-------------|
-- Cleanup --|
-------------|

drop table if exists input_ranges;
drop table if exists fresh;
drop table if exists ingredients;

-----------|
-- Setup --|
-----------|

create unlogged table input_ranges (
    low int8,
    high int8
);

-- Creating a single multirange sped up the part 1 query by a factor of 10!
create unlogged table fresh (
    range int8multirange
);

create unlogged table ingredients (
    id int8 primary key
);

-----------------|
-- Parse input --|
-----------------|

-- I don't understand these sed things
-- I just want the part before or after the empty line
-- Client side commands are terminatd with a newline, so no breaky for me
\copy input_ranges (low, high) from program'sed "/^$/Q" input/2025/5.txt' with delimiter '-'

-- Insert values as inclusive ranges (postgres normalizes them to be half-open)
-- because half-open ranges are objectively cooler.
insert into fresh (range)
select range_agg(int8range(range.low, range.high, '[]'))
from input_ranges as range;

-- We don't even need a temporary input table for the these
\copy ingredients (id) from program 'sed "1,/^$/d" input/2025/5.txt';

------------|
-- Part 1 --|
------------|

select count(ingredient.id) as part_1
from ingredients as ingredient
inner join fresh on fresh.range @> ingredient.id;

------------|
-- Part 2 --|
------------|

-- Even for this one the multirange is useful because it does
-- all the range merging for us!
with flat (range) as (
    select unnest(range) from fresh
)
select sum(upper(range) - lower(range)) as part_2
from flat;