Ajay Javiya
31518bc09b
In method `user_has_groups`, make "Technical Features" effective in debug mode. Make the group "Employees" inherit "Technical Features". Make the group "Technical Features" invisible in the user form view. Remove useless `ir.rule` attached on group "Technical Features". Fix `test_acl` by avoiding the tricks around the group "Technical Features".
451 lines
18 KiB
Python
451 lines
18 KiB
Python
import csv
|
|
import datetime
|
|
import io
|
|
import itertools
|
|
import logging
|
|
import operator
|
|
import os
|
|
|
|
try:
|
|
from cStringIO import StringIO
|
|
except ImportError:
|
|
from StringIO import StringIO
|
|
|
|
import psycopg2
|
|
|
|
from openerp.osv import orm, fields
|
|
from openerp.tools.translate import _
|
|
from openerp.tools import DEFAULT_SERVER_DATE_FORMAT, \
|
|
DEFAULT_SERVER_DATETIME_FORMAT
|
|
|
|
FIELDS_RECURSION_LIMIT = 2
|
|
ERROR_PREVIEW_BYTES = 200
|
|
_logger = logging.getLogger(__name__)
|
|
|
|
try:
|
|
import xlrd
|
|
try:
|
|
from xlrd import xlsx
|
|
except ImportError:
|
|
xlsx = None
|
|
except ImportError:
|
|
xlrd = xlsx = None
|
|
|
|
try:
|
|
import odf_ods_reader
|
|
except ImportError:
|
|
odf_ods_reader = None
|
|
|
|
FILE_TYPE_DICT = {
|
|
'text/csv': ('csv', True, None),
|
|
'application/vnd.ms-excel': ('xls', xlrd, 'xlrd'),
|
|
'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet': ('xlsx', xlsx, 'xlrd >= 0.8'),
|
|
'application/vnd.oasis.opendocument.spreadsheet': ('ods', odf_ods_reader, 'odfpy')
|
|
}
|
|
EXTENSIONS = {
|
|
'.' + ext: handler
|
|
for mime, (ext, handler, req) in FILE_TYPE_DICT.iteritems()
|
|
}
|
|
|
|
class ir_import(orm.TransientModel):
|
|
_name = 'base_import.import'
|
|
# allow imports to survive for 12h in case user is slow
|
|
_transient_max_hours = 12.0
|
|
|
|
_columns = {
|
|
'res_model': fields.char('Model'),
|
|
'file': fields.binary(
|
|
'File', help="File to check and/or import, raw binary (not base64)"),
|
|
'file_name': fields.char('File Name'),
|
|
'file_type': fields.char(string='File Type'),
|
|
}
|
|
|
|
def get_fields(self, cr, uid, model, context=None,
|
|
depth=FIELDS_RECURSION_LIMIT):
|
|
""" Recursively get fields for the provided model (through
|
|
fields_get) and filter them according to importability
|
|
|
|
The output format is a list of ``Field``, with ``Field``
|
|
defined as:
|
|
|
|
.. class:: Field
|
|
|
|
.. attribute:: id (str)
|
|
|
|
A non-unique identifier for the field, used to compute
|
|
the span of the ``required`` attribute: if multiple
|
|
``required`` fields have the same id, only one of them
|
|
is necessary.
|
|
|
|
.. attribute:: name (str)
|
|
|
|
The field's logical (Odoo) name within the scope of
|
|
its parent.
|
|
|
|
.. attribute:: string (str)
|
|
|
|
The field's human-readable name (``@string``)
|
|
|
|
.. attribute:: required (bool)
|
|
|
|
Whether the field is marked as required in the
|
|
model. Clients must provide non-empty import values
|
|
for all required fields or the import will error out.
|
|
|
|
.. attribute:: fields (list(Field))
|
|
|
|
The current field's subfields. The database and
|
|
external identifiers for m2o and m2m fields; a
|
|
filtered and transformed fields_get for o2m fields (to
|
|
a variable depth defined by ``depth``).
|
|
|
|
Fields with no sub-fields will have an empty list of
|
|
sub-fields.
