Feature / refresh History dtype upon encountering a new field

.github/workflows/basic.yml

@@ -98,7 +98,7 @@ jobs:
             pip install -r install/testing_requirements.txt
             pip install -r install/misc_feature_requirements.txt
 
-            git clone --recurse-submodules -b refactor/pounders_API https://github.com/POptUS/IBCDFO.git
+            git clone --recurse-submodules -b develop https://github.com/POptUS/IBCDFO.git
             pushd IBCDFO/minq/py/minq5/
             export PYTHONPATH="$PYTHONPATH:$(pwd)"
             echo "PYTHONPATH=$PYTHONPATH" >> $GITHUB_ENV

.github/workflows/extra.yml

@@ -166,7 +166,7 @@ jobs:
             sed -i -e "s/pyzmq>=22.1.0,<23.0.0/pyzmq>=23.0.0,<24.0.0/" ./balsam/setup.cfg
             cd balsam; pip install -e .; cd ..
 
-            git clone --recurse-submodules -b refactor/pounders_API https://github.com/POptUS/IBCDFO.git
+            git clone --recurse-submodules -b develop https://github.com/POptUS/IBCDFO.git
             pushd IBCDFO/minq/py/minq5/
             export PYTHONPATH="$PYTHONPATH:$(pwd)"
             echo "PYTHONPATH=$PYTHONPATH" >> $GITHUB_ENV

libensemble/ensemble.py

@@ -480,7 +480,7 @@
                 if old_spec.get("in") and old_spec.get("inputs"):
                     old_spec.pop("inputs")  # avoid clashes
                 elif old_spec.get("out") and old_spec.get("outputs"):
-                    old_spec.pop("inputs")  # avoid clashes
+                    old_spec.pop("outputs")  # avoid clashes
                 setattr(self, f, ClassType(**old_spec))
             else:  # None. attribute not set yet
                 setattr(self, f, ClassType(**loaded_spec))

libensemble/history.py

@@ -62,6 +62,7 @@ def __init__(
             # - then convert that type into a python type in the best way known so far...
             # - we need to make sure the size of string types is preserved
             # - if sub-array shape, save as 3-tuple
+
             H0_fields = []
             for i in range(len(H0.dtype.names)):
                 dtype = H0.dtype[i]
@@ -110,6 +111,7 @@ def __init__(
         self.using_H0 = len(H0) > 0
         self.index = len(H0)
         self.grow_count = 0
+        self.safe_mode = False
 
         self.sim_started_count = np.sum(H["sim_started"])
         self.sim_ended_count = np.sum(H["sim_ended"])
@@ -123,7 +125,15 @@ def __init__(
         self.last_started = -1
         self.last_ended = -1
 
-    def update_history_f(self, D: dict, safe_mode: bool, kill_canceled_sims: bool = False) -> None:
+    def _append_new_fields(self, H_f: npt.NDArray) -> None:
+        dtype_new = np.dtype(list(set(self.H.dtype.descr + H_f.dtype.descr)))
+        H_new = np.zeros(len(self.H), dtype=dtype_new)
+        old_fields = self.H.dtype.names
+        for field in old_fields:
+            H_new[field][: len(self.H)] = self.H[field]
+        self.H = H_new
+
+    def update_history_f(self, D: dict, kill_canceled_sims: bool = False) -> None:
         """
         Updates the history after points have been evaluated
         """
@@ -132,9 +142,12 @@ def update_history_f(self, D: dict, safe_mode: bool, kill_canceled_sims: bool =
         returned_H = D["calc_out"]
         fields = returned_H.dtype.names if returned_H is not None else []
 
+        if returned_H is not None and any([field not in self.H.dtype.names for field in returned_H.dtype.names]):
+            self._append_new_fields(returned_H)
+
         for j, ind in enumerate(new_inds):
             for field in fields:
-                if safe_mode:
+                if self.safe_mode:
                     assert field not in protected_libE_fields, "The field '" + field + "' is protected"
                 if np.isscalar(returned_H[field][j]) or returned_H.dtype[field].hasobject:
                     self.H[field][ind] = returned_H[field][j]
@@ -206,7 +219,7 @@ def update_history_to_gen(self, q_inds: npt.NDArray):
             self.H["gen_informed_time"][q_inds] = t
             self.gen_informed_count += len(q_inds)
 
-    def update_history_x_in(self, gen_worker: int, D: npt.NDArray, safe_mode: bool, gen_started_time: int) -> None:
+    def update_history_x_in(self, gen_worker: int, D: npt.NDArray, gen_started_time: int) -> None:
         """
         Updates the history (in place) when new points have been returned from a gen
 
@@ -221,6 +234,9 @@ def update_history_x_in(self, gen_worker: int, D: npt.NDArray, safe_mode: bool,
         if len(D) == 0:
             return
 
