add exporting of disclosures to json format

lcopt/export_disclosure.py

@@ -0,0 +1,117 @@
+import numpy as np
+import json
+from scipy.sparse import coo_matrix
+
+def matrix_to_coo(m):
+    m_coo = coo_matrix(m)
+    return [[[int(m_coo.row[i]), int(m_coo.col[i])], float(m_coo.data[i])] for i, _ in enumerate(m_coo.data)]
+
+def specify_matrix(model, ps_id):
+
+    eps = model.evaluated_parameter_sets
+
+    if isinstance(ps_id, str):
+        ps = eps[ps_id]
+    else:
+        ps = eps[list(eps.keys())[ps_id]]
+
+    matrix = model.matrix.copy()
+
+    for k, v in ps.items():
+        if k[:4] == "n_p_":
+            coords = [int(x) for x in k.split("_")[-2:]]
+            matrix[coords[0], coords[1]] = v
+
+    return matrix
+
+def export_disclosure(model, parameter_set=None):
+
+    if parameter_set is None:
+        matrix = model.matrix.copy()
+        efn = '{}_unspecified.json'.format(model.name.replace(" ", "_"))
+    else:
+        matrix = specify_matrix(model, parameter_set)
+        efn = '{}_ps_{}.json'.format(model.name.replace(" ", "_"), parameter_set)
+
+    background = [(i, x) for i, x in enumerate(model.names) if list(matrix.sum(axis=0))[i] == 0]
+    foreground = [(i, x) for i, x in enumerate(model.names) if list(matrix.sum(axis=0))[i] != 0]
+    fu = [(i, x) for i, x in enumerate(model.names) if list(matrix.sum(axis=1))[i] == 0 and list(matrix.sum(axis=0))[i] != 0]
+    unused = [(i, x) for i, x in enumerate(model.names) if list(matrix.sum(axis=1))[i] == 0 and list(matrix.sum(axis=0))[i] == 0]
+
+    background = sorted(list(set(background) - set(unused))) # get rid of unused items
+    foreground = sorted(list(set(foreground) - set(unused))) # get rid of unused items
+    foreground = fu + [x for x in foreground if x not in fu] # set fu to be the first item in the foreground matrix
+
+    #split background into technosphere and biosphere portions
+    technosphere = [x for x in background if model.database['items'][model.get_exchange(x[1])]['lcopt_type'] == "input"]
+    biosphere = [x for x in background if model.database['items'][model.get_exchange(x[1])]['lcopt_type'] == "biosphere"]
+
+    # Create Af
+    l = len(foreground)
+    Af = np.zeros((l,l))
+
+    for i, c in enumerate(foreground):
+        c_lookup = c[0]
+        for j, r in enumerate(foreground):
+            r_lookup = r[0]
+            Af[i, j] = matrix[c_lookup, r_lookup]
+
+    # Create Ad
+    Ad = np.zeros((len(background),l))
+
+    Ad = np.zeros((len(technosphere),l))
+    for i, c in enumerate(foreground):
+        c_lookup = c[0]
+        for j, r in enumerate(technosphere):
+            r_lookup = r[0]
+            Ad[j, i] = matrix[r_lookup,c_lookup ]
+
+    # Create Bf
+    Bf = np.zeros((len(biosphere),l))
+    for i, c in enumerate(foreground):
+        c_lookup = c[0]
+        for j, r in enumerate(biosphere):
+            r_lookup = r[0]
+            Bf[j, i] = matrix[r_lookup,c_lookup]
+
+    # Get technosphere and biosphere data from external links
+    technosphere_links = [model.database['items'][model.get_exchange(x[1])].get('ext_link',(None, '{}'.format(x[1]))) for x in background if model.database['items'][model.get_exchange(x[1])]['lcopt_type'] == "input"]
+    biosphere_links = [model.database['items'][model.get_exchange(x[1])]['ext_link'] for x in background if model.database['items'][model.get_exchange(x[1])]['lcopt_type'] == "biosphere"]
+
+    # Get technosphere ids
+    technosphere_ids = []
+    for t in technosphere_links:
+        y = t[0]
+        if y is None:
+            technosphere_ids.append((t[1], "cutoff exchange"))
+        else:
+            e = [i for i, x in enumerate (model.external_databases) if x['name'] == y][0]
+            technosphere_ids.append((model.external_databases[e]['items'][t]['name'], model.external_databases[e]['items'][t]['activity']))
+
+    # Get biosphere ids
+    biosphere_ids = []
+    for b in biosphere_links:
+        y = b[0]
+        e = [i for i, x in enumerate (model.external_databases) if x['name'] == y][0]
+        biosphere_ids.append((model.external_databases[e]['items'][b]))
+
+    # final preparations
+    foreground_names = [(i, x[1]) for i, x in enumerate(foreground)]
+    technosphere_names = [{'ecoinvent_name': technosphere_ids[i][0], 'ecoinvent_id':technosphere_ids[i][1], 'brightway_id':technosphere_links[i]} for i, x in enumerate(technosphere)]
+    biosphere_names = [{'name':"{}, {}, {}".format(biosphere_ids[i]['name'], biosphere_ids[i]['type'],  ",".join(biosphere_ids[i]['categories'])),'biosphere3_id': biosphere_links[i]} for i, x in enumerate(biosphere)]
+
+    # collate the data
+    data = {
+        'foreground flows':foreground_names,
+        'Af':matrix_to_coo(Af),
+        'background flows': technosphere_names,
+        'Ad':matrix_to_coo(Ad),
+        'Foreground emissions': biosphere_names,
+        'Bf':matrix_to_coo(Bf)
+    }
+
+    # export the data   
+    with open(efn, 'w') as f:
+        json.dump(data, f)
+
+    return efn

lcopt/model.py

@@ -11,6 +11,7 @@
 from lcopt.interact import FlaskSandbox
 from lcopt.bw2_export import Bw2Exporter
 from lcopt.analysis import Bw2Analysis
+from .export_disclosure import export_disclosure
 from .utils import check_for_config, lcopt_bw2_autosetup, DEFAULT_PROJECT_STEM, bw2_project_exists, write_search_index, FORWAST_PROJECT_NAME, upgrade_old_default, lcopt_bw2_forwast_setup
 # This is a copy straight from bw2data.query, extracted so as not to cause a dependency.
 #from lcopt.bw2query import Query, Dictionaries, Filter
@@ -853,3 +854,7 @@ def analyse(self, demand_item, demand_item_code):
         self.result_set = my_analysis.run_analyses(demand_item, demand_item_code, **self.analysis_settings)
 
         return True
+
+# << Disclosures >> #
+    def export_disclosure(self, parameter_set=None):
+        return export_disclosure(self, parameter_set)