SQL - Multiple conditionals with row_number() - sql

Question

+------------+------------------+-------+-----------------+------------------------------+
| product_id | date | STOCK | REAL_STOCK-DIFF | Couting one time? |
+------------+------------------+-------+-----------------+------------------------------+
| 1ab7 | 10/18/2022 18:30 | 6009 | 495 | 495 |
| 1ab7 | 10/18/2022 20:10 | 6003 | 495 | 0 |
| 1ab7 | 10/20/2022 10:05 | 5514 | 495 | 0 |
| 1ab7 | 10/20/2022 11:05 | 23856 | 0 | 0 |
| 1ab7 | 10/20/2022 12:05 | 25850 | 0 | 0 |
| 1ab7 | 10/20/2022 13:05 | 44160 | 0 | 0 |
| 1ab7 | 10/20/2022 14:05 | 48205 | 130 | 130 |
| 1ab7 | 10/20/2022 17:05 | 48122 | 130 | 0 |
| 1ab7 | 10/20/2022 18:05 | 48075 | 130 | 0 |
| 1ab7 | 10/20/2022 19:05 | 17438 | 128 | 128 |
| 1ab7 | 10/21/2022 1:38 | 17310 | 128 | 0 |
| 2ab7 | 10/18/2022 18:30 | 85692 | 0 | 0 |
| 2ab7 | 10/20/2022 14:05 | 84498 | | SUM DIF STOCK == 495+130+128 |
| 2ab7 | 10/20/2022 15:05 | 84477 | | |
| 2ab7 | 10/20/2022 16:05 | 0 | | |
| 2ab7 | 10/20/2022 23:38 | 0 | | |
| 2ab7 | 10/21/2022 0:05 | 0 | | |
+------------+------------------+-------+-----------------+------------------------------+
This data shows the select that I tried to do with partition by in SQL, I'm controlling the stock and I have to show the stock difference over the products. That said, I was doing something like partition by max - min, however several conditions are applied, like: The stock can suddenly grow or decrease and even be removed completely (stock = 0), the partition by clause won't solve it purely.
My real stock is the third column "STOCK" in black, as you can see it was dropping from 6009 to the minimum 5514, then it grew up to 23856 until 48205, then was deducted to 17438.
Here we could do something like if 23856 > last number (5514) then new minimum and maximum = 23856, Just don't know how to partition it though. For the 17438, something like if the previous row_number > 17438*1.2 (20%) then new minimum = 17438
The SQL that I made was giving me the "Dif_Stock" column, which is wrongly displaying 42691 as the difference.
All that I'm trying to do is achieve the values that I inserted here in this column "Couting one time?":
"SUM DIF STOCK == 495+130+128"
My SQL code:
SELECT DISTINCT
product_id,
date,
stock,
MaxStock,
MinStock,
(MaxStock - MinStock) AS Dif_Stock
FROM (SELECT product_id,
date,
stock,
MAX(stock) OVER (PARTITION BY product_id) AS MaxStock,
MIN(stock) OVER (PARTITION BY product_id) AS MinStock,
ROW_NUMBER() OVER (PARTITION BY product_id ORDER BY product_id) AS ROWN
FROM (SELECT product_id,
category,
product_name,
vol,
price,
CAST(stock AS int) AS stock,
date
FROM stock_control
WHERE 1 = 1) STOCK
GROUP BY date,
product_id,
category,
product_name,
vol,
price,
stock,
date
--HAVING STOCK != 0
) STOCK_2
--ORDER BY (MaxStock - MinStock) DESC
ORDER BY product_id,
date ASC;
Real image bellow from select * from mytable