beancount-chile 0.7.0

Beancount importers for Chilean banks

beancount-chile

Beancount importers for Chilean banks using the beangulp framework.

This project provides importers for various Chilean bank account statement formats, enabling automatic import of transactions into Beancount for double-entry bookkeeping.

Supported Banks and Formats

Bank	Format	Status	File Extension
Banco de Chile	Cartola (Account Statement)	✅ Supported	.xls, .xlsx, .pdf
Banco de Chile	Credit Card Statements (Facturado/No Facturado)	✅ Supported	.xls, .xlsx

Installation

Prerequisites

• Python 3.10 or higher
• Beancount 3.x

Install

pip install beancount-chile

Usage

Banco de Chile Importer

The Banco de Chile importer supports XLS/XLSX/PDF account statement files (cartola).

Basic Usage

Create a configuration file (e.g., import_config.py):

from beancount_chile import BancoChileImporter, BancoChileCreditImporter

CONFIG = [
    # Checking account
    BancoChileImporter(
        account_number="00-123-45678-90",  # Your account number
        account_name="Assets:BancoChile:Checking",
        currency="CLP",
    ),
    # Credit card
    BancoChileCreditImporter(
        card_last_four="1234",  # Last 4 digits of your card
        account_name="Liabilities:CreditCard:BancoChile",
        currency="CLP",
    ),
]

Import Transactions

Use beangulp to extract transactions:

# Identify which importers can handle your files
bean-extract import_config.py ~/Downloads/

# Extract transactions from a specific file (XLS, XLSX, or PDF)
bean-extract import_config.py ~/Downloads/cartola.xls
bean-extract import_config.py ~/Downloads/cartola.pdf

# Extract and append to your beancount file
bean-extract import_config.py ~/Downloads/cartola.pdf >> accounts.beancount

Note: The importer automatically detects the file format (XLS/XLSX/PDF) and uses the appropriate parser. PDF files are parsed using text extraction, which handles the same transaction types as XLS files.

Example Output

The importer will generate Beancount entries like:

2026-01-01 * "Supermercado Santa Isabel" "Supermercado Santa Isabel"
  channel: "Internet"
  Assets:BancoChile:Checking  -45000 CLP

2026-01-03 * "María González" "Traspaso A:María González"
  channel: "Internet"
  Assets:BancoChile:Checking  -125000 CLP

2026-01-05 balance Assets:BancoChile:Checking  1230000 CLP

Banco de Chile Credit Card Importer

The credit card importer supports both billed (Facturado) and unbilled (No Facturado) statements.

Basic Usage

The credit card importer automatically detects whether a file contains billed or unbilled transactions:

from beancount_chile import BancoChileCreditImporter

CONFIG = [
    BancoChileCreditImporter(
        card_last_four="1234",  # Last 4 digits of your card
        account_name="Liabilities:CreditCard:BancoChile",
        currency="CLP",
    ),
]

Example Output

Billed Transactions (Mov_Facturado.xls):

2026-01-08 note Liabilities:CreditCard:BancoChile "Credit Card Statement - FACTURADO (Billed) | Total Billed: $850,000 CLP | Minimum Payment: $42,500 CLP | Due Date: 2026-01-21"

2026-01-02 * "SUPERMERCADO JUMBO" "SUPERMERCADO JUMBO SANTIAGO"
  statement_type: "facturado"
  category: "Total de Pagos, Compras, Cuotas y Avance"
  installments: "01/01"
  Liabilities:CreditCard:BancoChile  75000 CLP

Unbilled Transactions (Saldo_y_Mov_No_Facturado.xls):

2026-01-16 note Liabilities:CreditCard:BancoChile "Credit Card Statement - NO FACTURADO (Unbilled) | Available Credit: $6,500,000 CLP | Total Limit: $7,000,000 CLP"

2026-01-16 ! "NETFLIX.COM" "NETFLIX.COM COMPRAS"
  statement_type: "no_facturado"
  city: "LAS CONDES"
  installments: "01/01"
  Liabilities:CreditCard:BancoChile  12000 CLP

Note: Billed transactions are marked as cleared (*) while unbilled transactions are marked as pending (!).

