def _y_axis_garbage_before_type(M, data):
i = 0
garbage = ""
data[0].append("name")
while i < len(M):
y_row = M[i][0]
i = i + 1 # XXX
if _is_garbage(data[i]):
garbage = y_row
continue
# do not add " " if crap is empty
value = garbage
if value:
value += " "
value += y_row
value = slugify(value)
data[i].append(value)
def _y_axis_type_before_garbage(M, data):
i = 0
last = -1
data[0].append("name")
data[0].append("description")
while i < len(M):
y_row = M[i][0]
i = i + 1 # XXX
if not _is_garbage(data[i]):
y_row = slugify(y_row)
data[i].append(y_row)
data[i].append("")
last = i
continue
if last == -1:
continue
data[last][-1] += y_row
def jsonize(data, fp, indent=None):
headers = data[0]
body = data[1:]
for row in body:
package = CKANPackage()
#
# The algorithm here matches loosely the one that has been
# implemented in <ckanload-italy-nexa>.
#
for j in range(0, len(row)):
cell = row[j]
header = headers[j]
if (header == "datasource" or header == "istituzione" or
header == "author"):
package.author = cell
continue
if header == "name":
package.name = cell
continue
if header == "url":
package.url = cell
continue
if (header == "tipologia di dati" or
header == "diritti sul database"):
package.notes.append(cell)
continue
if header == "title":
package.title = cell
continue
if header == "tag":
package.tags.append(cell.lower().replace(" ", "-"))
continue
if cell:
package.extras[header] = cell
# XXX
package.notes = "\n\n".join(package.notes)
#
# As suggested by steko, the machine readable name must
# be prepended with a slugified version of the name of the
# dataset author.
# While on that, make sure the author name is not all-
# uppercase because that looks ugly.
# Ensure that the package name is not too long or the server
# will have a boo.
#
name = slugify(package.author.lower())
if not package.name.startswith(name):
package.name = name + "_" + package.name
package.name = package.name[:100]
#
# AFAIK vars() here will work as long as all the variables of
# `package` have been initialized using __init__(). This is
# what the code above already does. Nonetheless I whish to add
# this comment for future robusteness of the code.
#
octets = json.dumps(vars(package), indent=indent)
fp.write(octets)
fp.write("\n")
