model-csv makes it easy to treat csv, tsv and excel file.
class Book(models.Model):
title = models.CharField(max_length=100)
price = models.PositiveIntegerField(null=True, blank=True)
is_on_sale = models.BooleanField(default=True)
description = models.TextField()
created_at = models.DateTimeField(auto_now_add=True)
updated_at = models.DateTimeField(auto_now=True)
from model_csv.csv.django import DjangoCsv
class BookCsv(DjangoCsv):
class Meta:
model = Book
# if you declare fields then ModelCsv create columns automatically.
fields = '__all__'@dataclasses.dataclass
class Book:
title: str
price: int
is_on_sale: bool
description: str
from model_csv.csv.dclass import DataClassCsv
class BookCsv(DataClassCsv):
class Meta:
dclass = Book
fields = '__all__'>>> mcsv = BookCsv.for_write(instances=book_instances)
>>> mcsv.get_table(header=True)
[['title', 'price', 'is_on_sale', 'description'], ['Book title', '540', 'YES', 'description'], ...]
>>> with Path("book.csv").open("w") as f:
writer = csv.write(f)
write.writerows(mcv.get_table())
# only DjangoCsv
# choose file type. (CsvWriter, TsvWriter, XlsxWriter)
>>> from model_csv.writers import CsvWriter
>>> writer = CsvWriter(file_name='book')
>>> mcsv.get_response(writer=writer)>>> from model_csv.readers import CsvReader
>>> reader = CsvReader(file=file, table_starts_from=1)
>>> headers, *table = reader.get_table(header=True)
>>> headers
['title', 'price', 'is_on_sale', 'description']
>>> if headers != BookCsv._meta.get_headers():
>>> raise ValueError('Invalid header.')
>>> mcsv = BookCsv.for_read(table=table)
>>> if not mcsv.is_valid():
>>> raise ValueError('Invalid table values.')
>>> mcsv.get_instances()
[<Book: Book object (1)>, <Book: Book object (2)>, ...]
# Only DjangoCsv
>>> mcsv.bulk_create()# Download csv file in django view.
class BookDownloadView(generic.View):
def get(self, **kwargs):
mcsv = BookCsv.for_write(instance=Book.objects.all())
writer = CsvWriter(file_name='book.csv')
return mcsv.get_response(writer=writer)
# Upload csv file in django view.
class BookUploadView(generic.View):
def form_valid(self, form, **kwargs):
file = form.cleaned_data['file']
reader = CsvReader(file=file, table_starts_from=1)
headers, *table = reader.get_table()
if headers != BookCsv._meta.get_headers():
form.add_error('file', 'Invalid header.')
return self.form_invalid(form)
# e.g. [['Book title', '540', 'YES', 'description'],]
mcsv = BookCsv.for_read(table=table)
if mcsv.is_valid():
# create book model.
mcsv.bulk_create()
return redirect(self.get_success_url)
# handling errors in templates
...Column class can define header, an order of columns and a type of value. And You can define, validate and fix values in ModelCsv methods. Therefore, you don't have to write spaghetti code in django view anymore.
class BookDownloadCsv(generic.View):
def post(self, **kwargs):
header = [
'Book Title',
'Now On sale',
'R18',
'Price',
'Publish Date',
'Memo',
'Description',
'Csv Output Date',
# Suppose columns over 20 ...
]
body = [header]
today = timezone.now().date()
for book in Book.objects.all():
title = book.title.replace(F_WORD, '*', len(F_WORD))
if book.is_restricted_under_18:
title = '[R18]' + title
price = f'${book.price or 0}'
if book.published_at:
published_at = book.published_at.strftime('%Y-%m-%d')
else:
published_at = 'TBD'
title = '(TBD)' + title # values are sometimes fixed at several points.
price += ' (Unsettled)'
