Advent Of Code '2021

My adventofcode.com adventures.
Older ones: 2020, 2015.

Diary

day 01: Joined the races, with just 16 hrs of delay 🐌🐌🐌
day 02: So far, its all trivial. Joined by afternoon 🐌, actual time 23min.
day 03: Actually simple stuff, but it really took some focussing (101min).
day 04: A nice one. Some time was lost for misreading the input format.
day 05: Simple, but... man, I hate numbers - I just can't see them!.
day 06: Now, this was a good lesson - I lost at least 30mins implementing an illustration of example, rather than the actual logic of the process. 🤬
day 07: This seemingly trivial puzzle left me clueless - I just could not 👁👁 visualize the problem!
Only when I figured out it is like finding a balance point for a bar with suspended weights⚖️, everything became clear, finally.
day 08: Wasted too much time working on description details. Succeeded debugging of code #2 only after using simplified datasets.
day 09: Made a lot of mistakes. 🤡
day 10: A recipe for immediate failure: "Oh, I know how to do that!".
day 11: Simulations are fun! Some code from day09 was reusable today.
day 12: Seemingly simple 🤓.
day 13: That was fun!
day 14: Damn, I was barely able to steal some time for AOC today! The puzzle was fun, simple simulation resulted in 💥 explosion on part 2 and the correct solution was of less code than the initial part 1 😎!
day 15: This was a clever one - at first, I fell into trap of best-route-search 💀; the working solution was to just flood the whole field with cumulative risk values. I bet the algorithm can still be improved.
day 16: That was fun!
day 17:Mistakenly left off one character when copy-pasting input data and was about to nuke my own brains trying to figure out how to solve the part #2! 🤬
day 18: Lost insane amount of time because of stupid mistake: the puzzle hung up, so I was sure it is because of complexity, but the real reason was I had chosen a wrong way for data preparation, so the actual puzzle did not even start... doh.
day 19: This was horrible - I just could not visualize the situation. 🥶 Looking at my code I guess I probably solved it hard way.
day 25: After struggling through massive maelström 🌪 of Santa, Sauerkraut and other stuff, I finally made it back to AoC on Dec-26th!
Now, the part #1 was surprisingly simple, but to open the #2, I need to clear all the pending puzzles from days 17 and 19-24 first! 😳

Track record

My ranking placements and final times for puzzle 1 and 2 by days.
Others is number of solvers of both puzzles and Just 1 of those who had solved only the first one, taken from global stats when I completed.

day	Rank 1	Time 1	Rank 2	Time 2	Others	Just 1
1	87733	16:09:30	77832	16:26:49	127445	14909
2	58863	08:21:25	55421	08:26:51	106584	4125
3	8808	00:25:36	10369	01:41:35	45825	23066
4	4087	00:41:37	4896	01:06:05	6191	1630
5	8539	01:18:17	7974	01:44:41	8093	1972
6	8567	00:30:44	7349	01:06:45	7484	4379
7	16732	01:54:24	15841	02:05:48	15930	1773
8	9645	00:41:25	17577	07:04:10	15930	1773
9	9219	00:51:40	8295	01:55:16	8308	5105
10	9915	01:12:20	10966	02:00:54	10982	1716
11	6808	01:22:26	6819	01:30:05	6882	331
12	8918	02:55:40	8762	03:45:21	8771	1601
13	8031	01:28:30	4321	01:35:31	7324	1034
14	11437	02:07:05	16267	08:17:38	18245	9021
15	11538	04:58:08	9372	06:00:22	9388	3618
16	6503	03:47:12	6251	04:41:06	6270	1290
17	13740	07:57:44	25251	>24h
18	8060	11:51:28	11314	17:57:49	11326	281
19	15318	>24h	15075	> 24h
25	10309	>24h			6259	4060

Code

The project codebase consists of daily puzzle files named like day01.js, universal runner code and module tests for runner code.

Performance report

The following report was generated using command ./run abm.

day	Main1	Main2	Demo1	Demo2	M1_µs	M2_µs	D1_µs	D2_µs
01	1759	1805	7	5	92	252	76	51
02	2073315	1840311528	150	900	1074	408	68	81
03	3309596	2981085	198	230	698	2872	120	962
04	82440	20774	4512	1924	318	51360	172	61
05	98486	163135	5	12	90402	62993	239	81
06	362346	1639643057051	5934	26984457539	19083	263	3021	348
07	345035	97038163	37	168	3100	4398	125	87
08	344	1048410	26	61229	106	457651	140	36204
09	417	1148965	15	1134	6902	14368	248	5247
10	341823	2801302861	26397	288957	4142	18103	222	576
11	1719	232	1656	195	1245	629	4499	558
12	4167	98441	10	36	9184	216149	357	277
13	827	EAHKRECP	17	O	2800	7831	455	1658
14	2988	3572761917024	1588	2188189693529	306	2730	374	328
15	613	2899	40	315	30914	1553300	1044	10114
16	923	258888628940	6,9,14,16,12,23,31	3,54,7,9,1,0,0,1	588	475	630	95
17	9180	3767	45	112	10698	7593	336	6362
18	4469	4770	4140	3993	15882	393053	8026	9140
19	332	8507	79	3621	2230650	215	141873	224
25	568		58		192101		2636

Usage

After installing with npm or yarn, just type ./run and the puzzles from the most recent puzzle file will be executed.

The following command line parameters apply:

integer - day number(s), (default: most recent day only);
a: all days;
b: both datasets (default: main data only);
d: example data only (mutually exclusive with 'b' option;
h: print help information and terminate;
j: generate output as JSON-formatted rows;
m: generate markdown output (default: text table for multiple, JSON for single puzzle).

If neither 'b' nor 'd' is present, only the main dataset is used. If 'd' is present, and demo data is not defined, it falls back to main dataset.

Coding

To start with a new puzzle for, say, day 14, do:

copy the contents of template.js to day14.js;
create day14.txt in ./data directory and populate it with data from the website;
have fun!

It is recommended to copy the puzzle data into a text file in data folder.

Debugging

To trap failed assertions when debugging, set a breakpoint to runner/index.js line #20.

valango / adventofcode_2021 Goto Github PK