2.5.92

Author: vlkong

examples/mp/docplexcloud/diet_pandas.py

@@ -19,31 +19,6 @@
 from docplex.mp.model import Model
 from docplex.util.environment import get_environment
 
-# We need a lock so that we can make write_all_outputs
-# atomic. All output operation (update of the ``outputs`` dict or write) should
-# lock this Lock to ensure that the output is not in an unfinished state when
-# the job is aborted.
-# We also need a stop callback that will lock until this Lock is released
-# to make sure that any write operation finishes.
-output_lock = threading.Lock()
-
-
-def set_stop_callback(cb):
-    env = get_environment()
-    try:
-        env.set_stop_callback(cb)
-    except AttributeError:
-        # env.set_stop_callback does not exists -> older version of docplex
-        # use work around
-        try:
-            import docplex.worker.solvehook as worker_env
-            hook = worker_env.get_solve_hook()
-            if hook:
-                hook.set_stop_callback(cb)
-        finally:
-            # ignore errors
-            pass
-
 
 def get_all_inputs():
     '''Utility method to read a list of files and return a tuple with all
@@ -55,49 +30,14 @@ def get_all_inputs():
     result = {}
     env = get_environment()
     for iname in [f for f in os.listdir('.') if splitext(f)[1] == '.csv']:
-        with env.get_input_stream(iname) as in_stream:
-            df = pandas.read_csv(in_stream)
-            datasetname, _ = splitext(iname)
-            result[datasetname] = df
+        df = env.read_df(iname, index_col=None)
+        datasetname, _ = splitext(iname)
+        result[datasetname] = df
     return result
 
-def callonce(f):
-    @wraps(f)
-    def wrapper(*args, **kwargs):
-        if not wrapper.called:
-            wrapper.called = True
-            return f(*args, **kwargs)
-        print('Function already called once.')
-    wrapper.called = False
-    return wrapper
-
-@callonce
-def write_all_outputs(outputs):
-    '''Write all dataframes in ``outputs`` as .csv.
-
-    Args:
-        outputs: The map of outputs 'outputname' -> 'output df'
-    '''
-    global output_lock
-    with output_lock:
-        for (name, df) in iteritems(outputs):
-            csv_file = '%s.csv' % name
-            with get_environment().get_output_stream(csv_file) as fp:
-                fp.write(df.to_csv(index=False))
-    if len(outputs) == 0:
-        print("Warning: no outputs written")
-
-
-
-
 
 def wait_and_save_all_cb(outputs):
-    global output_lock
-    # just wait for the output_lock to be available
-    with output_lock:
-        pass
-    # write outputs
-    write_all_outputs(outputs)
+    get_environment().store_solution(outputs)
 
 
 def mp_solution_to_df(solution):
@@ -157,7 +97,8 @@ def build_diet_model(inputs, **kwargs):
     inputs = get_all_inputs()
     outputs = {}
 
-    set_stop_callback(partial(wait_and_save_all_cb, outputs))
+    # The abort callbacks are called when the docplexcloud job is aborted
+    get_environment().abort_callbacks += [partial(wait_and_save_all_cb, outputs)]
 
     mdl = build_diet_model(inputs)
 
@@ -170,12 +111,5 @@ def build_diet_model(inputs, **kwargs):
         mdl.report_kpis()
         # Save the CPLEX solution as 'solution.csv' program output
         solution_df = mp_solution_to_df(mdl.solution)
-
-        with output_lock:
-            # makes sure the solution is fully assigned before we quit
-            outputs['solution'] = solution_df
-
-        # This allows the solution to be saved if this script
-        # is not aborted. If the script is aborted, the `wait_and_save_all_cb`
-        # takes care of the writing
-        write_all_outputs(outputs)
+        outputs['solution'] = solution_df
+        get_environment().store_solution(outputs)