# Suppose each values need to be fixed depends on other values ...
body += [
title,
'YES' if book.is_on_sale else 'NO',
'YES' if book.is_restricted_under_18 else 'NO',
price,
published_at,
'',
'',
today,
# You have to check if orders of values and headers are same ...
]
return # create csv file and return response.Cool Solution with model-csv
from model_csv import columns
from model_csv.csv.django import DjangoCsv
class BookCsv(DjangoCsv):
title = columns.MethodColumn(header='Book Title')
is_on_sale = columns.AttributeColumn(header='Now On Sale', to=bool)
is_restricted_under_18 = columns.AttributeColumn(header='R18', to=bool)
price = columns.MethodColumn(header='Price')
published_at = columns.AttributeColumn(header='Publish Date')
memo = columns.StaticColumn(header='Memo', static_value='')
description = columns.StaticColumn(header='Description')
csv_output_date = columns.StaticColumn(header='Csv Output Date')
# even if columns over 20, it's not so difficult to manage an order of columns.
class Meta:
model = Book
read_mode = False
auto_assign = True
def column_title(self, instance: Book, **kwargs) -> str:
# Special methods having prefix `column_` are called when write down to a csv file.
title = instance.title.replace(F_WORD, '*' * len(F_WORD))
if instance.is_restricted_under_18:
title = '[R18] ' + title
if instance.is_restricted_under_18:
title = '(TBD)' + title
return title
def column_published_at(self, instance: Book, **kwargs) -> Union[datetime, stre]:
pbl_at = instance.published_at
return pbl_at.strftime('%Y-%m-%d') if pbl_at else 'TBD'
def column_price(self, instance: Book, **kwargs) -> str:
price = f'${instance.price or 0}'
if not instance.published_at:
price += ' (Unsettled)'
return price
class BookDownloadView(generic.View):
# It's easy to keep view logic simple.
def post(self, **kwargs):
mcsv = BookCsv.for_write(instances=Book.objects.all())
# you can insert values after construct ModelCsv.
mcsv.set_static_column('csv_output_date', timezone.now().date())
writer = CsvWriter(file_name='book.csv')
return mcsv.get_response(writer=writer)There are 3 types of columns and 1 decorator.
All these columns work in a same way when column instances read values from a csv file.
(just return a value from row by using an index)
They work differently when column instances write down values to a csv file.
Write down an attribute value
class BookCsv(DataClassCsv):
title = columns.AttributeColumn(header='Book Title', index=0)
class Meta:
dclass = Book
# or
class BookCsv(DjangoCsv):
title = columns.AttributeColumn(header='Book Title', index=0)
class Meta:
model = BookThen, AttributeColumn returns book.title when create csv file.
If set attr_name then AttributeColumn changes an attribute it refers to.
class BookCsv(DjangoCsv):
title = columns.AttributeColumn(header='Book Title', index=0)
official_title = columns.AttributeColumn(
header='Official Title', attr_name='title', index=1) # same as title
class Meta:
model = BookWrite down a value which is returned by a method
class BookCsv(DjangoCsv):
full_title = columns.MethodColumn(header='Title', index=0)
class Meta:
model = Book
def column_full_title(self, instance: Book, **kwargs) -> str:
return f'{instance.title} {instance.subtitle}'
def field_title(self, value: dict, **kwargs) -> str:
# value of index 0 is stored with key "full_title".
return value["full_title"].split()[0]
def field_subtitle(self, value: dict, **kwargs) -> str:
return value["full_title"].split()[1]If MethodColumn exists then ModelCsv search methods named column_<name> and write down the return values.
@as_column decorator returns a column which works like MethodColumn.
class BookCsv(ModelCsv):
@columns.as_column(header='Title', index=0)
def full_title(self, instance: Book, **kwargs) -> str:
return f'{instance.title} {instance.subtitle}'StaticColumn simply returns a static value.
class BookCsv(ModelCsv):
check = columns.StaticColumn(header='Check Box') # always returns ''
title = columns.AttributeColumn(header='Title')
...
phone_number = columns.StaticColumn(header='Phone Number',
static_value='+81 080-0000-0000')
class Meta:
model = Book
auto_assign = Trueclass BookCsv(DjangoCsv):
class Meta:
model = Book # django model only DjangoCsv
dclass = Book # dataclass only DataClassCsv
# automatically assign indexes to columns
# so you don't have to write `index=0` for each column.
auto_assign = True
# automatically convert values to the type of column if Column(to=<type>)
auto_convert = True
# prohibit to use as read or write mode.
read_mode = False # raise Error if .for_read is called
write_mode = True
# Convert datetime to str / str to datetime.
# Column(to=datetime) and auto_convert=True
datetime_format = '%Y-%m-%d %H:%M:%S'
date_format = '%Y-%m-%d'
tzinfo = timezone.utc
# convert bool to str.
# Column(to=bool) and auto_convert=True
show_true = 'yes' # if instance.value is true then write down 'yes'
show_false = 'no'
# understand values as bool if the value in list.
# Column(to=bool) and auto_convert=True
as_true = ['yes', 'Yes']
as_false = ['no', 'No']
# The value ModelCsv write down if the value is None
default_if_none = ''
# if indexes are not sequence, insert blank columns.
insert_blank_column = True