Features

• Automatic payee extraction: Extracts payee names from transaction descriptions
• Balance assertions: Adds balance assertions to verify account balances
• Metadata tracking: Preserves channel information (Internet, Sucursal, etc.)
• Deduplication support: Works with beangulp's existing entry detection
• Custom categorization: Optional categorizer function for automatic transaction categorization

Advanced Features

Custom Categorization

Both importers support an optional categorizer parameter that allows you to automatically categorize transactions by providing a custom function. This is more flexible than pattern matching approaches as it gives you full control over the categorization logic.

Categorizer Function Signature

def categorizer(date, payee, narration, amount, metadata):
    """
    Categorize a transaction and optionally override payee/narration/subaccount.

    Args:
        date: Transaction date (datetime.date)
        payee: Extracted payee name (str)
        narration: Transaction description (str)
        amount: Transaction amount (Decimal, negative for debits)
        metadata: Dict with transaction-specific metadata

    Returns:
        Dict with optional fields (all fields are optional):
        - category: str - Single category account
        - payee: str - Override transaction payee
        - narration: str - Override transaction narration
        - subaccount: str - Subaccount suffix for main account
        - postings: List[Dict] - For splits, each with 'category' and 'amount' keys

        Or None for no categorization
    """
    # Your categorization logic here
    return {"category": "Expenses:Category"} or None

Metadata Available

For Checking Account (BancoChileImporter):

• channel: Transaction channel (e.g., "Internet", "Sucursal")
• debit: Debit amount (Decimal or None)
• credit: Credit amount (Decimal or None)
• balance: Account balance after transaction (Decimal)

For Credit Card (BancoChileCreditImporter):

• statement_type: "facturado" or "no_facturado"
• installments: Installment info (e.g., "01/12" or None)
• category: Transaction category for billed statements
• city: Transaction city for unbilled statements
• card_type: Card information for unbilled statements

Example: Simple Pattern Matching

def my_categorizer(date, payee, narration, amount, metadata):
    """Categorize based on payee name patterns."""
    # Internet services
    if "Starlink" in payee or "STARLINK" in narration:
        return {"category": "Expenses:Internet"}

    # Transportation
    if any(keyword in narration.upper() for keyword in ["UBER", "CABIFY", "TAXI"]):
        return {"category": "Expenses:Transportation"}

    # Groceries
    if any(store in payee.upper() for store in ["JUMBO", "UNIMARC", "SANTA ISABEL"]):
        return {"category": "Expenses:Groceries"}

    # Don't categorize this transaction
    return None

CONFIG = [
    BancoChileImporter(
        account_number="00-123-45678-90",
        account_name="Assets:BancoChile:Checking",
        currency="CLP",
        categorizer=my_categorizer,
    ),
]

Example: Amount-Based Categorization

def amount_based_categorizer(date, payee, narration, amount, metadata):
    """Categorize based on transaction amount."""
    # Large debits might be rent or major expenses
    if amount < -500000:  # More than 500k CLP debit
        return {"category": "Expenses:Major"}

    # Credits are income
    if amount > 0:
        return {"category": "Income:Salary"}

    return None

Example: Metadata-Based Categorization

def metadata_categorizer(date, payee, narration, amount, metadata):
    """Categorize based on transaction metadata."""
    # Only categorize online transactions
    if metadata.get("channel") == "Internet":
        if amount < 0:  # Debit
            return {"category": "Expenses:Online"}

    # For credit cards, use statement type
    if metadata.get("statement_type") == "facturado":
        # Already billed transactions
        return {"category": "Expenses:CreditCard"}

    return None

Example: Different Categorizers for Different Accounts

def checking_categorizer(date, payee, narration, amount, metadata):
    """Categorizer for checking account."""
    if "Starlink" in payee:
        return {"category": "Expenses:Internet"}
    return None

def credit_card_categorizer(date, payee, narration, amount, metadata):
    """Categorizer for credit card."""
    if "NETFLIX" in payee.upper():
        return {"category": "Expenses:Streaming"}
    if metadata.get("city") == "LAS CONDES":
        return {"category": "Expenses:Shopping"}
    return None

