Skip to content

Commit

Permalink
misc templates of (common) dutch suppliers
Browse files Browse the repository at this point in the history
  • Loading branch information
bosd committed Nov 18, 2022
1 parent 73dcc23 commit 54db86a
Show file tree
Hide file tree
Showing 20 changed files with 962 additions and 9 deletions.
47 changes: 47 additions & 0 deletions src/invoice2data/extract/templates/nl/nl.action.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,47 @@
# -*- coding: utf-8 -*-
# SPDX-License-Identifier: MIT
issuer: action
fields:
amount:
parser: regex
regex:
- 'TOTAAL\s+(\d{1,3}[,]\d{2})'
- PIN\s+(\d{1,3}.\d{2})
type: float
date: (\d{1,2}[-|\/]\d{2}[-|\/]\d{4})
invoice_number: (\d{1,2}[-|\/]\d{2}[-|\/]\d{4})
static_partner_name: action
payment_method:
- (AMEX)
- (VISA)
- (MCARD)
- (CONTANT)
- (KAS):\s.\s\d+\.\d+
lines:
start: ARTIKELEN
end: BTW SPECIFICATIE
line: '.(?P<product>\w+).?\s+.?(?P<price_unit>\d+.\d{1,2})'
types:
qty: float
price_unit: float
keywords:
- action
- ARTIKELEN
-
options:
currency: EUR
date_formats:
- '%d-%m-%Y'
languages:
- nl
decimal_separator: ','
replace:
- ['é' ,'€']
- ['Contant' ,'CONTANT']
- ['\s[l|&|@]' ,' ℓ']
- ['B.W', 'BTW']
- [', ', ',']
- ['°', ' ']
- ['©', ' ']
- [',.', ',']
- ['(\s)([0-9]+)\.([0-9]{2})(\s)', '\1\2,\3\4'] # workaround for ocr
58 changes: 58 additions & 0 deletions src/invoice2data/extract/templates/nl/nl.albron.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,58 @@
# -*- coding: utf-8 -*-
# SPDX-License-Identifier: MIT
issuer: Albron
fields:
amount:
parser: regex
regex:
- Totaal\s+[:]\s+(\d{1,3}[.|,]\d{2})
- '\d{1,2}[,|.]\d{1,2}\S?\s+(\d{1,3}[,|.]\d{2})\s+\d{1,3}[,]\d{2}'
type: float
amount_tax:
parser: regex
regex:
- '\d{1,2}[,|.]\d{1,2}\S?\s+\d{1,3}[,|.]\d{2}\s+\d{1,3}[,|.]\d{2}\s+(\d{1,3}[,]\d{2})'
type: float
amount_untaxed:
parser: regex
regex:
- '\d{1,2}[,|.]\d{1,2}\S?\s+\d{1,3}[,|.]\d{2}\s+(\d{1,3}[,]\d{2})'
type: float
date: (\d{1,2}[-|\/]\d{2}[-|\/]\d{2})
invoice_number: (\d{1,2}[-|\/]\d{2}[-|\/]\d{2})
static_partner_name: Albron
payment_method:
- (AMEX)
- (VISA)
- (MCARD)
- (CONTANT)
- (CHIP)
- (KAS):\s.\s\d+\.\d+
lines:
start: antal
end: Subtotaal
line: '(?P<qty>\d+)(?P<product>\w+)\s+(?P<price_unit>\d{1,4}[,|.]\d{2})'
types:
qty: float
price_unit: float
keywords:
- Albron
- EUR
- eten
options:
currency: EUR
date_formats:
- '%d-%m-%Y'
languages:
- nl
decimal_separator: ','
replace:
- ['é' ,'€']
- ['Contant' ,'CONTANT']
- ['\s[l|&|@]' ,' ℓ']
- ['B.W', 'BTW']
- [', ', ',']
- ['°', ' ']
- ['©', ' ']
- [',.', ',']
- ['(\s)([0-9]+)\.([0-9]{2})(\s)', '\1\2,\3\4'] # workaround for ocr
60 changes: 60 additions & 0 deletions src/invoice2data/extract/templates/nl/nl.begra.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,60 @@
# -*- coding: utf-8 -*-
# SPDX-License-Identifier: MIT
issuer: Begra Magazijninrichting B.V.
fields:
