Allocating forecast to deeper granularity in SQL

You're given the following tables, called 'sales_forecast' and 'current_inventory'. The 'sales_forecast' table shows projected sales volume (in terms of # of units) by location by year. The 'current_inventory' shows the current # of units available by location and by color. The tables are shown below:

Table: sales_forecast

year	location	fcst_num_sales
2021	New York	113456
2022	New York	115675
2023	New York	117894
2024	New York	120113
2025	New York	122332
2021	Los Angeles	56728
2022	Los Angeles	57838
2023	Los Angeles	58947
2024	Los Angeles	60057
2025	Los Angeles	61166

Table: current_inventory

location	product_color	current_total_inventory
New York	Red	45382
New York	Blue	79419
New York	Green	11346
Los Angeles	Red	28364
Los Angeles	Blue	17018
Los Angeles	Green	34037

Using these tables, write a SQL query to take the current sales forecast and allocate it down to a projection of colors that will be sold at each location. For the simplicity of this example, you can use the current_inventory table to map the projected color proportions by each location. (e.g. if Blue represents XX% of New York's current inventory you can use that percent to allocate New York's total sales forecast down to the color Blue, and so on).

Click here to view these tables in an interactive SQL fiddle.