CONFIG = [
    BancoChileImporter(
        account_number="00-123-45678-90",
        account_name="Assets:BancoChile:Checking",
        currency="CLP",
        categorizer=checking_categorizer,
    ),
    BancoChileCreditImporter(
        card_last_four="1234",
        account_name="Liabilities:CreditCard:BancoChile",
        currency="CLP",
        categorizer=credit_card_categorizer,
    ),
]

Example: Transaction Splitting

The categorizer can return a dict with postings to split one transaction into multiple postings with fixed amounts. This is useful for:

• Splitting shared expenses across multiple categories
• Allocating specific amounts to different accounts
• Handling transactions with multiple components

from decimal import Decimal

def split_categorizer(date, payee, narration, amount, metadata):
    """Split transactions into multiple fixed-amount postings."""
    # Split grocery shopping between food and household supplies
    if "JUMBO" in payee.upper() or "UNIMARC" in payee.upper():
        # For checking account, debits are negative
        if amount < 0:  # Debit
            return {
                "postings": [
                    {"category": "Expenses:Groceries", "amount": Decimal("40000")},
                    {"category": "Expenses:Household", "amount": Decimal("10000")},
                ]
            }

    # Split pharmacy purchase between medicine and personal care
    if "PHARMACY" in payee.upper():
        if amount < 0:
            return {
                "postings": [
                    {"category": "Expenses:Health:Medicine", "amount": Decimal("25000")},
                    {"category": "Expenses:Health:Personal", "amount": Decimal("8000")},
                ]
            }

    # For credit cards, amounts are positive (increase liability)
    # Split subscription service between personal and family
    if "NETFLIX" in payee.upper():
        return {
            "postings": [
                {"category": "Expenses:Streaming:Personal", "amount": Decimal("-6000")},
                {"category": "Expenses:Streaming:Family", "amount": Decimal("-6000")},
            ]
        }

    # No split needed
    return None

CONFIG = [
    BancoChileImporter(
        account_number="00-123-45678-90",
        account_name="Assets:BancoChile:Checking",
        currency="CLP",
        categorizer=split_categorizer,
    ),
]

Important Notes on Transaction Splitting:

1. Amount Signs:

   • For checking accounts: debits are negative, so split amounts should be positive
   • For credit cards: charges are positive, so split amounts should be negative
   • The split amounts balance the original transaction

2. Balance: The sum of split amounts should balance the original transaction amount. Beancount will flag unbalanced transactions.

3. Flexible Splits: You can split into any number of postings:

# Split into 3 categories with fixed amounts
return {
    "postings": [
        {"category": "Expenses:Groceries", "amount": Decimal("30000")},
        {"category": "Expenses:Household", "amount": Decimal("15000")},
        {"category": "Expenses:Personal", "amount": Decimal("5000")},
    ]
}

4. Mixed Amounts: You can combine fixed amounts with calculated remainders:

# Allocate fixed amount to one category, rest to another
if "PHARMACY" in payee.upper():
    medicine_amount = Decimal("15000")
    return {
        "postings": [
            {"category": "Expenses:Health:Medicine", "amount": medicine_amount},
            {"category": "Expenses:Health:Personal", "amount": -amount - medicine_amount},  # Remainder
        ]
    }

Example Output with Categorizer

Single Posting (String Return):

When a categorizer returns a single account string, transactions will have two postings:

2026-01-01 * "Starlink" "Pago:starlink Starlink"
  channel: "Internet"
  Assets:BancoChile:Checking  -48000 CLP
  Expenses:Internet            48000 CLP