examples/mp/jupyter/sktrans/transformers.ipynb

@@ -813,7 +813,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "### Rounding a fractional solution\n",
+    "### Rouding a fractional solution\n",
     "\n",
     "An idea that often comes up to deal with fractional solutions is to solve an LP and then round the fractional numbers in order to find an integer solution. However, because the optimal solution is always on the edge of the feasible region, rounding can lead to an infeasible solution, that is, a solution that lies outside the feasible region. In the case of the telephone problem, rounding would produce infeasible results for both types of phones. \n",
     "\n",
@@ -1146,7 +1146,7 @@
     "\n",
     "To optimize a portfolio in terms of risk and return, an investor will evaluate the sum of expected returns of the securities, the total variances of the securities, and the covariances of the securities.  A portfolio that contains a large number of positively covariant securities is more risky (and potentially more rewarding) than one that contains a mix of positively and negatively covariant securities. \n",
     "\n",
-    "### Uses of portfolio optimization\n",
+    "### Potfolio optimization: what use?\n",
     "\n",
     "Portfolio optimization is used to select securities to maximize the rate of return, while managing the volatility of the portfolio and remaining within the investment budget. \n",
     "\n",

examples/mp/jupyter/tutorials/Beyond_Linear_Programming.ipynb

@@ -68,16 +68,16 @@ def day_to_day_week(day):
           ("Emergency", "monday", 18, 2, 3, 7),
           ("Consultation", "monday", 8, 12, 10, 13),
           ("Consultation", "monday", 12, 18, 8, 12),
-          ("Cardiac Care", "monday", 8, 12, 10, 13),
-          ("Cardiac Care", "monday", 12, 18, 8, 12),
+          ("Cardiac_Care", "monday", 8, 12, 10, 13),
+          ("Cardiac_Care", "monday", 12, 18, 8, 12),
           ("Emergency", "tuesday", 8, 12, 4, 7),
           ("Emergency", "tuesday", 12, 18, 2, 5),
           ("Emergency", "tuesday", 18, 2, 3, 7),
           ("Consultation", "tuesday", 8, 12, 10, 13),
           ("Consultation", "tuesday", 12, 18, 8, 12),
-          ("Cardiac Care", "tuesday", 8, 12, 4, 7),
-          ("Cardiac Care", "tuesday", 12, 18, 2, 5),
-          ("Cardiac Care", "tuesday", 18, 2, 3, 7),
+          ("Cardiac_Care", "tuesday", 8, 12, 4, 7),
+          ("Cardiac_Care", "tuesday", 12, 18, 2, 5),
+          ("Cardiac_Care", "tuesday", 18, 2, 3, 7),
           ("Emergency", "wednesday", 2, 8, 3, 5),
           ("Emergency", "wednesday", 8, 12, 4, 7),
           ("Emergency", "wednesday", 12, 18, 2, 5),
@@ -105,18 +105,18 @@ def day_to_day_week(day):
           ("Geriatrics", "sunday", 8, 10, 2, 5)]
 
 NURSE_SKILLS = {"Anne": ["Anaesthesiology", "Oncology", "Pediatrics"],
-                "Betsy": ["Cardiac Care"],
+                "Betsy": ["Cardiac_Care"],
                 "Cathy": ["Anaesthesiology"],
                 "Cecilia": ["Anaesthesiology", "Oncology", "Pediatrics"],
-                "Chris": ["Cardiac Care", "Oncology", "Geriatrics"],
-                "Gloria": ["Pediatrics"], "Jemma": ["Cardiac Care"],
+                "Chris": ["Cardiac_Care", "Oncology", "Geriatrics"],
+                "Gloria": ["Pediatrics"], "Jemma": ["Cardiac_Care"],
                 "Joyce": ["Anaesthesiology", "Pediatrics"],
                 "Julie": ["Geriatrics"], "Juliet": ["Pediatrics"],
-                "Kate": ["Pediatrics"], "Nancy": ["Cardiac Care"],
+                "Kate": ["Pediatrics"], "Nancy": ["Cardiac_Care"],
                 "Nathalie": ["Anaesthesiology", "Geriatrics"],
                 "Patrick": ["Oncology"], "Suzanne": ["Pediatrics"],
                 "Wendie": ["Geriatrics"],
-                "Zoe": ["Cardiac Care"]
+                "Zoe": ["Cardiac_Care"]
                 }
 
