udst / urbansim_defaults Goto Github PK

this check here could maybe work better: https://github.com/synthicity/urbansim_defaults/blob/ec68a0b2fff79c66f00bc63c8232e55c580cce41/urbansim_defaults/utils.py#L78-L81

at line 78, it seems like it would print out which data frame has NA's. but i don't think it actually does. (see the output below). i ran into this while having a config file with NA's as parameters (as output by Estimation.ipynb)

it looks like this (at least in a notebook, not sure about python):

AssertionError Traceback (most recent call last)
in ()
27 #"pusher", # push results to the web client
28 #"clear_cache" # clear the cache each year
---> 29 ], years=range(in_year, out_year))
30 print "Finished", time.ctime()

d:\urbansim_windows\urbansim\urbansim\sim\simulation.pyc in run(models, years, data_out, out_interval)
1524 model = get_model(model_name)
1525 t2 = time.time()
-> 1526 model()
1527 print("Time to execute model '{}': {:.2f}s".format(
1528 model_name, time.time()-t2))

d:\urbansim_windows\urbansim\urbansim\sim\simulation.pyc in call(self)
679 kwargs = _collect_variables(names=self._argspec.args,
680 expressions=self._argspec.defaults)
--> 681 return self._func(**kwargs)
682
683 def _tables_used(self):

d:\urbansim_windows\urbansim_defaults\urbansim_defaults\models.pyc in hlcm_simulate(households, buildings, aggregations, settings)
46 "building_id","residential_units",
47 "vacant_residential_units",
---> 48 settings.get("enable_supply_correction", None))
49
50

d:\urbansim_windows\urbansim_defaults\urbansim_defaults\utils.pyc in lcm_simulate(cfg, choosers, buildings, join_tbls, out_fname, supply_fname, vacant_fname, enable_supply_correction)
337 additional_columns +=[enable_supply_correction["price_col"]]
338 locations_df = to_frame(buildings, join_tbls, cfg,
--> 339 additional_columns=additional_columns)
340
341 available_units = buildings[supply_fname]

d:\urbansim_windows\urbansim_defaults\urbansim_defaults\utils.pyc in to_frame(tbl, join_tbls, cfg, additional_columns)
193 else:
194 df = tables[0].to_frame(columns)
--> 195 check_nas(df)
196 return df
197

d:\urbansim_windows\urbansim_defaults\urbansim_defaults\utils.pyc in check_nas(df)
79 (df_cnt-s_cnt, df_cnt, col)
80
---> 81 assertnot fail, "NAs were found in dataframe, please fix"
82
83

I believe the value count results will be sorted based on frequency (descending) so I don't see how this would align with the buildings. Seems like you will need to reindex the count results before the subtract?

@orca.column('buildings', 'vacant_residential_units')
def vacant_residential_units(buildings, households):
    return buildings.residential_units.sub(
        households.building_id.value_counts(), fill_value=0)


@orca.column('buildings', 'vacant_job_spaces')
def vacant_job_spaces(buildings, jobs):
    return buildings.job_spaces.sub(
        jobs.building_id.value_counts(), fill_value=0)

Pass in params in utils.py rather than storing and accessing data using the simulation framework

/urbansim_defaults/urbansim_defaults/utils.py:222: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.
  model_type = yaml.load(open(cfg))["model_type"]

Might be changed to yaml.load(open(cfg), Loader=yaml.FullLoader)["model_type"]