2026-01-16 ! "NETFLIX.COM" "NETFLIX.COM COMPRAS"
  statement_type: "no_facturado"
  city: "LAS CONDES"
  installments: "01/01"
  Liabilities:CreditCard:BancoChile  12000 CLP
  Expenses:Streaming                -12000 CLP

Split Postings (List Return):

When a categorizer returns a list of (account, amount) tuples, transactions can have multiple category postings:

2026-01-15 * "Jumbo" "Supermercado Jumbo Santiago"
  channel: "Internet"
  Assets:BancoChile:Checking  -50000 CLP
  Expenses:Groceries           40000 CLP
  Expenses:Household           10000 CLP

2026-01-20 * "Restaurant Central" "Restaurant Central Santiago"
  channel: "Internet"
  Assets:BancoChile:Checking  -35000 CLP
  Expenses:Food:Restaurant     24500 CLP
  Expenses:Entertainment       10500 CLP

2026-01-16 ! "NETFLIX.COM" "NETFLIX.COM COMPRAS"
  statement_type: "no_facturado"
  city: "LAS CONDES"
  installments: "01/01"
  Liabilities:CreditCard:BancoChile  12000 CLP
  Expenses:Streaming:Personal        -6000 CLP
  Expenses:Streaming:Family          -6000 CLP

Best Practices

1. Start Simple: Begin with a few high-frequency patterns and expand over time
2. Return None: Always return None for transactions you don't want to categorize automatically
3. Case Insensitive: Use .upper() or .lower() for pattern matching to handle case variations
4. Test Thoroughly: Review categorized transactions to ensure accuracy
5. Use Metadata: Leverage the metadata dict for more precise categorization rules
6. Combine Strategies: Mix pattern matching, amount-based, and metadata-based logic as needed

Virtual Subaccounts (NEW in v0.6.0)

The categorizer can specify subaccounts along with other fields using the dict-based API. This is perfect for envelope budgeting or tracking earmarked funds.

Subaccount Examples

1. Subaccount + Category

def my_categorizer(date, payee, narration, amount, metadata):
    """Return dict with subaccount and category"""
    if "SHELL" in payee.upper() or "COPEC" in payee.upper():
        return {"subaccount": "Car", "category": "Expenses:Car:Gas"}
    return None

# Creates: Assets:BancoChile:Checking:Car -> Expenses:Car:Gas

2. Subaccount + Split Postings

def my_categorizer(date, payee, narration, amount, metadata):
    """Return dict with subaccount and split postings"""
    if "JUMBO" in payee.upper():
        return {
            "subaccount": "Household",
            "postings": [
                {"category": "Expenses:Groceries", "amount": Decimal("40000")},
                {"category": "Expenses:Household", "amount": Decimal("10000")},
            ]
        }
    return None

# Creates: Assets:BancoChile:Checking:Household -> multiple expense accounts

3. Subaccount Only (No Category)

def my_categorizer(date, payee, narration, amount, metadata):
    """Return dict with subaccount only"""
    # Large deposits go to emergency fund, but don't categorize
    if amount > 500000:
        return {"subaccount": "Emergency"}
    return None

# Creates: Assets:BancoChile:Checking:Emergency (single posting, no category)

4. Payee & Narration Overrides

def my_categorizer(date, payee, narration, amount, metadata):
    """Return dict with payee and narration overrides"""
    if "CGE" in payee.upper():
        return {
            "payee": "CGE",
            "narration": "Electricity bill",
            "category": "Expenses:Utilities:Electricity"
        }
    return None

# Overrides payee and narration in the transaction

Example: Combining All Features

from decimal import Decimal

def my_categorizer(date, payee, narration, amount, metadata):
    """Single function with all capabilities!"""
    # Subaccount + category
    if "SHELL" in payee.upper() or "COPEC" in payee.upper():
        return {"subaccount": "Car", "category": "Expenses:Car:Gas"}

    # Subaccount + split postings
    if "JUMBO" in payee.upper():
        return {
            "subaccount": "Household",
            "postings": [
                {"category": "Expenses:Groceries", "amount": Decimal("40000")},
                {"category": "Expenses:Household", "amount": Decimal("10000")},
            ]
        }