+        if any([field not in self.H.dtype.names for field in D.dtype.names]):
+            self._append_new_fields(D)
+
         t = time.time()
         rows_remaining = len(self.H) - self.index
 
@@ -251,7 +267,7 @@ def update_history_x_in(self, gen_worker: int, D: npt.NDArray, safe_mode: bool,
             update_inds = D["sim_id"]
 
         for field in D.dtype.names:
-            if safe_mode:
+            if self.safe_mode:
                 assert field not in protected_libE_fields, "The field '" + field + "' is protected"
             self.H[field][update_inds] = D[field]
 

libensemble/libE.py

@@ -266,7 +266,7 @@ def manager(
     persis_info,
     alloc_specs,
     libE_specs,
-    hist: np.ndarray,
+    hist: History,
     on_abort: Callable = None,
     on_cleanup: Callable = None,
 ):

libensemble/manager.py

@@ -66,7 +66,7 @@ def report_worker_exc(wrk_exc: Exception = None) -> None:
 
 
 def manager_main(
-    hist: npt.NDArray,
+    hist,
     libE_specs: dict,
     alloc_specs: dict,
     sim_specs: dict,
@@ -168,7 +168,7 @@ class Manager:
 
     def __init__(
         self,
-        hist: npt.NDArray,
+        hist,
         libE_specs: dict,
         alloc_specs: dict,
         sim_specs: dict,
@@ -183,6 +183,7 @@ def __init__(
         self.safe_mode = libE_specs.get("safe_mode")
         self.kill_canceled_sims = libE_specs.get("kill_canceled_sims")
         self.hist = hist
+        self.hist.safe_mode = self.safe_mode
         self.libE_specs = libE_specs
         self.alloc_specs = alloc_specs
         self.sim_specs = sim_specs
@@ -452,9 +453,9 @@ def _update_state_on_worker_msg(self, persis_info: dict, D_recv: dict, w: int) -
             final_data = D_recv.get("calc_out", None)
             if isinstance(final_data, np.ndarray):
                 if calc_status is FINISHED_PERSISTENT_GEN_TAG and self.libE_specs.get("use_persis_return_gen", False):
-                    self.hist.update_history_x_in(w, final_data, self.safe_mode, self.W[w - 1]["gen_started_time"])
+                    self.hist.update_history_x_in(w, final_data, self.W[w - 1]["gen_started_time"])
                 elif calc_status is FINISHED_PERSISTENT_SIM_TAG and self.libE_specs.get("use_persis_return_sim", False):
-                    self.hist.update_history_f(D_recv, self.safe_mode, self.kill_canceled_sims)
+                    self.hist.update_history_f(D_recv, self.kill_canceled_sims)
                 else:
                     logger.info(_PERSIS_RETURN_WARNING)
             self.W[w - 1]["persis_state"] = 0
@@ -467,9 +468,9 @@ def _update_state_on_worker_msg(self, persis_info: dict, D_recv: dict, w: int) -
             self._freeup_resources(w)
         else:
             if calc_type == EVAL_SIM_TAG:
-                self.hist.update_history_f(D_recv, self.safe_mode, self.kill_canceled_sims)
+                self.hist.update_history_f(D_recv, self.kill_canceled_sims)
             if calc_type == EVAL_GEN_TAG:
-                self.hist.update_history_x_in(w, D_recv["calc_out"], self.safe_mode, self.W[w - 1]["gen_started_time"])
+                self.hist.update_history_x_in(w, D_recv["calc_out"], self.W[w - 1]["gen_started_time"])
                 assert (
                     len(D_recv["calc_out"]) or np.any(self.W["active"]) or self.W[w - 1]["persis_state"]
                 ), "Gen must return work when is is the only thing active and not persistent."

