80dbf0f703
- Add profile source modules: apply_profile, extraction, extraction_loader, field_extractor, match_eval - Add profiles CLI subcommand (profiles_cmd.rs) - Update all 9 built-in profile YAMLs (invoice, receipt, contract, scientific_paper, slide_deck, form, bank_statement, legal_filing, book_chapter) - Add 50 invoice fixture PDFs - Add 2 receipt fixture PDFs Part of: pdftract-3a310 (Phase 7.10 coordinator)
75 lines
1.8 KiB
YAML
75 lines
1.8 KiB
YAML
# Bank Statement extraction profile
|
|
# Matches bank statements with account info, period, balances, transactions
|
|
name: bank_statement
|
|
description: Bank statement with account info, period, balances, transactions
|
|
priority: 42
|
|
|
|
match:
|
|
all:
|
|
- any:
|
|
- text_contains:
|
|
patterns: ["statement of account", "bank statement", "account statement", "transaction history"]
|
|
- text_contains:
|
|
patterns: ["opening balance", "closing balance", "statement period"]
|
|
- structural:
|
|
has_table: true
|
|
has_form_field: false
|
|
has_math: false
|
|
page_count:
|
|
min: 1
|
|
max: 10
|
|
|
|
extraction:
|
|
reading_order: line_dominant
|
|
table_detection: default
|
|
readability_threshold: 0.5
|
|
include_invisible: false
|
|
include_headers_footers: false
|
|
force_ocr: false
|
|
min_block_chars: 0
|
|
|
|
fields:
|
|
account_number:
|
|
type: string
|
|
extraction:
|
|
regex: "account\\s*(?:number|#|no)?\\s*:?,?\\s*(\\*?\\d[\\d\\*]{3,})"
|
|
parse: string
|
|
|
|
statement_period:
|
|
type: string
|
|
extraction:
|
|
near: ["Statement Period", "Period"]
|
|
parse: string
|
|
|
|
opening_balance:
|
|
type: decimal
|
|
extraction:
|
|
near: ["Opening Balance", "Beginning Balance"]
|
|
regex: "([\\d,]+\\.\\d{2})"
|
|
parse: decimal
|
|
|
|
closing_balance:
|
|
type: decimal
|
|
extraction:
|
|
near: ["Closing Balance", "Ending Balance", "Current Balance"]
|
|
regex: "([\\d,]+\\.\\d{2})"
|
|
parse: decimal
|
|
|
|
transactions:
|
|
type: array
|
|
extraction:
|
|
table_region: largest_table
|
|
schema:
|
|
- name: date
|
|
type: date
|
|
required: true
|
|
- name: description
|
|
type: string
|
|
required: true
|
|
- name: amount
|
|
type: decimal
|
|
required: false
|
|
- name: balance
|
|
type: decimal
|
|
required: false
|
|
fallback: []
|