- PWC2019
- Linear distance
- Time-based scoring (TBS)
- GAP
- Earth Model
- FAI sphere
- WGS84 ellipsoid
- Distance Method
- Pythagorus on a UTM plane
- Haversines on the sphere
- Vincenty on the ellipsoid
- Andoyer on the ellipsoid
- Type of Task
- Race
- Elapsed time
- Open distance (can be declared but not yet scored)
- Shape of Zone
- Cylinder
- Inverted cone (can be defined but treated as a cylinder)
- Shape of Goal
- Circle
- Line
- Final Glide Decelerator
- Conical end of speed section (CESS)
- Arrival altitude time bonus (AATB)
- Source of Altitude
- GPS
- Pressure (QNH)
- Validities
- Task (day quality)
- Launch
- Distance
- Time
- Stop
- Points
- Linear distance (reach)
- Distance difficulty (effort)
- Arrival position
- Arrival time
- Time (speed)
- Leading
- Departure
- Leading Area as a Function of Time and Distance Tweaks
- Use distance; a = t * d
- Use distance squared; a = t * d^2
- Use PWCA weighting; a = w(t, d)
- Parameter Tweaks
- Day quality override
- 1000 points for winner if no pilot made goal
- 1000 points for winner before day quality applied
- Double leading points weight
- Proportional leading points weight if no pilot made goal
- Adjustable stopped task bonus glide ratio (fixed at 4:1 for PG and 5:1 for HG)
- Special Cases
- End of the speed section but not goal
- Early start
- Stopped tasks
- Stopped Tasks
- Stopped task time as announcement minus score back
- Requirements checking, goal or duration
- Score time window
- Time points for pilots at or after the end of the speed section
- Distance points with altitude bonus
- Penalties
- Absolute
- Fractional
- Jump-the-gun factor
- Jump-the-gun maximum
- Made ESS but not goal
- Task Ranking
- Overall
- Female
- Country
- Competition Ranking
- Overall
- Female
- Country
- Ties
- Fixed Total Validity
Copyright © Phil de Joux 2017-2020
Copyright © Block Scope Limited 2017-2020
This software is subject to the terms of the Mozilla Public License, v2.0. If a copy of the MPL was not distributed with this file, you can obtain one at http://mozilla.org/MPL/2.0/.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent