Merge many Ramonco developments (#109)

* GIES and AVO Smeaheia in progress (#72)

Co-authored-by: Martha Økland Lien <mlie@bgo-1907.ad.norceresearch.no>
Co-authored-by: Rolf J. Lorentzen <rolo@norceresearch.no>

* Ramon co pet (#73)

* GIES and AVO Smeaheia in progress

* avo updted, grav included, method for joint grav and avo. calc_pem method in flow_rock that is read by flow_avo

* softsand.py represents soft sand rock physics model, added 4D gravity and 4D avo

* give overburden velocity values to inactive cells

* minor edits

---------

Co-authored-by: Martha Økland Lien <mlie@bgo-1907.ad.norceresearch.no>
Co-authored-by: Rolf J. Lorentzen <rolo@norceresearch.no>

* Rename keys variable

* Ramon co pet (#74)

* GIES and AVO Smeaheia in progress

* avo updted, grav included, method for joint grav and avo. calc_pem method in flow_rock that is read by flow_avo

* softsand.py represents soft sand rock physics model, added 4D gravity and 4D avo

* give overburden velocity values to inactive cells

* minor edits

* Compute reflection coefficients for active cells

---------

Co-authored-by: Martha Økland Lien <mlie@bgo-1907.ad.norceresearch.no>
Co-authored-by: Rolf J. Lorentzen <rolo@norceresearch.no>

* Bug fix of avo and gravity (#77)

* Bug fix of avo and gravity

* Bug fix

* Add functionality for updating pressure and saturations directly (#80)

* invert for dynamic variables first iteration (#94)

* NE in pipt file is now used instead of the whole loaded ensemble if this number is smaller+bug fix (#95)

* subsidence

* subsidence (#96)

* stash issue resolve

* Subsidence (#97)

* subsidence

* stash issue resolve

* Compression also for 2D data (#107)

* Smeaheia update and GIES modifications

* More updates after merge of Ramonco branch

---------

Co-authored-by: Martha Økland Lien <mlie@bgo-1907.ad.norceresearch.no>
Co-authored-by: Rolf J. Lorentzen <rolo@norceresearch.no>
Co-authored-by: Rolf Johan Lorentzen <rolo@bgo-1151.ad.norceresearch.no>
Co-authored-by: Kjersti <48013553+kjei@users.noreply.github.com>

Author: 5 people

ensemble/ensemble.py

@@ -15,6 +15,7 @@
 from tqdm.auto import tqdm
 from p_tqdm import p_map
 import logging
+from geostat.decomp import Cholesky  # Making realizations
 
 # Internal imports
 import pipt.misc_tools.analysis_tools as at
@@ -26,7 +27,6 @@
 from misc.system_tools.environ_var import OpenBlasSingleThread  # Single threaded OpenBLAS runs
 
 
-
 class Ensemble:
     """
     Class for organizing misc. variables and simulator for an ensemble-based inversion run. Here, the forecast step
@@ -139,15 +139,24 @@ def __init__(self, keys_en, sim, redund_sim=None):
                 # individually).
                 self.state = {key: val for key, val in tmp_load.items()}
 
-                # Find the number of ensemble members from state variable
+                # Find the number of ensemble members from loaded state variables
                 tmp_ne = []
                 for tmp_state in self.state.keys():
                     tmp_ne.extend([self.state[tmp_state].shape[1]])
-                if max(tmp_ne) != min(tmp_ne):
-                    print('\033[1;33mInput states have different ensemble size\033[1;m')
-                    sys.exit(1)
-                self.ne = min(tmp_ne)
-                
+
+                if 'ne' not in self.keys_en: # NE not specified in input file
+                    if max(tmp_ne) != min(tmp_ne): #Check loaded ensembles are the same size (if more than one state variable)
+                        print('\033[1;33mInput states have different ensemble size\033[1;m')
+                        sys.exit(1)
+                    self.ne = min(tmp_ne) # Use the number of ensemble members in loaded ensemble
+                else:
+                    # Use the number of ensemble members specified in input file (may be fewer than loaded)
+                    self.ne = int(self.keys_en['ne'])
+                    if self.ne <= min(tmp_ne):
+                        # pick correct number of ensemble members
+                        self.state = {key: val[:,:self.ne] for key, val in self.state.items()}
+                    else:
+                        print('\033[1;33mInput states are smaller than NE\033[1;m')
         if 'multilevel' in self.keys_en:
             ml_info = extract.extract_multilevel_info(self.keys_en)
             self.multilevel, self.tot_level, self.ml_ne, self.ML_error_corr, self.error_comp_scheme, self.ML_corr_done = ml_info
@@ -338,6 +347,20 @@ def calc_prediction(self, input_state=None, save_prediction=None):
             # Index list of ensemble members
             list_member_index = list(range(self.ne))
 
+            # modified by xluo, for including the simulation of the mean reservoir model
+            # as used in the RLM-MAC algorithm
+            if 'daalg' in self.keys_en and self.keys_en['daalg'][1] == 'gies':
+                list_state.append({})
+                list_member_index.append(self.ne)
+
+                for key in self.state.keys():
+                    tmp_state = np.zeros(list_state[0][key].shape[0])
+
+                    for i in range(self.ne):
+                        tmp_state += list_state[i][key]
+
+                    list_state[self.ne][key] = tmp_state / self.ne
+
             if no_tot_run==1: # if not in parallel we use regular loop
                 en_pred = [self.sim.run_fwd_sim(state, member_index) for state, member_index in
                            tqdm(zip(list_state, list_member_index), total=len(list_state))]
@@ -392,6 +415,7 @@ def calc_prediction(self, input_state=None, save_prediction=None):
             else: # Run prediction in parallel using p_map
                 en_pred = p_map(self.sim.run_fwd_sim, list_state,
                                 list_member_index, num_cpus=no_tot_run, disable=self.disable_tqdm)
+
             # List successful runs and crashes
             list_crash = [indx for indx, el in enumerate(en_pred) if el is False]
             list_success = [indx for indx, el in enumerate(en_pred) if el is not False]

pipt/loop/assimilation.py

@@ -529,8 +529,12 @@ def post_process_forecast(self):
                     for k in pred_data_tmp[i]:  # DATATYPE
                         if vintage < len(self.ensemble.sparse_info['mask']) and \
                                 len(pred_data_tmp[i][k]) == int(np.sum(self.ensemble.sparse_info['mask'][vintage])):
-                            self.ensemble.pred_data[i][k] = np.zeros(
-                                (len(self.ensemble.obs_data[i][k]), self.ensemble.ne))
+                            if self.ensemble.keys_da['daalg'][1] == 'gies':
+                                self.ensemble.pred_data[i][k] = np.zeros(
+                                    (len(self.ensemble.obs_data[i][k]), self.ensemble.ne+1))
+                            else:
+                                self.ensemble.pred_data[i][k] = np.zeros(
+                                    (len(self.ensemble.obs_data[i][k]), self.ensemble.ne))
                             for m in range(pred_data_tmp[i][k].shape[1]):
                                 data_array = self.ensemble.compress(pred_data_tmp[i][k][:, m], vintage,
                                                                     self.ensemble.sparse_info['use_ensemble'])

pipt/loop/ensemble.py

@@ -540,7 +540,24 @@ def _ext_scaling(self):
         self.state_scaling = at.calc_scaling(
             self.prior_state, self.list_states, self.prior_info)
 
-        self.Am = None
+        delta_scaled_prior = self.state_scaling[:, None] * \
+            np.dot(at.aug_state(self.prior_state, self.list_states), self.proj)
+
+        u_d, s_d, v_d = np.linalg.svd(delta_scaled_prior, full_matrices=False)
+
+        # remove the last singular value/vector. This is because numpy returns all ne values, while the last is actually
+        # zero. This part is a good place to include eventual additional truncation.
+        energy = 0
+        trunc_index = len(s_d) - 1  # inititallize
+        for c, elem in enumerate(s_d):
+            energy += elem
+            if energy / sum(s_d) >= self.trunc_energy:
+                trunc_index = c  # take the index where all energy is preserved
+                break
+        u_d, s_d, v_d = u_d[:, :trunc_index +
+                            1], s_d[:trunc_index + 1], v_d[:trunc_index + 1, :]
+        self.Am = np.dot(u_d, np.eye(trunc_index+1) *
+                         ((s_d**(-1))[:, None]))  # notation from paper
 
     def save_temp_state_assim(self, ind_save):
         """
@@ -694,7 +711,7 @@ def compress(self, data=None, vintage=0, aug_coeff=None):
 
             data_array = None
 
-        elif aug_coeff is None:
+        elif aug_coeff is None: # compress predicted data
 
             data_array, wdec_rec = self.sparse_data[vintage].compress(data)
             rec = self.sparse_data[vintage].reconstruct(
@@ -703,7 +720,7 @@ def compress(self, data=None, vintage=0, aug_coeff=None):
                 self.data_rec.append([])
             self.data_rec[vintage].append(rec)
 
-        elif not aug_coeff:
+        elif not aug_coeff: # compress true data, aug_coeff = false
 
             options = copy(self.sparse_info)
             # find the correct mask for the vintage

pipt/misc_tools/analysis_tools.py

@@ -542,8 +542,9 @@ def calc_objectivefun(pert_obs, pred_data, Cd):
     data_misfit : array-like
         Nex1 array containing objective function values.
     """
-    ne = pred_data.shape[1]
-    r = (pred_data - pert_obs)
+    #ne = pred_data.shape[1]
+    ne = pert_obs.shape[1]
+    r = (pred_data[:, :ne] - pert_obs)  # This is necessary to use to gies code that xilu has implemented
     if len(Cd.shape) == 1:
         precission = Cd**(-1)
         data_misfit = np.diag(r.T.dot(r*precission[:, None]))

pipt/misc_tools/extract_tools.py

@@ -271,8 +271,12 @@ def organize_sparse_representation(info: Union[dict,list]) -> dict:
     sparse['dim'] = [dim[2], dim[1], dim[0]]
 
     # Read mask_files
-    sparse['mask'] = [] 
-    for idx, filename in enumerate(info['mask'], start=1):
+    sparse['mask'] = []
+    m_info = info['mask']
+    # allow for one mask with filename given as string
+    if isinstance(m_info, str):
+        m_info = [m_info]
+    for idx, filename in enumerate(m_info, start=1):
         if not os.path.exists(filename):
             mask = np.ones(sparse['dim'], dtype=bool)
             np.savez(f'mask_{idx}.npz', mask=mask)