News

This is where I keep track of what I'm thinking and doing related to the Experimental Baseball League.

6:30 AM - Overnight cron jobs failed on Render. Starting there.

7:45 AM - Learning how to work with my prod db on Render without paying $7/month for pgAdmin.

8:30 AM - Starting new project, secret for now.

2:00 AM - I'm going to work on the roster change handling plan for 45 minutes starting now.

2:45 AM - Good progress. Next 45 minutes continue to focus on mlb roster change handling.

3:35 AM - Ran the mlb roster change handling plan. Taking a 10-minute break.

3:49 AM - Next 45-minutes to focus on simulating the 2025 season for testing.

4:35 AM - I didn't realize that simulating a season would be almost as complicated as implementing mlb roster change handling, but it is. I'm working on a plan, still. Going to end this session. Next time, continue working on 2025 simulation plan. I haven't done a push in a long time. I'm 33 commits ahead of main. I should definitely be on a branch.

7:07 AM - Settled in. I'm going to spend the next 40 minutes working on the 2025 season simulation.

8:40 AM - I forgot to start the timer. The simulation is finding bugs, and I'm fixing them. I'll keep going with that.

9:40 AM - Taking a break.

11:45 AM - Time for a prod deploy.

1:50 PM - We are live. People are logging in and submitting roster move requests. Well, not people, technically. A person.

8:20 PM - Going to work for 45 minutes on cleaning up little bugs and fixing documentation.

9:05 PM - Adding a count of active people on the website to the home page for large screens. 45 more minutes, then I'm done for the day.

9:30 PM - Prod deploy happening.

9:50 PM - That's a wrap.

2:40 AM - As I work, I have ideas for other projects that I want to do after the Experimental Baseball League is up and running. Currently, that's motivating me to work a lot. This morning, I'm going to focus on implementing the rest of the roster move process, then automated testing.

4:00 AM - I have a start on code to handle when players leave the MLB Phillies. Also, I have test data setup that creates user accounts for 8 team owners, and opens a browser window for each owner, logged in as that owner, navigated to their team page so that I can start testing roster moves. Next, I need to step through the 2025 season one week at a time write test scripts to handle MLB roster changes, EBL roster changes, open roster spots and the roster move request process.

5:15 AM - Progress on the handling of roster moves. Added a testing for roster move changes. Made a set of scripts to step through the 2025 season day by day and test roster moves and stats. Overhauled the league rules on roster moves.

3:30 PM - Automating testing.

4:40 PM - Reverted all automated testing. Starting over.

5:35 PM - Break for dinner.

6:00 PM - Back, time-boxing. Next 30 minutes focus on automated testing.

6:30 PM - next 30-minutes focus on simulating 2025.

7:00 PM - next 30-minutes overhaul how mlb transactions and roster moves are handled.

7:30 PM - transaction handling overhaul. next 30 minutes on strategy, planning and implementation.

8:00 PM - Generating local player ids was a choice with disadvantages. I'm changing to use the mlb player ids instead because they never change.

8:25 PM - Losing focus. Not going to push to prod now. Next time, start with the roster change handling plan.

4:05 AM - Obviously, this is not a sustainable rate of time spent on this project, but I'm awake and this is what I'm thinking about, so I might as well get to work.

4:35 AM - Did some security hardening and minor tweaks to my pycharm UI.

5:05 AM - Struggling to authenticate to an existing AWS account to set up email sending.

5:45 AM - Created new AWS account and spent 40 minutes trying to figure out how to add new CNAME records to robertbryan.net.

4:35 PM - I've decided that sending email can wait for later. Instead of that, I'm going to work on... what exactly? Not sure. I'll start with a review of the requirements and update my lists.

5:38 PM - I'm moving the rules and this lab notebook into the website.

6:25 PM - Fixing links in this lab notebook. Do I want to call it "Lab Notebook" or "News"? News is shorter, I'll go with that.

6:35 PM - Doing a prod deployment.

6:45 PM - Deprecated and removed the EBL content from robertbryan.net as it's all on the website now.

7:00 PM - Fixed a few little UI bugs, date display formats, added a case for players who don't have stats yet.

7:05 PM - Fixed a few places where I was still showing player positions instead of just "Pitcher" or "Hitter".

7:06 PM - Digging in to implementing user accounts, authentication, and logins. Almost everything on the site is going to be available to the public without login. You only need to log in to see the audit trail and to create roster move requests.

9:15 PM - Creating owner and admin accounts is done manually with scripts. Assigning owners to teams is a script. When a logged-in owner is on the team page for the team they own, they see a roster move button. That button opens a form where they choose the player they want to drop from their team and their first, second, and third choices for available players to add. Owners can see their pending roster move request after they submit it. If they don't like it or change their mind, they can just submit another roster move request and the old request is superseded. The new request is pending. Owners are limited to 50 roster move attempt requests per week.

9:30 PM - Doing a prod deploy. Going to db-init and populate stats for half of 2025, make an admin, check the UI, then call it a night.

4:30 PM - This is my list of things I want to work on in rough priority order:

6:05 PM - I'm satisfied with the web app overhaul to make it cell friendly.

