Feat/long-term schduler

[ishiko732]

@L-M-Sherlock I am uncertain whether the state transition from New to Review is correct; I need you to verify the LongTermScheduler.newState and LongTermScheduler.reviewState methods.

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 100.00%. Comparing base (ba57899) to head (3294c28). Report is 2 commits behind head on main.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #100 +/- ## ========================================== Coverage 100.00% 100.00% ========================================== Files 9 10 +1 Lines 439 557 +118 Branches 50 58 +8 ========================================== + Hits 439 557 +118 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[L-M-Sherlock]

How can I check them quickly? Do you have any unit tests?

[ishiko732]

Do you have any unit tests?

https://github.com/open-spaced-repetition/ts-fsrs/blob/11429fff6568dc49ed2d549f2b751bfdecc7c5c9/__tests__/impl/long-term_schduler.test.ts

  const parameters = {
    request_retention: 0.9,
    maximum_interval: 36500,
    w: [
      0.4197, 1.1869, 3.0412, 15.2441, 7.1434, 0.6477, 1.0007, 0.0674, 1.6597,
      0.1712, 1.1178, 2.0225, 0.0904, 0.3025, 2.1214, 0.2498, 2.9466, 0.4891,
      0.6468,
    ],
    enable_fuzz: false,
    enable_short_term: false,
  }

first_time: 2022/11/29 12:30:00

test1

Grades:

[Good, Good, Good, Good, Good, Good, Again, Again, Good, Good, Good, Good, Good]

ivl_history:

[3, 6, 17, 42, 95, 200, 8, 2, 3, 5, 8, 14, 23]

s_history:

[
      0.57587467, 6.28341418, 16.83356103, 41.95128557, 95.07063986,
      199.53765138, 8.31519008, 1.96276113, 3.06877302, 4.90880017, 8.15177579,
      13.50873393, 22.92901865,
    ]

d_history:

[
      7.1434, 7.1434, 7.1434, 7.1434, 7.1434, 7.1434, 9.00990564, 10,
      9.80746516, 9.62790717, 9.46045139, 9.30428213, 9.15863867,
    ]

test2

Grades:

[Again, Hard, Good, Easy, Again, Hard, Good, Easy]

ivl_history:

[1, 2, 3, 8, 2, 2, 4, 10]

s_history:

[
      0.21652154, 0.51780862, 1.8183783, 8.18593986, 1.96087115, 2.23717242,
      3.33185406, 10.31008873,
    ]

d_history:

[
      9.00990564, 9.81735598, 9.63713135, 8.53580104, 10, 10, 9.80746516,
      8.69465435,
    ]

test3

Grades:

[Hard, Good, Easy, Again, Hard, Good, Easy, Again]

ivl_history:

[2, 3, 17, 3, 4, 6, 18, 3]

s_history:

[
      0.35311368, 3.40179546, 16.86596974, 2.9223039, 3.73601023, 6.27595217,
      18.057595, 2.96541083,
    ]

d_history:

[
      8.07665282, 8.01375158, 7.02183706, 8.89653604, 9.71162749, 9.53852896,
      8.44384445, 10,
    ]

test4

Grades:

[Good, Easy, Again, Hard, Good, Easy, Again, Hard]

ivl_history:

[3, 17, 3, 4, 8, 30, 4, 4]

s_history:

[
      0.57587467, 17.3937106, 2.98322161, 4.08808234, 7.99405969, 29.76078201,
      3.78204811, 4.44627015,
    ]

d_history:

[
      7.1434, 6.21014718, 8.13955406, 9.0056661, 8.88014936, 7.82983963,
      9.65007924, 10,
    ]

test5

Grades:

[Easy, Again, Hard, Good, Easy, Again, Hard, Good]

ivl_history:

[4, 1, 2, 3, 13, 3, 3, 5]

s_history:

[
      0.93916394, 0.70772253, 1.16119435, 3.42301186, 12.97577367, 2.53486926,
      3.06202367, 4.60523893,
    ]

d_history:
[
      6.21014718, 8.13955406, 9.0056661, 8.88014936, 7.82983963, 9.65007924, 10,
      9.80746516,
    ]

[ishiko732]

How can I check them quickly?

I will write a preview page when I have time.

[L-M-Sherlock]

test1

Grades:

[Good, Good, Good, Good, Good, Good, Again, Again, Good, Good, Good, Good, Good]

ivl_history:

[3, 6, 17, 42, 95, 200, 8, 2, 3, 3, 6, 10, 17]

s_history:

[
      0.57587467, 6.28341418, 16.83356103, 41.95128557, 95.07063986,
      199.53765138, 8.31519008, 1.59859456, 1.59859456, 3.49350932, 5.58821385,
      9.83641092, 16.95422162,
    ]

d_history:
[
      7.1434, 7.1434, 7.1434, 7.1434, 7.1434, 7.1434, 9.00990564, 10,
      9.80746516, 9.62790717, 9.46045139, 9.30428213, 9.15863867,
    ]

Why the first Good after the last Again doesn't change the stability?

[ishiko732]

Why the first Good after the last Again doesn't change the stability?

I don’t understand why all the results from the next_recall_stability calculation are exactly the same.

[ishiko732]

Why the first Good after the last Again doesn't change the stability?

https://github.com/open-spaced-repetition/ts-fsrs/blob/f838a179f9371e8acf8703a028aa04995f5cdfa3/src/fsrs/impl/long_term_schduler.ts#L213-L214 scheduled it by minutes here, which was a mistake. I’ll make the adjustments tonight. 😢

[ishiko732]

switch long-term schduler test

Grades:

[Good(short), Good(long), Again(long), Good(long), Good(short), Again(short)]

ivl_history:

[0, 4, 1, 4, 12, 0]

s_history:

[
      3.0412, 3.0412, 1.20788692, 3.83856852, 12.23542321, 2.48288917,
    ]

d_history:

[4.49094334, 4.66971892, 6.70295066, 6.73263695,6.76032238,8.65264745]

state_history:

['Learning', 'Review', 'Review', 'Review','Review','Relearning']

[L-M-Sherlock]

Hi, @remnoteio, @sobjornstad, @bjsi. This PR will allow user to skip the short-term schedule. I guess you will be interested in it.