• Sunday

    • Got back from evansville indiana (via nashville tennessee) for the DII march madness finals. Amazing stage/outcome.
    • Got the 8-pod trays, reached out to Rise for (any) rearrangement.
    • American Cup finally finished. Hikaru and Krush.
    • Private work.
  • Friday

    • Private work.
    • Kwon charges, coinbase regulatory fights, tiktok congress, dorsey’s block tumble.
    • With both github and gpt4 under microsoft, Copilot now integrates its AI with developer workflows.
    • Flatiron auctioned for $190M.
    • Played with the new route53 console UX. Also transferred a registered domain between aws accounts.
      • Gotcha: the name servers of the hosted zone and the registered domain must match. Since these were created at different times, I had to sync them.
    • Gitlab web IDE can run a terminal, but you need to provide the remote connection. They don’t give you a free compute env.
    • Created a TrueBlue account (one of the few I didn’t have yet). Bought a ticket to Nashville – heading to Indiana early tomorrow morning for the DII March Madness finals.
    • Talked with https://www.pomtech.net/ and https://www.capzona.mx/ about sports/quant, more.
    • Little bit of sbsc dev.
  • Thursday

    • LOTS of private work.
  • Wednesday

    • Private work.
    • Japan won WBC.
    • DII final four: NSU, West Lib, San Bernardino, Black Hills.
    • Bard waitlisted (man, people are addicted to this stuff right now).
    • JPow quarter point hike.
    • Played with all calendly add-to-website options (embed, button, link) – they’re pretty great. But I don’t want to advertise on my site, already get way too many cold calls and reachouts from random people.
    • More gardening.
    • SBSC.
    • Parametrized the “iterate over” for each dimension. Basically games for all, but wanted to make it variable.
    • Also parametrized the aggregation capability. Only average is used right now, and only for line/margin.
  • Tuesday

    • Renewed ducati NY registration. $42.
    • Gmail notification bubble on iphone wouldn’t go away. 0 unread. Restarted app. Restarted phone. Nothing worked – until a new email came in and was read. That seemed to flush.
    • Anemones can live for hundreds of years.
    • Spoke with Rise Gardens again. I planted a few new lettuces/herbs today so they’ll be lush by the time the Good Morning America taping occurs. They’re also sending me a few 8-slot lids and will have me rearrange the desired plants to settle before bringing to the studio. If I deliver, they don’t have to fly out.
    • SBSC. Great progress on the dynamic looping of the statistical analysis/plotting based on dimension classes.
  • Monday

    • Private work.
    • Vernal equinox, first day of spring.
    • UBS bought Credit Suisse, 3.2B.
    • FRB still troubled, getting help from other banks.
    • Bought another tortilla press.
    • SBSC.
    • Updated the progression graph (plotly scatter) to follow the same format as the hbars of stats. Moved to plotting module.
    • Moved the calc_percentages functionality to the form_hbar plotter. It now assumes a multi-trace bar graph has complementary values.
    • Wrote some strategies for the dynamic looping of variables/adjacency/statistics. One using the “by” variable, the other using games directly. So far, the latter is superior.
  • Sunday

    • Private work.
    • Lots of gardening.
    • Remember that SQLA relationships define the join conditions and allow convenience ORM access to cols on instances (objects). As opposed to classes (tables), which is used for queries – only has and any exist for those.
    • Confirmed that alembic’s alter_column for a rename does NOT update the names of constraints (like FKs). Same is true for tables (the default prefix of FK constraints). It does update the references and functional pointers, just not the name.
    • SBSC.
    • Continued on the rewrite.
    • Graphql SQLA2 support coming out “soon” as of 5 days ago: https://github.com/graphql-python/graphene-sqlalchemy/pull/368
    • The Contestant change made it so line commit doesn’t have to do score commit. It’s just an initial insert of the 1 Game obj and 2 Contestant objs.
    • Added the functions to the Game class: get_contestants_by_prediction/location/coverage. Also added get_opponent to the Contestant class. All useful. Better located here with the models than elsewhere (queries module, results module, stats, wherever).
    • Removed all the get_la_matchups logic of the scores flow. Much simpler now that we can just look it up on the Game and Contestant objects.
    • Basically got the app working again after all the model changes.
    • Now just need to (1) dynamically loop the stats builders then (2) final optimization.