 VACATIONS = [("Anne", "friday"),
@@ -195,7 +195,7 @@ def day_to_day_week(day):
                            ("Joan", "Anne")
                            ]
 
-SKILL_REQUIREMENTS = [("Emergency", "Cardiac Care", 1)]
+SKILL_REQUIREMENTS = [("Emergency", "Cardiac_Care", 1)]
 
 DEFAULT_WORK_RULES = TWorkRules(40)
 
@@ -218,7 +218,8 @@ def __str__(self):
         dept2 = self.department[0:4].upper()
         # keep 3 days of weekday
         dayname = self.day[0:3]
-        return '{}_{}_{:02d}'.format(dept2, dayname, self.start_time)
+        return '{}_{}_{:02d}'.format(dept2, dayname, self.start_time).replace(" ", "_")
+
 
 
 class ShiftActivity(object):
@@ -296,6 +297,7 @@ def load_data(model, shifts_, nurses_, nurse_skills, vacations_=None,
     # computed
     model.departments = set(sh.department for sh in model.shifts)
 
+
     print('#nurses: {0}'.format(len(model.nurses)))
     print('#shifts: {0}'.format(len(model.shifts)))
     print('#vacations: {0}'.format(len(model.vacations)))
@@ -370,7 +372,7 @@ def setup_constraints(model):
             v += 1
             model.add_constraint(nurse_assigned[vac_n, shift] == 0,
                                  "medium_vacations_{0!s}_{1!s}_{2!s}".format(vac_n, vac_day, shift))
-    print('#vacation cts: {0}'.format(v))
+    #print('#vacation cts: {0}'.format(v))
 
     # a nurse cannot be assigned overlapping shifts
     # post only one constraint per couple(s1, s2)
@@ -385,7 +387,7 @@ def setup_constraints(model):
                 for n in all_nurses:
                     model.add_constraint(nurse_assigned[n, s1] + nurse_assigned[n, s2] <= 1,
                                          "high_overlapping_{0!s}_{1!s}_{2!s}".format(s1, s2, n))
-    print('# overlapping cts: {0}'.format(number_of_overlaps))
+    #print('# overlapping cts: {0}'.format(number_of_overlaps))
 
     for s in all_shifts:
         demand_min = s.min_requirement
@@ -441,7 +443,7 @@ def setup_constraints(model):
 def setup_objective(model):
     model.add_kpi(model.total_salary_cost, "Total salary cost")
     model.add_kpi(model.total_number_of_assignments, "Total number of assignments")
-    model.add_kpi(model.average_nurse_work_time)
+    model.add_kpi(model.average_nurse_work_time, "average work time")
 
     total_over_average_worktime = model.sum(model.nurse_over_average_time_vars[n] for n in model.nurses)
     total_under_average_worktime = model.sum(model.nurse_under_average_time_vars[n] for n in model.nurses)

examples/mp/modeling/nurses.py

@@ -47,8 +47,8 @@ def build_production_problem(products, resources, consumptions, **kwargs):
     """
     mdl = Model(name='production', **kwargs)
     # --- decision variables ---
-    mdl.inside_vars = mdl.continuous_var_dict(products, name='inside')
-    mdl.outside_vars = mdl.continuous_var_dict(products, name='outside')
+    mdl.inside_vars  = mdl.continuous_var_dict(products, name=lambda p: 'inside_%s' % p[0])
+    mdl.outside_vars = mdl.continuous_var_dict(products, name=lambda p: 'outside_%s' % p[0])
 
     # --- constraints ---
     # demand satisfaction
@@ -60,7 +60,9 @@ def build_production_problem(products, resources, consumptions, **kwargs):
 
     # --- objective ---
     mdl.total_inside_cost = mdl.sum(mdl.inside_vars[p] * p[2] for p in products)
+    mdl.add_kpi(mdl.total_inside_cost, "inside cost")
     mdl.total_outside_cost = mdl.sum(mdl.outside_vars[p] * p[3] for p in products)
+    mdl.add_kpi(mdl.total_outside_cost, "outside cost")
     mdl.minimize(mdl.total_inside_cost + mdl.total_outside_cost)
     return mdl