Containerized travis builds
terraform commit history
vinyl stickers for putting somewhere physically at a track denoting slow/neutral/fast
color pallette i like: https://www.colorcombos.com/color-schemes/375/ColorCombo375.html
stickermule looks nice. 100 stickers x 3 designs. 1x1” circles. $118 total.
quantity per 3 designs total per unit 50 $105.00 $0.70 100 $118.00 $0.39
beverly hills high school is an interesting one. is there a way i can get a better geometry label? And is there a way i can get an accurate straights vs. turns measurements?
37.84m x 2 = 75.68
68.43m x 2 = 136.86
yields: - turn distance: 302.19 - straight distance: 97.81
Then table of all 1500m races with columns.
Leave a build payload in main website as is ci.json
Record build payload and deploy payload. Push to s3 bucket before finish. Very final step
Trigger secondary job which rebuilds ops website based on contents of s3 bucket
Travis pipeline, triggered job
Or lambda job triggered by s3 event. Concurrency 1 based on arrival of new build or deploy in the s3 bucket.
This looks promising: https://docs.travis-ci.com/user/build-stages
Ops site will contain build history, deploy history, and review apps
Would be good to get the build sha into the review app url, which I tried to do before. Which I think I can do I just have to script a deploy in the Pr build. Remove the existing deploy block. Do the deploy in the same script where I add the Pr comment. Needs an install of aws cli. This has to happen in the Pr build otherwise I don’t have the Pr number to build the url of the pr to comment on. The sha isn’t meaningful since it is a transient Pr sha but it will result in distinct urls and meaningful Pr comments which will continue to be usable after a subsequent Pr build happens. In case you want to review the previous iterations. Is there really no link between branch sha and Pr sha?
The ops build will pull down all of the build/deploy/review data in the s3 bucket. Then a hugo config will build the site off of that data.
Single and list templates * Build * Review apps * Review app deploys (nested) if no sha in review app url * Deploy
Index * Recent builds * Recent review apps * Recent deploys
Analytics * Activity summary by week, month, year * Duration trends * Success/failure trends
Current build produces all data, and prod, staging deployments. Into a new bucket. Laps run ops data bucket. Turn bucket versioning on.
Secondary Travis stage build consumes from this bucket, builds ops site, pushes ops site to s3 bucket website. Eventually, remove ci.json from main website. Footer sha on home page links to ops website
Ops.laps.run for my data Status and ops links on footer
Status updates rss auto tweet? Deployments rss to twitter
Get a basic site skeleton in place with a few of each data files pulled down. Sync down the data bucket to data dir See site locally Create bucket, website Create policy Create dns Deploy site manually See a few iterations go
Stretch create time series data datasource for grafana https://grafana.com/plugins/grafana-simple-json-datasource The simple json api needs to be able to respond to queries defined in a post body. That will be tough or impossible to do. Querystring might be possible If you can precompile queries based on common boundaries like days. Then limit the filtering you can do, it might be possible if you can get grafana to query with querystring or path. Holding for now
What cloudwatch metrics could be pulled in?
Csv load to google sheet. Pull in google sheet data via grafana or another dashboarding tool.
Time to revisit. What candidate categories?
Can we derive taxonomies? Like bucket groups by elevation range based on an elevation value?
Most of the fields I’ve seen involve a single specific value be framed as part of a list of values
Probably the best way I can grow viewership is to engage on social
Follow track teams, athletes, brands
Reply to tweets referencing venues with a link to my speed page
I want to have the physics component done by the start of next outdoor track season. Speed ratings by lane for 200/400m.
Setup insta Setup Facebook
Follow teams. Reply after meet results
Reply to flotrack.
Setup a calendar of meets to cover. Saturdays run down my checklist
latest: post to instagram with a screenshot of the new track, post to twitter, and facebook page via instagram app.
[ ] Tagging and taxonomies. This is how I will accomplish filtering. Templates
High school state championship venues
Qr code on stickers with a link to fast/neutral/slow list or definition
Physics of the turn
Pretty track right on sac river. Up/down stream of the iconic bridge. Adjacent to a marina off the river
Colors could be a good first taxonomy. multiple colors are possible (georgia tech, ucla). lets you group still. “give me every track with blue as a primary or secondary color”
I think my goal would be to maintain translations for key vocabulary, glossary terms. Journals would not be translated. glossary definitions could be translated be me or others
laps.run as a testbed for a11y
How to do an accessibility review
Lighthouse chrome dev tools audits gives a decent scan
Chrome Extensions, which give even more detail: - axe - wave
interesting, structural things to fix on the page
hard of hearing. limit audio cues.
Minimal static theme
https://teespring.com/about Sticker mule
Sticker mule links for fast/neutral/slow could link to top-level pages for those.
A definition followed by a table of tracks in that category
This track Is ranked
Fast/neutral/slow is a fixed quality based on speed rating tiers/ranges, not on the percentile they happen to fall into currently.
How to handle tracks shorter than 400 (franklin field)?
this feels similar to how we build speed rating for list views. we build a slice of tracks enriched with speed rating. then order by.
and this is related to how i can get country lists, which i still havent solved.
the key is that i want the ranking data available on the single.html view.
my first attempt will be to do the ranking for a single track in the single.html layout.
My hunch is coefficient of friction on an all weather track in metal spikes is 1
things to track on all builds:
find public/ -name '*.html' | xargs wc -l | grep total 329716 total
compare line count to count on current production. shouldn’t be greater than x percent.