|
|
|
|
:param str model: name of the model to get fields form
|
|
:param int landing: depth of recursion into o2m fields
|
|
"""
|
|
model_obj = self.pool[model]
|
|
fields = [{
|
|
'id': 'id',
|
|
'name': 'id',
|
|
'string': _("External ID"),
|
|
'required': False,
|
|
'fields': [],
|
|
}]
|
|
fields_got = model_obj.fields_get(cr, uid, context=context)
|
|
blacklist = orm.MAGIC_COLUMNS + [model_obj.CONCURRENCY_CHECK_FIELD]
|
|
for name, field in fields_got.iteritems():
|
|
if name in blacklist:
|
|
continue
|
|
# an empty string means the field is deprecated, @deprecated must
|
|
# be absent or False to mean not-deprecated
|
|
if field.get('deprecated', False) is not False:
|
|
continue
|
|
if field.get('readonly'):
|
|
states = field.get('states')
|
|
if not states:
|
|
continue
|
|
# states = {state: [(attr, value), (attr2, value2)], state2:...}
|
|
if not any(attr == 'readonly' and value is False
|
|
for attr, value in itertools.chain.from_iterable(
|
|
states.itervalues())):
|
|
continue
|
|
|
|
f = {
|
|
'id': name,
|
|
'name': name,
|
|
'string': field['string'],
|
|
# Y U NO ALWAYS HAS REQUIRED
|
|
'required': bool(field.get('required')),
|
|
'fields': [],
|
|
}
|
|
|
|
if field['type'] in ('many2many', 'many2one'):
|
|
f['fields'] = [
|
|
dict(f, name='id', string=_("External ID")),
|
|
dict(f, name='.id', string=_("Database ID")),
|
|
]
|
|
elif field['type'] == 'one2many' and depth:
|
|
f['fields'] = self.get_fields(
|
|
cr, uid, field['relation'], context=context, depth=depth-1)
|
|
if self.user_has_groups(cr, uid, 'base.group_no_one'):
|
|
f['fields'].append({'id' : '.id', 'name': '.id', 'string': _("Database ID"), 'required': False, 'fields': []})
|
|
|
|
fields.append(f)
|
|
|
|
# TODO: cache on model?
|
|
return fields
|
|
|
|
def _read_file(self, file_type, record, options):
|
|
(file_extension, handler, req) = FILE_TYPE_DICT.get(file_type, (None, None, None))
|
|
if handler:
|
|
return getattr(self, '_read_' + file_extension)(record, options)
|
|
# fallback on file extensions as mime types can be unreliable (e.g.
|
|
# software setting incorrect mime types, or non-installed software
|
|
# leading to browser not sending mime types)
|
|
if record.file_name:
|
|
p, ext = os.path.splitext(record.file_name)
|
|
if ext and EXTENSIONS.get(ext):
|
|
return getattr(self, '_read_' + ext[1:])(record, options)
|
|
|
|
if req:
|
|
raise ImportError(_("Unable to load \"{extension}\" file: requires Python module \"{modname}\"").format(extension=file_extension, modname=req))
|
|
raise ValueError(_("Unsupported file format \"{}\", import only supports CSV, ODS, XLS and XLSX").format(file_type))
|
|
|
|
def _read_xls(self, record, options):
|
|
book = xlrd.open_workbook(file_contents=record.file)
|
|
sheet = book.sheet_by_index(0)
|
|
# emulate Sheet.get_rows for pre-0.9.4
|
|
for row in itertools.imap(sheet.row, range(sheet.nrows)):
|
|
values = []
|
|
for cell in row:
|
|
if cell.ctype is xlrd.XL_CELL_NUMBER:
|
|
is_float = cell.value % 1 != 0.0
|
|
values.append(
|
|
unicode(cell.value)
|
|
if is_float
|
|
else unicode(int(cell.value))
|
|
)
|
|
elif cell.ctype is xlrd.XL_CELL_DATE:
|
|
is_datetime = cell.value % 1 != 0.0
|
|
# emulate xldate_as_datetime for pre-0.9.3
|
|
dt = datetime.datetime(*xlrd.xldate.xldate_as_tuple(
|
|
cell.value, book.datemode))
|
|
values.append(
|
|
dt.strftime(DEFAULT_SERVER_DATETIME_FORMAT)
|
|
if is_datetime
|
|
else dt.strftime(DEFAULT_SERVER_DATE_FORMAT)
|
|
)
|
|
elif cell.ctype is xlrd.XL_CELL_BOOLEAN:
|
|
values.append(u'True' if cell.value else u'False')
|
|
elif cell.ctype is xlrd.XL_CELL_ERROR:
|
|
raise ValueError(
|
|
_("Error cell found while reading XLS/XLSX file: %s") %
|
|
xlrd.error_text_from_code.get(
|
|
cell.value, "unknown error code %s" % cell.value)
|
|
)
|
|
else:
|
|
values.append(cell.value)
|
|
if any(x for x in values if x.strip()):
|
|
yield values
|
|
_read_xlsx = _read_xls
|
|
|
|
def _read_ods(self, record, options):
|
|
doc = odf_ods_reader.ODSReader(file=io.BytesIO(record.file))
|
|
|
|
return (
|
|
row
|
|
for row in doc.getFirstSheet()
|
|
if any(x for x in row if x.strip())
|
|
)
|
|
|
|
def _read_csv(self, record, options):
|
|
""" Returns a CSV-parsed iterator of all empty lines in the file
|
|
|
|
:throws csv.Error: if an error is detected during CSV parsing
|
|
:throws UnicodeDecodeError: if ``options.encoding`` is incorrect
|
|
"""
|
|
csv_iterator = csv.reader(
|
|
StringIO(record.file),
|
|
quotechar=str(options['quoting']),
|
|
delimiter=str(options['separator']))
|
|
|
|
# TODO: guess encoding with chardet? Or https://github.com/aadsm/jschardet
|
|
encoding = options.get('encoding', 'utf-8')
|
|
return (
|
|
[item.decode(encoding) for item in row]
|
|
for row in csv_iterator
|
|
if any(x for x in row if x.strip())
|
|
)
|
|
|
|
def _match_header(self, header, fields, options):
|
|
""" Attempts to match a given header to a field of the
|
|
imported model.
