SearchLibrium 0.0.136__tar.gz → 0.0.138__tar.gz

{searchlibrium-0.0.136 → searchlibrium-0.0.138}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: SearchLibrium
-Version: 0.0.136
+Version: 0.0.138
 Summary: A Python package for econometric models driven by search
 Author: Mario Fernando Taco Morales, Alexander Paz, Prithvi Beeramole, Robert Burdett
 Author-email: Zeke Ahern <z.ahern@qut.edu.au>

{searchlibrium-0.0.136 → searchlibrium-0.0.138}/pyproject.toml

@@ -73,7 +73,7 @@ Documentation = "https://pypi.org/project/SearchLibrium/"
 realpython = "SearchLibrium.__main__:main"
 
 [tool.bumpver]
-current_version = "0.0.136"
+current_version = "0.0.138"
 version_pattern = "MAJOR.MINOR.PATCH"
 commit_message = "[skip ci] Bump version {old_version} -> {new_version}"
 commit = true

{searchlibrium-0.0.136 → searchlibrium-0.0.138}/src/SearchLibrium/MixedLogit.py

@@ -459,12 +459,13 @@ class MixedLogit(DiscreteChoiceModel):
         print(f"[MXL] Starting beta seed length={n_coeff}, first_values={betas[:min(8, len(betas))]!r}")
 
         if self.de_init:
-            print(f"[MXL] DE init enabled: popsize={self.de_popsize}, maxiter={self.de_maxiter}, tol={self.de_tol}, polish={self.de_polish}")
+            _v = getattr(self, 'verbose_de', False)
+            if _v: print(f"[MXL] DE init enabled: popsize={self.de_popsize}, maxiter={self.de_maxiter}, tol={self.de_tol}, polish={self.de_polish}")
             try:
                 before_de_obj = self.get_loglik_gradient(betas, self.X, self.y, self.panel_info,
                                                         draws, drawstrans, self.weights,
                                                         self.avail, self.batch_size)[0]
-                print(f"[MXL] DE before: obj={before_de_obj:.6g}")
+                if _v: print(f"[MXL] DE before: obj={before_de_obj:.6g}")
 
                 def _de_obj(x):
                     return self.get_loglik_gradient(x, self.X, self.y, self.panel_info,
@@ -495,23 +496,22 @@ class MixedLogit(DiscreteChoiceModel):
                     workers=1,
                 )
                 de_improved = False
-                print(f"[MXL] DE completed: success={de_result.success}, nit={de_result.nit}, message={de_result.message}")
+                if _v: print(f"[MXL] DE completed: success={de_result.success}, nit={de_result.nit}, message={de_result.message}")
                 if de_result.success:
                     de_obj = self.get_loglik_gradient(de_result.x, self.X, self.y, self.panel_info,
                                                      draws, drawstrans, self.weights,
                                                      self.avail, self.batch_size)[0]
-                    print(f"[MXL] DE after: obj={de_obj:.6g}")
+                    if _v: print(f"[MXL] DE after: obj={de_obj:.6g}")
                     de_improved = de_obj < before_de_obj
                     if de_improved:
                         betas = de_result.x
-                        print(f"[MXL] DE best seed first_values={betas[:min(8, len(betas))]!r}")
-                        print("[MXL] Differential evolution initialization improved the objective; starting minimise from DE seed.")
-                    else:
+                        if _v: print(f"[MXL] DE best seed first_values={betas[:min(8, len(betas))]!r}")
+                    elif _v:
                         print(f"[MXL] Differential evolution did not beat the original start ({de_obj:.6g} >= {before_de_obj:.6g}); keeping original start.")
-                if not de_improved:
+                if not de_improved and _v:
                     print(f"[MXL] Differential evolution initialization rejected because it did not improve the starting objective. Using original seed.")
             except Exception as e:
-                print(f"[MXL] Differential evolution initialization failed: {e}")
+                if _v: print(f"[MXL] Differential evolution initialization failed: {e}")
 
         # '''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
         if dev.using_gpu:  # {

{searchlibrium-0.0.136 → searchlibrium-0.0.138}/src/SearchLibrium/_choice_model.py

@@ -357,10 +357,33 @@ class DiscreteChoiceModel(ABC):
         # {
             if self.method == "bfgs":
                 self.stderr = np.sqrt(np.abs(np.diag(result.hess_inv)))
+                std_err_estimated = True
 
             if self.method == "l-bfgs-b":
                 hess = result['hess_inv'].todense()
                 self.stderr = np.sqrt(np.abs(np.diag(np.array(hess))))
+                std_err_estimated = True
+        # }
+
+        # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+        # Fallback: some models (e.g. MixedLogit with the default
+        # "slsqp" method) compute and stash a numerical Hessian inverse
+        # directly on the result dict (result['hess_inv']) instead of
+        # setting self.Hinv or using scipy's bfgs/l-bfgs-b machinery.
+        # The two checks above only recognise "bfgs"/"l-bfgs-b", so without
+        # this fallback self.stderr silently stays all-zero for every other
+        # method even though a usable Hessian was already computed.
+        # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+        if not std_err_estimated and 'hess_inv' in result and result['hess_inv'] is not None:
+        # {
+            hess_inv = result['hess_inv']
+            hess_inv = hess_inv.todense() if hasattr(hess_inv, 'todense') else hess_inv
+            diag_arr_tmp = np.diag(np.array(hess_inv))
+            pos_vals_idx = [ii for ii, el in enumerate(diag_arr_tmp) if el > 0]
+            diag_arr = np.zeros(len(diag_arr_tmp))
+            diag_arr[pos_vals_idx] = diag_arr_tmp[pos_vals_idx]
+            self.stderr = np.sqrt(np.abs(diag_arr))
+            std_err_estimated = not np.isnan(self.stderr).any()
         # }
 
         # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

{searchlibrium-0.0.136 → searchlibrium-0.0.138}/src/SearchLibrium/latent_class.py

@@ -42,7 +42,7 @@ class LatentClassMixedLogit:
         optimise_membership=True,
         membership_maxiter=50,
         l1_penalty=0.0,
-        l2_penalty=0.1,
+        l2_penalty=0.5,
     ):
         self.n_classes = int(n_classes)
         self.maxiter = int(maxiter)

{searchlibrium-0.0.136 → searchlibrium-0.0.138}/src/SearchLibrium/search.py

@@ -686,7 +686,7 @@ class Parameters:
 
         # ── Regularisation (primarily for latent class) ──────────────
         self.l1_penalty = kwargs.get('l1_penalty', 0.0)
-        self.l2_penalty = kwargs.get('l2_penalty', 0.1)
+        self.l2_penalty = kwargs.get('l2_penalty', 0.5)
 
         self.intercept_opts = intercept_opts
         self.base_alt = base_alt

searchlibrium-0.0.138/src/SearchLibrium/version.txt

@@ -0,0 +1 @@
+0.0.138

{searchlibrium-0.0.136 → searchlibrium-0.0.138}/src/SearchLibrium.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: SearchLibrium
-Version: 0.0.136
+Version: 0.0.138
 Summary: A Python package for econometric models driven by search
 Author: Mario Fernando Taco Morales, Alexander Paz, Prithvi Beeramole, Robert Burdett
 Author-email: Zeke Ahern <z.ahern@qut.edu.au>

searchlibrium-0.0.136/src/SearchLibrium/version.txt

@@ -1 +0,0 @@
-0.0.136