s464924/csi_rekrutacja
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Added regex for postal-codes

Browse Source
This commit is contained in:
Zuzanna Rachuba 2024-11-03 12:15:47 +01:00
parent c32112799b
commit 364798ca52
1 changed files with 8 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
projekt.py
View File

@ -2,9 +2,7 @@ import fitz
import re import re
my_path = r"C:\Users\DELL\Downloads\A-24VU-00511.PDF.pdf" my_path = r"C:\Users\DELL\Downloads\A-24VU-00511.PDF.pdf"
postal_code_pattern = r"(?<!\S)(?:(?:[A-Za-z]{3,}\s+)(\d{5}|\d{2}-\d{3})|(\d{5}|\d{2}-\d{3})(?:\s+[A-Za-z]{3,}))(?!\S)"
postal_code_pattern = r'\b(?!0)([1-9]\d?-\d{3}|[1-9]\d{4}|[1-9]\d? \d{3})\b'
with fitz.open(my_path) as doc: with fitz.open(my_path) as doc:
order_numbers = [] order_numbers = []
order_dates = [] order_dates = []
@ -34,14 +32,12 @@ with fitz.open(my_path) as doc:
line_above = lines[i - 1] if i > 0 else '' line_above = lines[i - 1] if i > 0 else ''
line_below = lines[i + 1] if i + 1 < len(lines) else '' line_below = lines[i + 1] if i + 1 < len(lines) else ''
buyers.append({
if not re.search(r'[^0-9a-zA-Z \-]', line): 'line_2': line_2,
buyers.append({ 'line_above': line_above,
'line_2': line_2, 'postal_code_line': line,
'line_above': line_above, 'line_below': line_below
'postal_code_line': line, })
'line_below': line_below
})
print("Reference numbers:", order_numbers) print("Reference numbers:", order_numbers)
print("Document dates:", order_dates) print("Document dates:", order_dates)