|
|
|
|
:param str header: header name from the CSV file
|
|
:param fields:
|
|
:param dict options:
|
|
:returns: an empty list if the header couldn't be matched, or
|
|
all the fields to traverse
|
|
:rtype: list(Field)
|
|
"""
|
|
string_match = None
|
|
for field in fields:
|
|
# FIXME: should match all translations & original
|
|
# TODO: use string distance (levenshtein? hamming?)
|
|
if header.lower() == field['name'].lower():
|
|
return [field]
|
|
if header.lower() == field['string'].lower():
|
|
# matching string are not reliable way because
|
|
# strings have no unique constraint
|
|
string_match = field
|
|
if string_match:
|
|
# this behavior is only applied if there is no matching field['name']
|
|
return [string_match]
|
|
|
|
if '/' not in header:
|
|
return []
|
|
|
|
# relational field path
|
|
traversal = []
|
|
subfields = fields
|
|
# Iteratively dive into fields tree
|
|
for section in header.split('/'):
|
|
# Strip section in case spaces are added around '/' for
|
|
# readability of paths
|
|
match = self._match_header(section.strip(), subfields, options)
|
|
# Any match failure, exit
|
|
if not match: return []
|
|
# prep subfields for next iteration within match[0]
|
|
field = match[0]
|
|
subfields = field['fields']
|
|
traversal.append(field)
|
|
return traversal
|
|
|
|
def _match_headers(self, rows, fields, options):
|
|
""" Attempts to match the imported model's fields to the
|
|
titles of the parsed CSV file, if the file is supposed to have
|
|
headers.
|
|
|
|
Will consume the first line of the ``rows`` iterator.
|
|
|
|
Returns a pair of (None, None) if headers were not requested
|
|
or the list of headers and a dict mapping cell indices
|
|
to key paths in the ``fields`` tree
|
|
|
|
:param Iterator rows:
|
|
:param dict fields:
|
|
:param dict options:
|
|
:rtype: (None, None) | (list(str), dict(int: list(str)))
|
|
"""
|
|
if not options.get('headers'):
|
|
return None, None
|
|
|
|
headers = next(rows)
|
|
return headers, {
|
|
index: [field['name'] for field in self._match_header(header, fields, options)] or None
|
|
for index, header in enumerate(headers)
|
|
}
|
|
|
|
def parse_preview(self, cr, uid, id, options, count=10, context=None):
|
|
""" Generates a preview of the uploaded files, and performs
|
|
fields-matching between the import's file data and the model's
|
|
columns.
|
|
|
|
If the headers are not requested (not options.headers),
|
|
``matches`` and ``headers`` are both ``False``.
|
|
|
|
:param id: identifier of the import
|
|
:param int count: number of preview lines to generate
|
|
:param options: format-specific options.
|
|
CSV: {encoding, quoting, separator, headers}
|
|
:type options: {str, str, str, bool}
|
|
:returns: {fields, matches, headers, preview} | {error, preview}
|
|
:rtype: {dict(str: dict(...)), dict(int, list(str)), list(str), list(list(str))} | {str, str}
|
|
"""
|
|
(record,) = self.browse(cr, uid, [id], context=context)
|
|
fields = self.get_fields(cr, uid, record.res_model, context=context)
|
|
|
|
try:
|
|
rows = self._read_file(record.file_type, record, options)
|
|
|
|
headers, matches = self._match_headers(rows, fields, options)
|
|
# Match should have consumed the first row (iif headers), get
|
|
# the ``count`` next rows for preview
|
|
preview = list(itertools.islice(rows, count))
|
|
assert preview, "CSV file seems to have no content"
|
|
return {
|
|
'fields': fields,
|
|
'matches': matches or False,
|
|
'headers': headers or False,
|
|
'preview': preview,
|
|
}
|
|
except Exception, e:
|
|
# Due to lazy generators, UnicodeDecodeError (for
|
|
# instance) may only be raised when serializing the
|
|
# preview to a list in the return.
|
|
_logger.debug("Error during parsing preview", exc_info=True)
|
|
preview = None
|
|
if record.file_type == 'text/csv':
|
|
preview = record.file[:ERROR_PREVIEW_BYTES].decode('iso-8859-1')
|
|
return {
|
|
'error': str(e),
|
|
# iso-8859-1 ensures decoding will always succeed,
|
|
# even if it yields non-printable characters. This is
|
|
# in case of UnicodeDecodeError (or csv.Error
|
|
# compounded with UnicodeDecodeError)
|
|
'preview': preview,
|
|
}
|
|
|
|
def _convert_import_data(self, record, fields, options, context=None):
|
|
""" Extracts the input browse_record and fields list (with
|
|
``False``-y placeholders for fields to *not* import) into a
|
|
format Model.import_data can use: a fields list without holes
|
|
and the precisely matching data matrix
|
|
|
|
:param browse_record record:
|
|
:param list(str|bool): fields
|
|
:returns: (data, fields)
|
|
:rtype: (list(list(str)), list(str))
|
|
:raises ValueError: in case the import data could not be converted
|
|
"""
|
|
# Get indices for non-empty fields
|
|
indices = [index for index, field in enumerate(fields) if field]
|
|
if not indices:
|
|
raise ValueError(_("You must configure at least one field to import"))
|
|
# If only one index, itemgetter will return an atom rather
|
|
# than a 1-tuple
|
|
if len(indices) == 1: mapper = lambda row: [row[indices[0]]]
|
|
else: mapper = operator.itemgetter(*indices)
|
|
# Get only list of actually imported fields
|
|
import_fields = filter(None, fields)
|
|
|
|
rows_to_import = self._read_file(record.file_type, record, options)
|
|
if options.get('headers'):
|
|
rows_to_import = itertools.islice(
|
|
rows_to_import, 1, None)
|
|
data = [
|
|
row for row in itertools.imap(mapper, rows_to_import)
|
|
# don't try inserting completely empty rows (e.g. from
|
|
# filtering out o2m fields)
|
|
if any(row)
|
|
]
|
|
|
|
return data, import_fields
|
|
|
|
def do(self, cr, uid, id, fields, options, dryrun=False, context=None):
|
|
""" Actual execution of the import
|
|
|
|
:param fields: import mapping: maps each column to a field,
|
|
``False`` for the columns to ignore
|
|
:type fields: list(str|bool)
|
|
:param dict options:
|
|
:param bool dryrun: performs all import operations (and
|
|
validations) but rollbacks writes, allows
|
|
getting as much errors as possible without
|
|
the risk of clobbering the database.
|
|
:returns: A list of errors. If the list is empty the import
|
|
executed fully and correctly. If the list is
|
|
non-empty it contains dicts with 3 keys ``type`` the
|
|
type of error (``error|warning``); ``message`` the
|
|
error message associated with the error (a string)
|
|
and ``record`` the data which failed to import (or
|
|
``false`` if that data isn't available or provided)
|
|
:rtype: list({type, message, record})
|
|
"""
|
|
cr.execute('SAVEPOINT import')
|
|
|
|
(record,) = self.browse(cr, uid, [id], context=context)
|
|
try:
|
|
data, import_fields = self._convert_import_data(
|
|
record, fields, options, context=context)
|
|
except ValueError, e:
|
|
return [{
|
|
'type': 'error',
|
|
'message': unicode(e),
|
|
'record': False,
|
|
}]
|
|
|
|
_logger.info('importing %d rows...', len(data))
|
|
import_result = self.pool[record.res_model].load(
|
|
cr, uid, import_fields, data, context=context)
|
|
_logger.info('done')
|
|
|
|
# If transaction aborted, RELEASE SAVEPOINT is going to raise
|
|
# an InternalError (ROLLBACK should work, maybe). Ignore that.
|
|
# TODO: to handle multiple errors, create savepoint around
|
|
# write and release it in case of write error (after
|
|
# adding error to errors array) => can keep on trying to
|
|
# import stuff, and rollback at the end if there is any
|
|
# error in the results.
|
|
try:
|
|
if dryrun:
|
|
cr.execute('ROLLBACK TO SAVEPOINT import')
|
|
else:
|
|
cr.execute('RELEASE SAVEPOINT import')
|
|
except psycopg2.InternalError:
|
|
pass
|
|
|
|
return import_result['messages']
|