amount: Totaal EUR incl[.] btw\s+(\d{1,4}[,]\d{2})
amount_tax: BTW\s+(\d{1,4}[,]\d{2})
amount_untaxed: Totaal EUR excl[.] btw\s+(\d{1,4}[,]\d{2})
date: Factuurdatum\s+(\d{1,2}[.]\s\w+\s\d{4})
invoice_number: Factuurnr[.]\s+(\S+)
static_vat: NL818038524B01
bic: SWIFT[\/]BIC[:]\s+(\w{8,11})
iban: [A-Z]{2}\d{2}?\w{4}?\d{4}?\d{4}?\d{0,2}
telephone:
parser: regex
regex: '[+]\d{11}'
partner_website:
parser: static
value: begra.nl
partner_email:
parser: static
value: info@begra.nl
partner_name:
parser: static
value: Begra Magazijninrichting B.V.
country_code:
parser: static
value: nl
partner_zip:
parser: regex
regex: 5347\sKG
partner_city:
parser: regex
regex: Oss
partner_street:
parser: static
value: Ijsselstraat 9
partner_coc:
parser: regex
regex: KvK[:]\s\d{8}
narration:
parser: regex
regex: Ordernr[.]\s+(\w+)
lines:
start: Product
end: Totaal[:]
first_line: (?P<code>\S{8,12})\s+(?P<name>(\w+(?:\s\S+){1,2}))\s+(?P<qty>\d+)Stuks\s+(?P<price_unit>\d+[,]\d{2})\s+(?P<price_subtotal>\d+[,]\d{2})
line: '(?P<name>(\w+(?:\s\S+){1,2}))'
types:
qty: float
price_unit: float
price_subtotal: float
keywords:
- Begra
- 'Factuur'
- NL818038524B01
options:
currency: EUR
languages:
- nl
decimal_separator: ','
79 changes: 79 additions & 0 deletions src/invoice2data/extract/templates/nl/nl.blokker.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,79 @@
# -*- coding: utf-8 -*-
# SPDX-License-Identifier: MIT
issuer: blokker
fields:
amount:
parser: regex
regex:
- Totaal [(]incl. BTW[)]\s+[€]\s=(\d{1,3}[.]\d{2})
type: float
amount_untaxed:
parser: regex
regex:
- Subtotaal\s+[€]\s+(\d{1,3}[.]\d{2})
type: float
group: sum
partner_coc:
parser: regex
regex:
- KvK\s+(\d{8})
telephone:
parser: regex
regex:
- WhatsApp. \s+(\d{2}[-]\d{8})
partner_website:
parser: regex
regex:
- (blokker.nl)
partner_name:
parser: static
value: Blokker B.V.
country_code:
parser: static
value: nl
partner_zip:
parser: static
value: 1114 AM
partner_city:
parser: static
value: Amsterdam-Duivendrecht
partner_street:
parser: static
value: Van der Madeweg 13
bic: BIC.(\w{8,11})
iban: IBAN.(\w{10,30})
date: Besteldatum\s+.?\s+[:]\s+(\d{1,2}[-]\d{1,2}[-]\d{2})
invoice_number: Factuurnummer\s+.?\s+(\d{7,10})
static_vat: NL854525154B01
payment_method:
- (iDEAL)
- (AMEX)
- (AMERICAN EXPRESS)
- (VISA)
- (MCARD)
- (CONTANT)
- (?i)(Paypal)
lines:
- start: Artikel
end: Totaal [(]incl. BTW[)]\s+[€]
first_line: (?P<barcode>\w+.*)\s+(?P<product>\w+.*)\s+(?P<qty>\d+)\s+(?P<price_unit>[€]\s+(\d+.\d{2}))\s+(?P<price_subtotal>[€]\s+(\d+.\d{2}))
line: (?P<product>\w+.*) # needs check
types:
qty: float
price_unit: float
price_subtotal: float
- start: Subtotaal
end: Totaal [(]incl. BTW[)]\s+[€]
line: (?P<name>(\w+(?:\s\S+)*))\s+(?P<price_unit>(\d+.\d+))?
keywords:
- blokker
- NL854525154B01
-
options:
currency: EUR
languages:
- nl
decimal_separator: '.'
replace:
- ['é' ,'€']

