def get_rows(json_all):
json_tree = defaultdict(dict)
for json_f in json_all:
json_tree[json_f['end']][json_f['project']] = json_f['countries']
# log.debug('f: %s' % (json.dumps(json_all, indent=2)))
# expand tree structure of dictionaries into list of dicts with named fields
rows = list(flatten(json_tree, [], ['date', 'project', 'country', 'cohort', 'count']))
by_date = Nest().key(itemgetter('date')).map(rows)
# everything is by date, so everyone wants things sorted
by_date = OrderedDict(sorted(by_date.items()))
return by_date
def write_overall_datasource(projects, json_all, args):
log.info('writing overall datasource')
_id = 'overall'
name = 'Overall Editors by Language'
# build rows
keys = ['end', 'project', 'world']
json_tree = defaultdict(lambda : defaultdict(int))
for json_f in json_all:
json_tree[json_f['end']][json_f['project']] = json_f['world']
# expand cohorts
rows = list(flatten(json_tree, [], ['date', 'project', 'cohort', 'count']))
# group by date
by_date = Nest().key(itemgetter('date')).map(rows)
by_date = OrderedDict(sorted(by_date.items()))
csv_name = args.basename + '_' + _id + '.csv'
csv_path = os.path.join(args.datafile_dir, csv_name)
csv_file = open(csv_path, 'w')
# remove rows that don't interest us and then grab the row id (country-cohort) and count
csv_rows = []
for date, row_batch in by_date.items():
# TODO: need to be extracting the top level field 'world' (note the lowercase)
csv_row = {'date' : date}
for row in row_batch:
csv_row['%s (%s)' % (row['project'], row['cohort'])] = row['count']
csv_rows.append(csv_row)
# normalize fields
all_fields = sorted(reduce(set.__ior__, map(set,map(dict.keys, csv_rows)), set()))
all_fields.remove('date')
all_fields.insert(0,'date')
writer = csv.DictWriter(csv_file, all_fields, restval='', extrasaction='ignore')
writer.writeheader()
for csv_row in csv_rows:
writer.writerow(csv_row)
csv_file.close()
#def write_yaml(_id, name, fields, csv_name, rows, args):
return write_yaml(_id, name, all_fields, csv_name, by_date, args)
