Almost all of my solutions were reworked afterwards and commented on.
>>> t="""
... --------Part 1-------- --------Part 2--------
... Day Time Rank Score Time Rank Score
... 22 03:28:06 3853 0 13:08:58 7743 0
... 21 05:24:22 10319 0 16:40:50 6784 0
... 20 01:01:34 1643 0 09:16:21 5792 0
... 19 00:52:15 3546 0 01:46:01 2163 0
... 18 01:54:04 5213 0 01:59:36 2527 0
... 17 00:58:49 1471 0 03:33:22 3419 0
... 16 00:42:46 2595 0 00:52:41 2434 0
... 15 00:06:18 1901 0 00:26:05 1668 0
... 14 00:14:18 2099 0 01:45:06 4343 0
... 13 00:32:53 2442 0 00:45:00 1890 0
... 12 00:51:16 3665 0 >24h 19646 0
... 11 00:23:25 2747 0 00:50:36 4371 0
... 10 00:33:42 1685 0 01:13:54 1047 0
... 9 00:14:11 2365 0 00:31:36 4758 0
... 8 00:16:04 4602 0 00:35:56 2308 0
... 7 00:51:45 5757 0 01:19:43 5484 0
... 6 00:11:37 3462 0 00:15:35 3055 0
... 5 00:28:16 2703 0 15:17:52 29752 0
... 4 00:10:02 3175 0 00:27:40 3877 0
... 3 00:26:14 2490 0 00:38:20 2263 0
... 2 00:12:47 2120 0 00:18:10 2323 0
... 1 01:03:15 13818 0 01:22:51 8780 0""".splitlines()
>>> {f"Mean {t[2].split()[2]}":sum(a:=[int((i:=l.split())[2])+int(i[5]) for l in t[3:]])/(n:=2*len(a)), "*":n}
{'Mean Rank': 4774.954545454545, '*': 44}Day 1 - Trebuchet Woke up too late. From now on: wake-up at 5:45.
Days 2 to 4 - Cube Conundrum/Gear Ratios/Scratchcards I knew I was slow. It's confirmed. These initial results call for modesty. My goal is still to get 50 stars with code I can understand in a year's time. When I trained this summer with the year 2022 challenges, I saw that things were getting tougher around day 15. So far, my aoc.Input tool hasn't proved its worth. We'll see.
Day 5 - If You Give A Seed A Fertilizer The problems started earlier this year. I had to come back to it at the end of the day to find the 2nd *.
Day 6 - Wait For It Elegant mathematical solution for 2nd * found afterwards.
Day 7 - Camel Cards Nothing difficult here. Depressing to have spent so much time when comparing hands of the same type was so simplified compared to poker rules. Important rework.
Day 8 - Haunted Wasteland The idea of least common multiple for 2nd * came to me instantly. Still 19 minutes to write it down, partly because I did not know the math.lcm() function, not even the LCM acronym (it is PPCM in french). Despite everything, I am happy with the progress in ranking from 4602 to 2308!
Day 9 - Mirage Maintenance A laborious prediction in the past, when all you had to do was reverse the list.
Days 10 & 11- Pipe Maze/Cosmic Expansion Idea was good, difficult to set up.
Day 12 - Hot Springs Enumerating and testing all possible arrangements was not too expansive for 1st *, and made easier with the help of itertools.combinations(). But I had to cheat and to look at a published solution to get the 2nd *. I'd never have thought of using a recursive function on my own.
Day 13 - Point Of Incidence Not quite awake this morning. More complicated to understand than to realize.
Day 14 - Parabolic Reflector Dish Spent a lot of time debugging. I had to insert what I thought was a calculation error to get the right answer from the test vector and finally the 2nd *!
Day 15 - Lens Library Second part takes some time to understand and solved a funny bug: I wanted to reuse the hash() function defined in first part. But that function was not visible in scope of part 2, so I was using the buily-in hash() function!
Day 16 - The Floor Will Be Lava This is the first time that the Structural Pattern Matching paradigm, which I try to use from time to time, has really come in handy. As a result, I've got a better grasp of the syntax. ex: case (_, _, '.') | (_, 0, '-') | (0, _, '|'):. Need a better algorithm actually since 3min for part 2 is too much.
Day 17 - Clumsy Crucible My solution was a complete depth-first graph walk with some heuristic and pruning with help of bisectmodule. Happy to have the 2* at last, but not satisfied with how long it took (6min and 15min), I throw a eye to winning solutions. And learn two main things: 1) the right algorithm is the Dijkstra's algorithm 2) the heapqmodule can help.
Day 18 - Lavaduct Lagoon After wandering for a long time over a two-dimensional map of the trench, trying to reuse the method of day10, I ended up remembering that we could very simply calculate the area delimited by a closed curve. It still took some time to include the half width of the digger in the calculation with complex solutions at the beginning until the final simplification. In the end the 2nd * came in 5 minutes.
Day 19 - Aplenty Wanted a breakfast before going for a walk into the workflow graph. Teleworking allow me to keep to finish.
Day 20 - Pulse Propagation No telework this day but a little time in the afternoon to craft a dirty specific code that gives the 2nd*. Submitted code is a bit more generic, but no idea how to prove it is the solution.
Day 21 - Step Counter It is time to go in vacation! First * was so easy, should not have taken 5h. 2nd much less. Dirty code.
Day 22 - Sand Slabs class make code easier to understand