27 changes: 27 additions & 0 deletions src/invoice2data/extract/templates/nl/nl.bunq.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,27 @@
# -*- coding: utf-8 -*-
# SPDX-License-Identifier: MIT
issuer: bunq
fields:
amount: Totaal\s+[€]\s+(\d+.\d{2})
date: Datum.\s+(\d{4}-\d{2}-\d{2})
invoice_number: Nummer.\s+(\d{7})
static_vat: NL851519945B01
lines:
start: Bedrag
end: Totaal
last_line: '(?P<description>(\w+(?:\S|[ ]\w\w+)*))\s+(?P<qty>\d+)\s+(?P<price_unit>€\s\d+.\d{2})\s+(?P<amountcurrency>€\s\d+.\d{2})'
line: '^(?P<description2>\w+(?:\S|[ ]\w\w+){1})$' # te veel hits
first_line: '^(?P<title>(Subscription))$'
types:
qty: float
price_unit: float
keywords:
- bunq
- NL851519945B01
options:
date_formats:
- '%Y-%m-%d'
currency: EUR
languages:
- nl
decimal_separator: ','
74 changes: 74 additions & 0 deletions src/invoice2data/extract/templates/nl/nl.farnell.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,74 @@
# -*- coding: utf-8 -*-
# SPDX-License-Identifier: MIT
issuer: Farnell (Netherlands) B.V.
fields:
amount: Eindtotaal\s+EUR\s+(\d{1,4}[.]\d{2})
amount_tax: BTW\s+(\d{1,4}[.]\d{2})
amount_untaxed: Subtotaal\s+(\d{1,4}[.]\d{2})
date: Factuurdatum\s+(\d{2}\s+\w+\s+\d{4})
date_due: 'Niet Later Dan[:]\s+\S+\n(\d{2}\s\w+\s\d{4})'
invoice_number: Factuur-Nr.\s+(\d+)
static_vat: NL800391007B01
bic:
parser: regex
regex: 'BIC[:]\s+(\w{8,11})'
iban:
parser: regex
regex: '[A-Z]{2}\d{2}?\w{4}?\d{4}?\d{4}?\d{0,2}'
telephone:
parser: static
value: '0302417373'
partner_website:
parser: static
value: nl.farnell.com
partner_email:
parser: regex
regex: \w+[@]farnell.com
partner_name:
parser: static
value: Farnell (Netherlands) B.V.
country_code:
parser: static
value: nl
partner_zip:
parser: static
value: 3600 BV
partner_city:
parser: regex
regex: Maarssen
partner_street:
parser: static
value: Postbus 1815
partner_coc:
parser: regex
regex: Handelsregister Utrecht nr.(\S{10})
narration:
parser: regex
regex: 'Uw Ordernummer\s+(\S+)'
note:
parser: regex
regex: '(Order geplaatst door \S+(?:\s+\S+)*)\n\s+[*]'
lines:
start: 'Lijn Nr'
end: BELANGRIJK
first_line: '\d+\s+(?P<code>\d{7})\s+(?P<uom>\w+)\s+(?P<qty>\d+)\s+(?P<price_unit>\d+[.]\d{2,4})\s+(?P<netto_price>\d+[.]\d{2,4})\s+(?P<btw_percent>\d+[.]\d{2})\s+(?P<price_subtotal>\d+[.]\d{2})'
line: '^\s{9,11}(?P<name>(\S+(?:\s\S+)*))\s+'
last_line: '\s+(?P<name>(Tariff Code[:]\s+\d+))'
types:
qty: float
price_unit: float
price_subtotal: float
netto_price: float
price_subtotal: float
btw_percent: float
keywords:
- Farnell
- 'Factuur'
- NL800391007B01
options:
currency: EUR
languages:
- nl
decimal_separator: '.'
replace:
- ['NL31 BOFA 0266 5038 53' ,'NL31BOFA0266503853']
47 changes: 47 additions & 0 deletions src/invoice2data/extract/templates/nl/nl.ferbox.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,47 @@
# -*- coding: utf-8 -*-
# SPDX-License-Identifier: MIT
issuer: FERBOX
fields:
amount:
parser: regex
regex:
- Totaal\s+[€](\d?.?\d{1,3}[.]\d{2})
type: float
amount_untaxed:
parser: regex
regex:
- Subtotaal\s+[€](\d?[,]?\d{1,4}.\d{2})
type: float
group: sum
bic: BIC.(\w{8,11})
iban: IBAN.(\w{10,30})
date: Besteldatum.?\s+(\d{1,2} \w+ \d{4})
invoice_number: Bestelnummer.?\s+(\d{4})
static_vat: NL861141350B01
payment_method:
- (iDEAL)
- (AMEX)
- (VISA)
- (MCARD)
- (CONTANT)
- (?i)(Paypal)
lines:
start: Product
end: Subtotaal
first_line: (?P<product>\w+.*)\s+(?P<qty>\d+)\s+(?P<price_unit>[€]\s+(\d+[.]\d{2}))
line: (?P<product>\w+.*)
types:
qty: float
price_unit: float
keywords:
- FERBOX
- NL861141350B01
-
options:
currency: EUR
languages:
- nl
decimal_separator: '.'
replace:
- ['é' ,'€']

Loading

0 comments on commit 54db86a

Please sign in to comment.