FIX: Equilibrator update + reqs for building Dockerfile

docker/requirements.txt

@@ -1,5 +1,7 @@
 bokeh>=0.12.1
 cobra>0.13
+equilibrator-api
+equilibrator-cache
 ipdb
 lxml
 networkx

pytfa/thermo/equilibrator.py

@@ -40,7 +40,8 @@ def build_thermo_from_equilibrator(model, T=TEMPERATURE_0):
         ccache = create_compound_cache_from_quilt()
         logger.debug("eQuilibrator compound cache is loaded.")
 
-    cc = ComponentContribution(temperature=Q_(str(T) + "K"))
+    cc = ComponentContribution()
+    cc.temperature=Q_(str(T) + "K")
 
     thermo_data = {"name": "eQuilibrator", "units": "kJ/mol", "cues": {}}
     met_to_comps = compat.map_cobra_metabolites(ccache, model.metabolites)

tests/test_equilibrator.py

@@ -3,48 +3,56 @@
 """Tests for the equilibrator integration."""
 import pytest
 import sys
+from pytfa.io.json import json_dumps_model, json_loads_model
 
 
-t_model = None
-
 
 # previous version aren't compatible for the required equlibrator_cache version
 # this annotation is not necessary since this file shoud be ignored in the test
 # collection phase for previous versions, but it makes it more clear.
 @pytest.mark.skipif(sys.version_info < (3, 6), reason="requires >= python3.6")
-def test_load_with_equi_thermo():
+def test_load_with_equi_thermo(request):
     """Build thermo data structure with equilibrator."""
+    t_model = request.config.cache.get('model', None)
     from pytfa.thermo.equilibrator import build_thermo_from_equilibrator
     from pytfa import ThermoModel
     from settings import cobra_model
-
-    global t_model
     cmodel = cobra_model.copy()
     for met in cmodel.metabolites:
         # normalize but don't overwrite
         met.annotation["seed.compound"] = met.annotation["seed_id"]
     thermo_data = build_thermo_from_equilibrator(cmodel)
     t_model = ThermoModel(thermo_data, cmodel)
     t_model.solver = "optlang-glpk"
+    request.config.cache.set('model', json_dumps_model(t_model))
+
 
 
+@pytest.mark.dependency(depends=['test_load_with_equi_thermo'])
 @pytest.mark.skipif(sys.version_info < (3, 6), reason="requires >= python3.6")
-def test_equilibrator_preparation():
+def test_equilibrator_preparation(request):
     """Prepare using equilibrator_api."""
-    global t_model
+    t_model = json_loads_model(request.config.cache.get('model',None))
+    assert t_model is not None
     t_model.prepare()
+    request.config.cache.set('model', json_dumps_model(t_model))
 
 
+@pytest.mark.dependency(depends=['test_equilibrator_preparation'])
 @pytest.mark.skipif(sys.version_info < (3, 6), reason="requires >= python3.6")
-def test_equilibrator_conversion():
+def test_equilibrator_conversion(request):
     """Convert by the usual method in `ThermoModel`."""
-    global t_model
+    t_model = json_loads_model(request.config.cache.get('model',None))
+    assert t_model is not None
     t_model.convert()
+    request.config.cache.set('model', json_dumps_model(t_model))
 
 
+@pytest.mark.dependency(depends=['test_equilibrator_conversion'])
 @pytest.mark.skipif(sys.version_info < (3, 6), reason="requires >= python3.6")
-def test_equilibrator_optim():
+def test_equilibrator_optim(request):
     """LP optimization by the usual method in `ThermoModel`."""
-    global t_model
+    t_model = json_loads_model(request.config.cache.get('model',None))
+    assert t_model is not None
     solution = t_model.optimize()
     return sum(solution.fluxes)