libensemble/tests/functionality_tests/test_new_field.py

@@ -0,0 +1,61 @@
+"""
+Runs libEnsemble with Latin hypercube sampling on a simple 1D problem
+
+Execute via one of the following commands (e.g. 3 workers):
+   mpiexec -np 4 python test_1d_sampling.py
+   python test_1d_sampling.py --nworkers 3 --comms local
+   python test_1d_sampling.py --nworkers 3 --comms tcp
+
+The number of concurrent evaluations of the objective function will be 4-1=3.
+"""
+
+# Do not change these lines - they are parsed by run-tests.sh
+# TESTSUITE_COMMS: mpi local
+# TESTSUITE_NPROCS: 2 4
+
+import numpy as np
+
+from libensemble.gen_funcs.sampling import latin_hypercube_sample as gen_f
+
+# Import libEnsemble items for this test
+from libensemble.libE import libE
+from libensemble.tools import add_unique_random_streams, parse_args, save_libE_output
+
+
+def sim_f(In):
+    Out = np.zeros(1, dtype=[("f", float), ("N", int)])
+    Out["f"] = np.linalg.norm(In)
+    Out["N"] = 123
+    return Out
+
+
+if __name__ == "__main__":
+    nworkers, is_manager, libE_specs, _ = parse_args()
+
+    sim_specs = {
+        "sim_f": sim_f,
+        "in": ["x"],
+        "out": [("f", float)],
+    }
+
+    gen_specs = {
+        "gen_f": gen_f,
+        "out": [("x", float, (1,))],
+        "user": {
+            "gen_batch_size": 500,
+            "lb": np.array([-3]),
+            "ub": np.array([3]),
+        },
+    }
+
+    persis_info = add_unique_random_streams({}, nworkers + 1, seed=1234)
+
+    exit_criteria = {"gen_max": 501}
+
+    H, persis_info, flag = libE(sim_specs, gen_specs, exit_criteria, persis_info, libE_specs=libE_specs)
+
+    if is_manager:
+        assert len(H) >= 501
+        assert "N" in H.dtype.names, "New datatype not added to history"
+        print("\nlibEnsemble with random sampling has generated enough points")
+        save_libE_output(H, persis_info, __file__, nworkers)

libensemble/tests/functionality_tests/test_persistent_uniform_sampling_running_mean.py

@@ -45,7 +45,11 @@
     gen_specs = {
         "gen_f": gen_f,
         "persis_in": ["f", "x", "corner_id", "sim_id"],
-        "out": [("sim_id", int), ("corner_id", int), ("x", float, (n,)), ("f_est", float)],
+        "out": [
+            ("sim_id", int),
+            ("corner_id", int),
+            ("x", float, (n,)),
+        ],  # expect ("f_est", float) from gen - test ability for gen to send back "unexpected" field.
         "user": {
             "initial_batch_size": 20,
             "lb": np.array([-3, -2, -1]),

libensemble/tests/regression_tests/test_persistent_aposmm_dfols.py

@@ -122,9 +122,6 @@ def combine_component(x):
             #     assert np.all(grad[l <= 1e-7] <= 0)
 
             #     if not np.all(grad[np.logical_and(u >= 1e-7, l >= 1e-7)] <= 1e-5):
-            #         import ipdb
-
-            #         ipdb.set_trace()
             # else:
             #     d = np.linalg.solve(np.dot(J.T, J), np.dot(J.T, F))
             #     assert np.linalg.norm(d) <= 1e-5

libensemble/tests/unit_tests/setup.py

@@ -62,6 +62,7 @@ def hist_setup2(sim_max=10, H0_in=[]):
     alloc_specs = specs_dump(AllocSpecs())
     H0 = H0_in
     hist = History(alloc_specs, sim_specs, gen_specs, exit_criteria, H0)
+    hist.safe_mode = True
     return hist, sim_specs, gen_specs, exit_criteria, alloc_specs
 
 

libensemble/tests/unit_tests/test_history.py

@@ -71,8 +71,6 @@
 
 exp_x_in_setup2["x"] = x
 
-safe_mode = True
-
 
 def isclose(a, b, rel_tol=1e-09, abs_tol=0.0):
     return abs(a - b) <= max(rel_tol * max(abs(a), abs(b)), abs_tol)
@@ -149,7 +147,7 @@ def test_update_history_x_in_Oempty():
     hist, sim_specs, gen_specs, _, _ = setup.hist_setup2()
     H_o = np.zeros(0, dtype=gen_specs["out"])
     gen_worker = 1
-    hist.update_history_x_in(gen_worker, H_o, safe_mode, np.inf)
+    hist.update_history_x_in(gen_worker, H_o, np.inf)
 
     compare_hists(hist.H, wrs2)
 
@@ -173,7 +171,7 @@ def test_update_history_x_in():
     H_o = np.zeros(size, dtype=gen_specs["out"])
     H_o["x"] = single_rand
 
-    hist.update_history_x_in(gen_worker, H_o, safe_mode, np.inf)
+    hist.update_history_x_in(gen_worker, H_o, np.inf)
     assert isclose(single_rand, hist.H["x"][0])
     assert hist.sim_started_count == 0
     assert hist.index == 1
@@ -185,7 +183,7 @@ def test_update_history_x_in():
     H_o = np.zeros(size, dtype=gen_specs["out"])
     H_o["x"] = gen_specs["gen_f"](size=size)
 
-    hist.update_history_x_in(gen_worker, H_o, safe_mode, np.inf)
+    hist.update_history_x_in(gen_worker, H_o, np.inf)
     # Compare by column
     exp_x = exp_x_in_setup2[: size + 1]
 
@@ -202,7 +200,7 @@ def test_update_history_x_in():
     H_o = np.zeros(size, dtype=gen_specs["out"])
     H_o["x"] = gen_specs["gen_f"](size=size)
 
