Param mapping

python/ngen_cal/src/ngen/cal/calibratable.py

@@ -48,9 +48,9 @@ def id(self) -> str:
     @property
     def variables(self) -> Series:
         """
-            Index series of variables
+            Index of unique variables to permute
         """
-        return Series(self.df.index.values)
+        return Series(self.df.index.unique().values)
 
     @property
     def bounds(self) -> tuple[Series, Series]:

python/ngen_cal/src/ngen/cal/ngen.py

@@ -46,13 +46,13 @@ def _params_as_df(params: Mapping[str, Parameters], name: str = None):
             df['model'] = k
             df.rename(columns={'name':'param'}, inplace=True)
             dfs.append(df)
-        return pd.concat(dfs)
+        return pd.concat(dfs).set_index('alias')
     else:
         p = params.get(name, [])
         df = pd.DataFrame([s.__dict__ for s in p])
         df['model'] = name
         df.rename(columns={'name':'param'}, inplace=True)
-        return df
+        return df.set_index('alias')
 
 def _map_params_to_realization(params: Mapping[str, Parameters], realization: Realization):
     # don't even think about calibration multiple formulations at once just yet..

python/ngen_cal/src/ngen/cal/parameter.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
-from pydantic import BaseModel, Field
-from typing import Sequence
+from pydantic import BaseModel, Field, root_validator
+from typing import Sequence, Mapping, Optional
 
 class Parameter(BaseModel, allow_population_by_field_name = True):
     """
@@ -11,5 +11,13 @@ class Parameter(BaseModel, allow_population_by_field_name = True):
     min: float
     max: float
     init: float
+    alias: Optional[str]
+
+    @root_validator
+    def _set_alias(cls, values: dict) -> dict:
+        alias = values.get('alias', None)
+        if alias is None:
+            values['alias'] = values['name']
+        return values
 
 Parameters = Sequence[Parameter]

python/ngen_cal/tests/conftest.py

@@ -1,5 +1,7 @@
+from __future__ import annotations
+
 import pytest
-from typing import Generator, List
+from typing import Generator, List, Mapping
 from pathlib import Path
 from copy import deepcopy
 import json
@@ -11,6 +13,7 @@
 from ngen.cal.calibration_cathment import CalibrationCatchment
 from ngen.cal.model import EvaluationOptions
 from ngen.cal.agent import Agent
+from ngen.cal.parameter import Parameter
 from hypy import Nexus
 
 from .utils import *
@@ -219,3 +222,32 @@ def explicit_catchments(nexus, fabric, workdir) -> Generator[ List[ CalibrationC
         cat = CalibrationCatchment(workdir, id, nexus, start, end, fabric, 'Q_Out', eval_options, data)
         catchments.append(cat)
     yield catchments
+
+@pytest.fixture
+def multi_model_shared_params() -> Mapping[str, list[Parameter]]:
+    p1 = Parameter(name='a', min=0, max=1, init=0)
+    p2 = Parameter(name='d', min=2, max=3, init=0)
+    p3 = Parameter(name='c', min=0, max=1, init=0)
+    p4 = Parameter(name='a', min=0, max=1, init=0)
+    params = {'A':[p1], 'B':[p2, p4], 'C':[p3, p4]}
+
+    return params
+
+@pytest.fixture
+def multi_model_shared_params2() -> Mapping[str, list[Parameter]]:
+    p1 = Parameter(name='a', min=0, max=1, init=0)
+    p2 = Parameter(name='a', min=0, max=1, init=0)
+    p3 = Parameter(name='c', min=0, max=1, init=0)
+    params = {'A':[p1, p3], 'B':[p2]}
+
+    return params
+
+@pytest.fixture
+def multi_model_alias_params() -> Mapping[str, list[Parameter]]:
+    p1 = Parameter(name='a', alias='c', min=0, max=1, init=0)
+    p2 = Parameter(name='b', alias='c', min=0, max=1, init=0)
+    p3 = Parameter(name='d', min=0, max=1, init=0)
+
+    params = {'A':[p1,p3], 'B':[p2, p3], 'C':[p3]}
+
+    return params

python/ngen_cal/tests/test_params.py

@@ -0,0 +1,60 @@
+from __future__ import annotations
+
+from ngen.cal.ngen import _params_as_df
+import pandas as pd
+
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from typing import Mapping
+    from ngen.cal.parameter import Parameter
+
+def test_multi_params(multi_model_shared_params: Mapping[str, list[Parameter]]):
+    # This is essentially the path the params go through from
+    # creation in model.py to update in search.py
+    params = _params_as_df(multi_model_shared_params)
+    params = pd.DataFrame(params).rename(columns={'init':'0'})
+    # create new iteration from old
+    params['1'] = params['0']
+    #update the parameters by index
+    idx1 = 'a'
+    idx2 = 'c'
+    params.loc[idx1, '1'] = 0.5
+    pa = params[ params['model'] == 'A' ].loc[idx1]
+    pb = params[ params['model'] == 'B' ].loc[idx1]
+    pc = params[ params['model'] == 'C' ].loc[idx2]
+
+    assert pa.drop('model').equals( pb.drop('model') )
+    # ensure unique params/alias are not modifed by selection
+    assert pa.loc['1'] != pc.loc['1']
+
+def test_multi_params2(multi_model_shared_params2: Mapping[str, list[Parameter]]):
+    # This is essentially the path the params go through from
+    # creation in model.py to update in search.py
+    params = _params_as_df(multi_model_shared_params2)
+    params = pd.DataFrame(params).rename(columns={'init':'0'})
+    # create new iteration from old
+    params['1'] = params['0']
+    #update the parameters by index
+    params.loc['a', '1'] = 0.5
+    pa = params[ params['model'] == 'A' ].drop('model', axis=1).loc['a']
+    pb = params[ params['model'] == 'B' ].drop('model', axis=1).loc['a']
+    assert pa.equals( pb )
+
+def test_alias_params(multi_model_alias_params: Mapping[str, list[Parameter]]):
+    # This is essentially the path the params go through from
+    # creation in model.py to update in search.py
+    assert multi_model_alias_params['C'][0].alias == multi_model_alias_params['C'][0].name
+    params = _params_as_df(multi_model_alias_params)
+    params = pd.DataFrame(params).rename(columns={'init':'0'})
+    # create new iteration from old
+    params['1'] = params['0']
+    #update the parameters by index
+
+    params.loc['c', '1'] = 0.5
+    pa = params[ params['model'] == 'A' ]
+    pb = params[ params['model'] == 'B' ]
+
+    assert pa.drop(['model', 'param'], axis=1).equals( pb.drop(['model', 'param'], axis=1) )
+
+# TODO test/document case where params have same name but different min/max/init values