Raw data files not included, they are large...
-
Generate initial spin files with commands such as:
-
java -jar sfinder.jar spin -p "*p7" -fb 0 -ft 4 -m 6 -c 1 -o "output/t_spc_spin_26.html" -t "v115@HhwwIexwHewwSeAgH"
-
Place them in
/spin_files/
. -
spin_to_fumen.js
is designed to read in the html files and grab relevant fumens (non mini TSS).
It should generate a large text fileoutput.txt
with a relevant spin fumen on each line.
When satisfied, rename the generatedoutput.txt
tostep_a.txt
-
parity_check.js
is designed to readstep_a.txt
and filter for good parity and output tostep_b.txt
. This may or may not be a useful filter. -
divider_check.js
is designed to readstep_b.txt
and filter out 0% 6L fields via divider rule, and output tostep_b_2.txt
. This is a useful filter. -
missing_pieces.js
is designed to readstep_b_2.txt
and process the missing pieces, outputting tostep_c.txt
.
This uses the fumens' queue comments, and assumes the setup queue was*p7
.
If thespin
command was run with a different setup queue, you may want to adjust this function. -
process_congruents.js
reads the fumens fromstep_c.txt
and processes congruents + mirror congruents, and outputs tostep_c_2.txt
This uses a dictionary to track for duplicates. If the file is too large, there may be memory issues. -
pc_chance_filter.js
reads fromstep_c_2.txt
and makes sfinder calls and filters for 98%+ fields.
This may take a long time to finish. If you want to do this in batches, you can adjust start position on line 51. -
Given that the previous step may have been split up into batches, I've included a quick
combine_d.js
script to combine results into a single file. EditfileList
. -
recursive_to_7p.js
reads fromreal_d_2.txt
(you may just name itd.txt
for simplicity); it generates all 7ps out of the list (which may include 4-7p setups). Outputs tostep_e.txt
. -
Apply filters again to these 7ps.
-
parity_filter_2.js
filters down the 7p list fromstep_e.txt
intostep_f.txt
. -
divider_filter_2.js
filters fromstep_f.txt
intostep_f_2.txt
. -
process_congruents_2.js
reads fromstep_f_2.txt
and processes congruents + mirror congruents intostep_f_3.txt
. -
pc_chance_filter_2.js
reads fromstep_f_3.txt
and makes sfinder calls and filters for 98%+ fields.
This may take a long time to finish. If you want to do this in batches, you can adjust start position on line 46. -
next steps: run score on final set.
tetris-fumen, readline, htmlparser2
I think