⬆️ Upgrade build software and dependencies (#768)

* ⬆️ Upgrade build software and dependencies

- Upgrade anaconda to 2025.12
- Upgrade quantecon-book-theme to 0.15.1
- Add jupyter-book version constraint (<2.0)
- Configure Dependabot for conda ecosystem with jupyter-book restriction
- Temporarily disable build cache for full execution check

* Fix YAML syntax error in ci.yml

* FIX: Use p_solution variable in exercise 2 to avoid dimension mismatch

The variable p was being reused from the high-dimensional problem section
(3000-dim vector) when the raises-exception cell failed, causing a shape
mismatch with the 3x3 matrix A defined in the exercise.

* ⬆️ Configure dependabot to ignore Python version updates

Python version should be constrained by the anaconda distribution version
specified in environment.yml, not updated independently by dependabot.

* Fix dependabot python ignore rule to include versions constraint

* FIX: Compute p_solution explicitly with converging initial value

The raises-exception cell stops at the first failure, so p_solution
may not be assigned. Now we explicitly compute p_solution using
initLs[1] which is known to converge.

Author: mmcky

.github/dependabot.yml

@@ -11,3 +11,16 @@ updates:
       prefix: ⬆️
     schedule:
       interval: weekly
+
+  - package-ecosystem: "conda"
+    directory: "/"
+    commit-message:
+      prefix: ⬆️
+    schedule:
+      interval: weekly
+    ignore:
+      - dependency-name: "jupyter-book"
+        versions: [">=2.0"]
+      - dependency-name: "python"
+        # Python version should be constrained by the anaconda distribution version
+        versions: [">0"]

.github/workflows/ci.yml

@@ -43,13 +43,13 @@ jobs:
       - name: Display Pip Versions
         shell: bash -l {0}
         run: pip list
-      - name: Download "build" folder (cache)
-        uses: dawidd6/action-download-artifact@v11
-        with:
-          workflow: cache.yml
-          branch: main
-          name: build-cache
-          path: _build
+      # - name: Download "build" folder (cache)
+      #   uses: dawidd6/action-download-artifact@v11
+      #   with:
+      #     workflow: cache.yml
+      #     branch: main
+      #     name: build-cache
+      #     path: _build
       # Build Assets (Download Notebooks and PDF via LaTeX)
       - name: Build Download Notebooks (sphinx-tojupyter)
         shell: bash -l {0}

environment.yml

@@ -3,11 +3,11 @@ channels:
   - default
 dependencies:
   - python=3.13
-  - anaconda=2025.06
+  - anaconda=2025.12
   - pip
   - pip:
-    - jupyter-book==1.0.4post1
-    - quantecon-book-theme==0.15.0
+    - jupyter-book>=1.0.4post1,<2.0
+    - quantecon-book-theme==0.15.1
     - sphinx-tojupyter==0.4.0
     - sphinxext-rediraffe==0.2.7
     - sphinx-exercise==1.2.0

lectures/newton_method.md

@@ -1074,22 +1074,21 @@ Let's run through each initial guess and check the output
 ```{code-cell} ipython3
 :tags: [raises-exception]
 
-attempt = 1
-for init in initLs:
+for attempt, init in enumerate(initLs, start=1):
     print(f"Attempt {attempt}: Starting value is {init} \n")
     %time p = newton(lambda p: e(p, A, b, c), init, tol=1e-15, max_iter=15)
     print("-" * 64)
-    attempt += 1
 ```
 
 We can see that Newton's method may fail for some starting values.
 
 Sometimes it may take a few initial guesses to achieve convergence.
 
-Substitute the result back to the formula to check our result
+Substitute the result back to the formula to check our result using the second initial guess which converges
 
 ```{code-cell} ipython3
-e(p, A, b, c)
+p_solution = newton(lambda p: e(p, A, b, c), initLs[1], tol=1e-15, max_iter=15)
+e(p_solution, A, b, c)
 ```
 
 We can see the result is very accurate.