.github/workflows/ruff.yml

@@ -0,0 +1,8 @@
+name: Ruff
+on: [ push, pull_request ]
+jobs:
+  ruff:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: chartboost/ruff-action@v1

.github/workflows/test.yml

@@ -0,0 +1,27 @@
+name: Test Python
+
+on: [push, pull_request]
+
+jobs:
+    test:
+
+        runs-on: ubuntu-latest
+
+        steps:
+        - uses: actions/checkout@v4
+        - name: Set up Python
+          uses: actions/setup-python@v3
+          with:
+            python-version: '3.12'
+
+        - name: Install uv
+          uses: astral-sh/setup-uv@v5
+          with:
+            enable-cache: true
+
+        - name: Install the project
+          run: uv sync --all-extras --dev
+
+        - name: Run tests
+          # For example, using `pytest`
+          run: uv run pytest tests

README.md

@@ -1,5 +1,7 @@
 # FSRS Optimizer
 
+[![PyPi](https://img.shields.io/pypi/v/FSRS-Optimizer)](https://pypi.org/project/FSRS-Optimizer/) [![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
+
 The FSRS Optimizer is a Python library capable of utilizing personal spaced repetition review logs to refine the FSRS algorithm. Designed with the intent of delivering a standardized, universal optimizer to various FSRS implementations across numerous programming languages, this tool is set to establish a ubiquitous standard for spaced repetition review logs. By facilitating the uniformity of learning data among different spaced repetition softwares, it guarantees learners consistent review schedules across a multitude of platforms.
 
 Delve into the underlying principles of the FSRS Optimizer's training process at: https://github.com/open-spaced-repetition/fsrs4anki/wiki/The-mechanism-of-optimization
@@ -62,3 +64,9 @@ python -m fsrs_optimizer "revlog.csv"
 ![image](https://github.com/open-spaced-repetition/fsrs-optimizer/assets/32575846/fad7154a-9667-4eea-b868-d94c94a50912)
 
 ![image](https://github.com/open-spaced-repetition/fsrs-optimizer/assets/32575846/f868aac4-2e9e-4101-b8ad-eccc1d9b1bd5)
+
+---
+
+## Alternative
+
+Are you getting tired of installing torch? Try [fsrs-rs-python](https://github.com/open-spaced-repetition/fsrs-rs-python)!

pyproject.toml

@@ -4,19 +4,29 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "FSRS-Optimizer"
-version = "4.25.2"
+version = "5.7.2"
 readme = "README.md"
 dependencies = [
     "matplotlib>=3.7.0",
     "numpy>=1.22.4",
     "pandas>=1.5.3",
     "pytz>=2022.7.1",
-    "scikit_learn>=1.2.2",
+    "scikit_learn>=1.4.0",
     "torch>=1.13.1",
     "tqdm>=4.64.1",
     "statsmodels>=0.13.5",
+    "scipy<1.14.1"
 ]
-requires-python = ">=3.9"
+requires-python = ">=3.9,<3.13"
 
 [project.urls]
 Homepage = "https://github.com/open-spaced-repetition/fsrs-optimizer"
+[tool.ruff.lint]
+ignore = ["F405", "F403", "E712", "F541", "E722", "E741"]
+[project.optional-dependencies]
+test = [
+    "ruff",
+    "mypy",
+    "pytest",
+    "pytest-cov",
+]

src/fsrs_optimizer/__main__.py

@@ -40,13 +40,16 @@ def process(filepath, filter_out_flags: list[int]):
             "revlog_start_date": "2006-10-05",
             "preview": "y",
             "filter_out_suspended_cards": "n",
+            "enable_short_term": "y",
         }
 
     # Prompts the user with the key and then falls back on the last answer given.
     def remembered_fallback_prompt(key: str, pretty: str = None):
         if pretty is None:
             pretty = key
-        remembered_fallbacks[key] = prompt(f"input {pretty}", remembered_fallbacks[key])
+        remembered_fallbacks[key] = prompt(
+            f"input {pretty}", remembered_fallbacks.get(key, None)
+        )
 
     print("The defaults will switch to whatever you entered last.\n")
 
@@ -60,11 +63,15 @@ def remembered_fallback_prompt(key: str, pretty: str = None):
 
         remembered_fallback_prompt("next_day", "used next day start hour")
         remembered_fallback_prompt(
-            "revlog_start_date", "the date at which before reviews will be ignored"
+            "revlog_start_date",
+            "the date at which before reviews will be ignored | YYYY-MM-DD",
         )
         remembered_fallback_prompt(
             "filter_out_suspended_cards", "filter out suspended cards? (y/n)"
         )
+        remembered_fallback_prompt(
+            "enable_short_term", "enable short-term component in FSRS model? (y/n)"
+        )
 
         graphs_input = prompt("Save graphs? (y/n)", remembered_fallbacks["preview"])
     else:
@@ -81,8 +88,9 @@ def remembered_fallback_prompt(key: str, pretty: str = None):
         json.dump(remembered_fallbacks, f)
 
     save_graphs = graphs_input != "n"
+    enable_short_term = remembered_fallbacks["enable_short_term"] == "y"
 
-    optimizer = fsrs_optimizer.Optimizer()
+    optimizer = fsrs_optimizer.Optimizer(enable_short_term=enable_short_term)
     if filepath.endswith(".apkg") or filepath.endswith(".colpkg"):
         optimizer.anki_extract(
             f"{filepath}",
@@ -107,18 +115,23 @@ def remembered_fallback_prompt(key: str, pretty: str = None):
     for i, f in enumerate(figures):
         f.savefig(f"pretrain_{i}.png")
         plt.close(f)
-    figures = optimizer.train(verbose=save_graphs)
+    figures = optimizer.train(verbose=save_graphs, recency_weight=True)
     for i, f in enumerate(figures):
         f.savefig(f"train_{i}.png")
         plt.close(f)
 
     optimizer.predict_memory_states()
-    figures = optimizer.find_optimal_retention(verbose=save_graphs)
-    for i, f in enumerate(figures):
-        f.savefig(f"find_optimal_retention_{i}.png")
-        plt.close(f)
+    try:
+        figures = optimizer.find_optimal_retention(verbose=save_graphs)
+        for i, f in enumerate(figures):
+            f.savefig(f"find_optimal_retention_{i}.png")
+            plt.close(f)
+    except Exception as e:
+        print(e)
+        print("Failed to find optimal retention")
+        optimizer.optimal_retention = 0.9
 
-    optimizer.preview(optimizer.optimal_retention)
+    print(optimizer.preview(optimizer.optimal_retention))
 
     profile = f"""{{
     // Generated, Optimized anki deck settings
@@ -139,7 +152,13 @@ def remembered_fallback_prompt(key: str, pretty: str = None):
     loss_before, loss_after = optimizer.evaluate()
     print(f"Loss before training: {loss_before:.4f}")
     print(f"Loss after training: {loss_after:.4f}")
-    metrics, figures = optimizer.calibration_graph()
+    metrics, figures = optimizer.calibration_graph(verbose=False)
+    for partition in metrics:
+        print(f"Last rating = {partition}")
+        for metric in metrics[partition]:
+            print(f"{metric}: {metrics[partition][metric]:.4f}")
+        print()
+
     metrics["Log loss"] = loss_after
     if save_graphs:
         for i, f in enumerate(figures):