-    hist.update_history_x_in(gen_worker, H_o, safe_mode, np.inf)
+    hist.update_history_x_in(gen_worker, H_o, np.inf)
     # Compare by column
     exp_x = exp_x_in_setup2
 
@@ -215,7 +213,7 @@ def test_update_history_x_in():
     # Test libE errors when a protected field appears in output from a gen_worker
     H_o = np.zeros(size, dtype=gen_specs["out"] + [("sim_started", bool)])
     try:
-        hist.update_history_x_in(gen_worker, H_o, safe_mode, np.inf)
+        hist.update_history_x_in(gen_worker, H_o, np.inf)
     except AssertionError:
         assert 1, "Failed like it should have"
     else:
@@ -224,7 +222,7 @@ def test_update_history_x_in():
     # Test libE errors when a protected field appears in output from a gen_worker
     H_o = np.zeros(size, dtype=gen_specs["out"] + [("sim_started", bool)])
     try:
-        hist.update_history_x_in(gen_worker, H_o, safe_mode, np.inf)
+        hist.update_history_x_in(gen_worker, H_o, np.inf)
     except AssertionError:
         assert 1, "Failed like it should have"
     else:
@@ -246,7 +244,7 @@ def test_update_history_x_in_sim_ids():
     H_o["x"] = single_rand
     H_o["sim_id"] = 0
 
-    hist.update_history_x_in(gen_worker, H_o, safe_mode, np.inf)
+    hist.update_history_x_in(gen_worker, H_o, np.inf)
     assert isclose(single_rand, hist.H["x"][0])
     assert hist.sim_started_count == 0
     assert hist.index == 1
@@ -258,7 +256,7 @@ def test_update_history_x_in_sim_ids():
     H_o = np.zeros(size, dtype=gen_specs["out"])
     H_o["x"] = gen_specs["gen_f"](size=size)
     H_o["sim_id"] = range(1, 7)
-    hist.update_history_x_in(gen_worker, H_o, safe_mode, np.inf)
+    hist.update_history_x_in(gen_worker, H_o, np.inf)
 
     # Compare by column
     exp_x = exp_x_in_setup2[: size + 1]
@@ -277,7 +275,7 @@ def test_update_history_x_in_sim_ids():
     H_o["x"] = gen_specs["gen_f"](size=size)
     H_o["sim_id"] = range(7, 10)
 
-    hist.update_history_x_in(gen_worker, H_o, safe_mode, np.inf)
+    hist.update_history_x_in(gen_worker, H_o, np.inf)
     # Compare by column
     exp_x = exp_x_in_setup2
 
@@ -357,7 +355,7 @@ def test_update_history_f():
         "calc_type": 2,
     }
 
-    hist.update_history_f(D_recv, safe_mode)
+    hist.update_history_f(D_recv)
     assert isclose(exp_vals[0], hist.H["g"][0])
     assert np.all(hist.H["sim_ended"][0:1])
     assert np.all(~hist.H["sim_ended"][1:10])  # Check the rest
@@ -383,7 +381,7 @@ def test_update_history_f():
         "calc_type": 2,
     }
 
-    hist.update_history_f(D_recv, safe_mode)
+    hist.update_history_f(D_recv)
     assert np.allclose(exp_vals, hist.H["g"])
     assert np.all(hist.H["sim_ended"][0:3])
     assert np.all(~hist.H["sim_ended"][3:10])  # Check the rest
@@ -405,7 +403,7 @@ def test_update_history_f():
         "calc_type": 2,
     }
 
-    hist.update_history_f(D_recv, safe_mode)
+    hist.update_history_f(D_recv)
     assert np.all(hist.H["sim_ended"][0:1])
     assert np.all(~hist.H["sim_ended"][3:10])  # Check the rest
     assert hist.sim_ended_count == 2
@@ -435,7 +433,7 @@ def test_update_history_f_vec():
         "calc_type": 2,
     }
 
-    hist.update_history_f(D_recv, safe_mode)
+    hist.update_history_f(D_recv)
 
     assert isclose(exp_fs[0], hist.H["f"][0])
     assert np.allclose(exp_fvecs[0], hist.H["fvec"][0])
@@ -470,7 +468,7 @@ def test_update_history_f_vec():
         "calc_type": 2,
     }
 
-    hist.update_history_f(D_recv, safe_mode)
+    hist.update_history_f(D_recv)
 
     assert np.allclose(exp_fs, hist.H["f"])
     assert np.allclose(exp_fvecs, hist.H["fvec"])
@@ -507,7 +505,7 @@ def test_update_history_f_vec():
         "calc_type": 2,
     }
 
-    hist.update_history_f(D_recv, safe_mode)
+    hist.update_history_f(D_recv)
 
     assert np.allclose(exp_fs, hist.H["f"])
     assert np.allclose(exp_fvecs, hist.H["fvec"])