Set mean reversion target to D0(4)

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "FSRS-Optimizer"
-version = "5.0.2"
+version = "5.0.3"
 readme = "README.md"
 dependencies = [
     "matplotlib>=3.7.0",

src/fsrs_optimizer/fsrs_optimizer.py

@@ -95,9 +95,13 @@ def stability_short_term(self, state: Tensor, rating: Tensor) -> Tensor:
         new_s = state[:, 0] * torch.exp(self.w[17] * (rating - 3 + self.w[18]))
         return new_s
 
+    def init_d(self, rating: Tensor) -> Tensor:
+        new_d = self.w[4] - torch.exp(self.w[5] * (rating - 1)) + 1
+        return new_d
+
     def next_d(self, state: Tensor, rating: Tensor) -> Tensor:
         new_d = state[:, 1] - self.w[6] * (rating - 3)
-        new_d = self.mean_reversion(self.w[4], new_d)
+        new_d = self.mean_reversion(self.init_d(4), new_d)
         return new_d
 
     def step(self, X: Tensor, state: Tensor) -> Tensor:
@@ -113,7 +117,7 @@ def step(self, X: Tensor, state: Tensor) -> Tensor:
             # first learn, init memory states
             new_s = torch.ones_like(state[:, 0])
             new_s[index[0]] = self.w[index[1]]
-            new_d = self.w[4] - torch.exp(self.w[5] * (X[:, 1] - 1)) + 1
+            new_d = self.init_d(X[:, 1])
             new_d = new_d.clamp(1, 10)
         else:
             r = power_forgetting_curve(X[:, 0], state[:, 0])

src/fsrs_optimizer/fsrs_simulator.py

@@ -117,14 +117,16 @@ def stability_short_term(s, init_rating=None):
         return new_s
 
     def init_d(rating):
-        new_d = w[4] - np.exp(w[5] * (rating - 1)) + 1
+        return w[4] - np.exp(w[5] * (rating - 1)) + 1
+
+    def init_d_with_short_term(rating):
         rating_offset = np.choose(rating - 1, first_rating_offset)
-        new_d -= w[6] * rating_offset
+        new_d = init_d(rating) - w[6] * rating_offset
         return np.clip(new_d, 1, 10)
 
     def next_d(d, rating):
         new_d = d - w[6] * (rating - 3)
-        new_d = mean_reversion(w[4], new_d)
+        new_d = mean_reversion(init_d(4), new_d)
         return np.clip(new_d, 1, 10)
 
     def mean_reversion(init, current):
@@ -202,7 +204,7 @@ def mean_reversion(init, current):
             card_table[col["stability"]][true_learn],
             init_rating=card_table[col["rating"]][true_learn].astype(int),
         )
-        card_table[col["difficulty"]][true_learn] = init_d(
+        card_table[col["difficulty"]][true_learn] = init_d_with_short_term(
             card_table[col["rating"]][true_learn].astype(int)
         )