6:35 PM - In the future, should remember to put the server in maintenance mode before pushing experimental code.

6:45 PM - I've migrated from SQLite to Postgres in prod. The new web layout looks good on cell.

7:02 PM - I've loaded all 2025 season stats into the prod db and run scoring. Everything looks good.

7:50 PM - I might need to install Postgres locally. That might actually be simpler than having two different database types.
I also need automated testing.

8:00 PM - Installing Postgres locally and ripping out all the SQLite code.

8:55 PM - I should not have gone down the path of trying to run SQLite locally and Postgres on Render.

9:25 PM - My local environment is stable and running on Postgres.

9:35 PM - Troubleshooting db-init.py on Render. I need to remember to time-box future development sessions.

9:45 PM - That's a wrap. The Render server is up and running with new data, looks good on cell.
my Render Postgres DB is on the Free level, need to remember to upgrade it if it continues to work.

6:35 AM - Designing screens before work.

4:35 PM - Resolved some security vulnerabilities in the website thanks to dependabot.
Drastically simplified the website by removing basically everything except the rules. Everything else was helpful in getting me started, but development is going so fast, just keeping the documentation up to date was slowing me down, so I cut it. I have a web front-end already, so I'm never going to serve the league as a set of texts lists, so we don't need stubs for where those lists will go.

5:45 PM - I have a scoring function that awards points on Sunday night according to the rules and appears to handle ties correctly.

6:45 PM - I have working prototypes of all the main screens.

7:00 PM - Made a few last changes. I'm ready to have people look at this and give feedback. Working on deployment.

7:30 PM - I am now a Render Professional license owner. The Experimental Baseball League, which was already operating at a loss because of my ChatGPT license, is now setting money on fire as quickly as we can light it up.

7:40 PM - We are live to the world. Our url is: https://ebl-k2sj.onrender.com/. Now updating the website and sending out comms.

8:05 PM - I sent out an email to the league.

5:30 AM - Fixed a few broken links and formatting annoyances on the website.
Now I'm going to start adding triggers to the database so that all changes get written to the audit log.

6:10 AM - AI is helpful. I'm using free ChatGPT. My db-make.py now creates triggers so that all table updates write to the audit trail. Later, I might want to split db-make into schema, triggers, and indexes.

6:40 AM - I'm working on a function to get player data from MLB and write it into my player table. There's a much better Python wrapper for the MLB API, but I need to upgrade Python to use it.

6:56 AM - That worked well. I'm on Python 3.14 now. Upgraded to the latest PyCharm while I was at it.
End session. Next time: Use the new Python wrapper to sync MLB API roster data to the players table.

3:50 PM - I upgraded to ChatGPT plus for $20/month. I use it for other things as well. It's been extremely helpful so far with coding, so the Experimental Baseball League is now operating at a loss.

4:00 PM - Oooooh, my ChatGPT plus subscription comes with Codex that integrates right into the IDE, but not PyCharm... or wait, does it? This is probably worth doing. This is what I'm doing now.

4:30 PM - Wow, okay. Wow. I have the Codex AI agent integrated into my IDE. It has read and understood the requirements I documented in the EBL Constitution. It understands how the MLB API works. It understands how my database works. It obviously knows how to do everything in Python.

4:45 PM - My role is different now. I don't have to figure out how to do the things I want to do. It knows how to do everything. I just need to be clear about telling it exactly what I want.

5:00 PM - This changes a lot. I don't have to limit myself to what I think I can code. That was holding me back. For now, I'm going to keep moving forward with my plan, but I'm open to the possibility that having AI available to do things for me means I can do different things, not just the same things faster.

5:20 PM - This is going to take some getting used to.

5:55 PM - More progress than I can summarize. I have a web front end now, running locally. Taking a break to walk and eat dinner.

7:05 PM - I have roster-sync so that it asks for a date and sets the player table to that date.

8:05 PM - I have a working web app that reads from my database and a prototype landing page.

8:55 PM - I have a prototype of a leaderboard. The drop-down lets you select any week. All the data is from the database. All the queries work. I'm using 2025 stats for testing. Players are assigned randomly to EBL teams for testing. Team names are randomly assigned from a list of fantasy baseball team names I found online.
That's it for today.

Worked on documentation. It's time to make a database.
All of my experience is with Postgres.
In this case, though, I'm going to start with SQLite because:

I'm going to invest an hour into installing and working with SQLite.
An hour? More like, what, 8 minutes?
That could not possibly have been any easier.
Cool. Okay, now what?
Database design, probably.

8:55 PM - Design is complete. I have a db-make.py that creates all the tables, fields, and foreign key constraints.
Do I want my db file in git? Oh, it's binary, probably not.

I'm removing the .pdf's I used to launch the league before moving to the web.
The website is okay for now.
The rules have stabilized enough.
I'm going to put an hour into trying to write a Python program to fetch the Phillies 40-man roster from the MLB API and write that data to a local text file that gets served to the website. No database. Wipe any previous results, replace with the new results.

That was easy.