    # Subaccount only (no category)
    if amount > 500000:
        return {"subaccount": "Emergency"}

    # Just category (no subaccount)
    if "NETFLIX" in payee.upper():
        return {"category": "Expenses:Streaming"}

    # Payee override
    if "CGE" in narration.upper():
        return {
            "payee": "CGE",
            "narration": "Monthly electricity",
            "category": "Expenses:Utilities:Electricity"
        }

    return None  # No categorization

CONFIG = [
    BancoChileImporter(
        account_number="00-123-45678-90",
        account_name="Assets:BancoChile:Checking",
        currency="CLP",
        categorizer=my_categorizer,
    ),
]

Example Output

Subaccount + Category:

2026-01-05 * "Shell" "Shell Costanera"
  channel: "Internet"
  Assets:BancoChile:Checking:Car  -45000 CLP
  Expenses:Car:Gas                 45000 CLP

Subaccount + Split Postings:

2026-01-08 * "Jumbo" "Supermercado Jumbo"
  channel: "Internet"
  Assets:BancoChile:Checking:Household  -50000 CLP
  Expenses:Groceries                     40000 CLP
  Expenses:Household                     10000 CLP

Subaccount Only:

2026-01-10 * "Salary Deposit" "Monthly Salary"
  channel: "Internet"
  Assets:BancoChile:Checking:Emergency  600000 CLP

Use Cases

• Envelope Budgeting: Track money earmarked for different purposes (Car, Vacation, Emergency)
• Savings Goals: Separate virtual accounts for different savings objectives
• Business/Personal Separation: Split credit card expenses by purpose
• Budget Categories: Map transactions to budget envelopes automatically

Development

Running Tests

# Activate virtual environment
source venv/bin/activate

# Run all tests
pytest

# Run with coverage
pytest --cov=beancount_chile

# Run specific test file
pytest tests/test_banco_chile.py -v

Code Quality

# Format code with ruff
ruff format .

# Lint code
ruff check .

# Fix auto-fixable issues
ruff check --fix .

Project Structure

beancount-chile/
├── beancount_chile/                   # Main package
│   ├── __init__.py
│   ├── banco_chile.py                 # Checking account importer
│   ├── banco_chile_credit.py          # Credit card importer
│   ├── helpers.py                     # Shared utilities
│   └── extractors/                    # File format parsers
│       ├── __init__.py
│       ├── banco_chile_xls.py         # Checking account XLS parser
│       ├── banco_chile_pdf.py         # Checking account PDF parser
│       └── banco_chile_credit_xls.py  # Credit card parser
├── tests/                             # Test suite
│   ├── __init__.py
│   ├── test_banco_chile.py            # Checking account tests
│   ├── test_banco_chile_credit.py     # Credit card tests
│   ├── test_banco_chile_pdf.py        # PDF parser tests
│   └── fixtures/                      # Test data (anonymized)
│       ├── banco_chile_cartola_sample.xls
│       ├── banco_chile_credit_facturado_sample.xls
│       └── banco_chile_credit_no_facturado_sample.xls
├── pyproject.toml
├── requirements.txt
└── README.md

Contributing

Contributions are welcome! To add support for a new bank:

1. Fork the repository
2. Create a feature branch
3. Add the importer in beancount_chile/
4. Add an extractor in beancount_chile/extractors/
5. Create anonymized test fixtures in tests/fixtures/
6. Write comprehensive tests
7. Update this README
8. Submit a pull request

Guidelines

• Privacy: Never commit real bank data. All test fixtures must use anonymized data.
• Testing: Every importer must have comprehensive tests.
• Documentation: Update README.md and CLAUDE.md with new features.
• Code Quality: Follow PEP 8 and use ruff for linting.

License

MIT License

Disclaimer

This project is not affiliated with any bank. Use at your own risk. Always verify imported transactions against your bank statements.

Support

For issues, questions, or contributions, please open an issue on GitHub.