Enable save_cmdstan_config

cmdstanpy/cmdstan_args.py

@@ -866,6 +866,7 @@ def compose_command(
                 cmd.append(f'init={self.inits[idx]}')
         cmd.append('output')
         cmd.append(f'file={csv_file}')
+        cmd.append('save_cmdstan_config=1')
         if diagnostic_file:
             cmd.append(f'diagnostic_file={diagnostic_file}')
         if profile_file:

cmdstanpy/stanfit/runset.py

@@ -56,20 +56,29 @@ def __init__(
         )
         self._stdout_files, self._profile_files = [], []
         self._csv_files, self._diagnostic_files = [], []
+        self._config_files = []
 
         # per-process output files
         if one_process_per_chain and chains > 1:
             self._stdout_files = [
                 self.gen_file_name(".txt", extra="stdout", id=id)
                 for id in self._chain_ids
             ]
+            self._config_files = [
+                os.path.join(
+                    self._outdir, f"{self._base_outfile}_{id}_config.json"
+                )
+                for id in self._chain_ids
+            ]
+
             if args.save_profile:
                 self._profile_files = [
                     self.gen_file_name(".csv", extra="profile", id=id)
                     for id in self._chain_ids
                 ]
         else:
             self._stdout_files = [self.gen_file_name(".txt", extra="stdout")]
+            self._config_files = [self.gen_file_name(".json", extra="config")]
             if args.save_profile:
                 self._profile_files = [
                     self.gen_file_name(".csv", extra="profile")
@@ -93,25 +102,21 @@ def __init__(
                 ]
 
     def __repr__(self) -> str:
-        repr = 'RunSet: chains={}, chain_ids={}, num_processes={}'.format(
-            self._chains, self._chain_ids, self._num_procs
-        )
-        repr = '{}\n cmd (chain 1):\n\t{}'.format(repr, self.cmd(0))
-        repr = '{}\n retcodes={}'.format(repr, self._retcodes)
-        repr = f'{repr}\n per-chain output files (showing chain 1 only):'
-        repr = '{}\n csv_file:\n\t{}'.format(repr, self._csv_files[0])
+        lines = [
+            f"RunSet: chains={self._chains}, chain_ids={self._chain_ids}, "
+            f"num_processes={self._num_procs}",
+            f" cmd (chain 1):\n\t{self.cmd(0)}",
+            f" retcodes={self._retcodes}",
+            " per-chain output files (showing chain 1 only):",
+            f" csv_file:\n\t{self._csv_files[0] if self._csv_files else ''}",
+        ]
         if self._args.save_latent_dynamics:
-            repr = '{}\n diagnostics_file:\n\t{}'.format(
-                repr, self._diagnostic_files[0]
-            )
+            lines.append(f" diagnostics_file:\n\t{self._diagnostic_files[0]}")
         if self._args.save_profile:
-            repr = '{}\n profile_file:\n\t{}'.format(
-                repr, self._profile_files[0]
-            )
-        repr = '{}\n console_msgs (if any):\n\t{}'.format(
-            repr, self._stdout_files[0]
-        )
-        return repr
+            lines.append(f" profile_file:\n\t{self._profile_files[0]}")
+        lines.append(f" console_msgs (if any):\n\t{self._stdout_files[0]}")
+        lines.append(f" config_files:\n\t{self._config_files[0]}")
+        return '\n'.join(lines)
 
     @property
     def model(self) -> str:
@@ -196,6 +201,13 @@ def stdout_files(self) -> list[str]:
         """
         return self._stdout_files
 
+    @property
+    def config_files(self) -> list[str]:
+        """
+        List of paths to CmdStan config json files.
+        """
+        return self._config_files
+
     def _check_retcodes(self) -> bool:
         """Returns ``True`` when all chains have retcode 0."""
         return all(retcode == 0 for retcode in self._retcodes)

test/test_cmdstan_args.py

@@ -808,3 +808,15 @@ def test_args_pathfinder_bad(arg: str, require_int: bool) -> None:
         args = PathfinderArgs(**{arg: 1.1})  # type: ignore
         with pytest.raises(ValueError):
             args.validate()
+
+
+def test_save_cmdstan_config() -> None:
+    sampler_args = SamplerArgs()
+    cmdstan_args = CmdStanArgs(
+        model_name='bernoulli',
+        model_exe='',
+        chain_ids=[1, 2, 3, 4],
+        method_args=sampler_args,
+    )
+    command = cmdstan_args.compose_command(0, csv_file="foo")
+    assert "save_cmdstan_config=1" in command

test/test_runset.py

@@ -30,6 +30,7 @@ def test_check_repr() -> None:
     assert 'csv_file' in repr(runset)
     assert 'console_msgs' in repr(runset)
     assert 'diagnostics_file' not in repr(runset)
+    assert 'config_file' in repr(runset)
 
 
 def test_check_retcodes() -> None:
@@ -106,6 +107,11 @@ def test_output_filenames_one_proc_per_chain() -> None:
         stdout_file.endswith(f"_stdout_{id}.txt")
         for id, stdout_file in zip(chain_ids, runset.stdout_files)
     )
+    assert len(runset.config_files) == len(chain_ids)
+    assert all(
+        config_file.endswith(f"_{id}_config.json")
+        for id, config_file in zip(chain_ids, runset.config_files)
+    )
 
     cmdstan_args_other_files = CmdStanArgs(
         model_name='bernoulli',
@@ -153,6 +159,8 @@ def test_output_filenames_threading() -> None:
     )
     assert len(runset.stdout_files) == 1
     assert runset.stdout_files[0].endswith("_stdout.txt")
+    assert len(runset.config_files) == 1
+    assert runset.config_files[0].endswith("_config.json")
 
     cmdstan_args_other_files = CmdStanArgs(
         model_name='bernoulli',
@@ -198,6 +206,7 @@ def test_output_filenames_single_chain() -> None:
     runset = RunSet(args=cmdstan_args, chains=1, one_process_per_chain=True)
     base_file = runset._base_outfile
     assert runset.stdout_files[0].endswith(f"{base_file}_stdout.txt")
+    assert runset.config_files[0].endswith(f"{base_file}_config.json")
 
     cmdstan_args_other_files = CmdStanArgs(
         model_name='bernoulli',

test/test_sample.py

@@ -2204,3 +2204,27 @@ def test_no_output_draws() -> None:
     mcmc = model.sample(data=data, iter_sampling=0, save_warmup=False, chains=2)
     draws = mcmc.draws()
     assert np.array_equal(draws, np.empty((0, 2, len(mcmc.column_names))))
+
+
+def test_config_output() -> None:
+    stan = os.path.join(DATAFILES_PATH, 'bernoulli.stan')
+    jdata = os.path.join(DATAFILES_PATH, 'bernoulli.data.json')
+    model = CmdStanModel(stan_file=stan)
+    fit = model.sample(
+        data=jdata,
+        chains=2,
+        seed=12345,
+        iter_warmup=100,
+        iter_sampling=200,
+    )
+    assert all(os.path.exists(cf) for cf in fit.runset.config_files)
+
+    # Config file naming differs when only a single chain is output
+    fit_one_chain = model.sample(
+        data=jdata,
+        chains=1,
+        seed=12345,
+        iter_warmup=100,
+        iter_sampling=200,
+    )
+    assert all(os.path.exists(cf) for cf in fit_one_chain.runset.config_files)