Refactors user balance calculation to directly parse journal
entries, enhancing accuracy and efficiency. This change
eliminates reliance on direct database queries and provides a
more reliable mechanism for determining user balances.
Adds logging for debugging purposes.
Also extracts and uses fiat metadata from invoice/payment extras.
Changes the displayed balance perspective to reflect the castle's point of view instead of the user's.
This involves:
- Displaying balances as positive when the user owes the castle
- Displaying balances as negative when the castle owes the user.
This change affects how balances are calculated and displayed in both the backend logic and the frontend templates.
Ensures unique identification for receivable and
revenue entries by generating a UUID and incorporating
it into a castle reference.
This enhances tracking and linking capabilities by
providing a consistent and easily identifiable
reference across the system.
Refactors user balance calculation to use journal entries
instead of querying Fava's query endpoint.
This change allows for exclusion of voided transactions
(tagged with #voided) in addition to pending transactions
when calculating user balances, providing more accurate
balance information.
Additionally the change improves parsing of the amounts in journal entries by using regular expressions.
Updates the pending entries API to correctly parse the amount and fiat values from the amount string, which can now contain both SATS and fiat information.
This change handles different formats of the amount string, including cases where the fiat amount is present within curly braces.
Improves handling of the amount field in user entries by parsing string formats that include both SATS and fiat currency information.
This change allows extracting the SATS amount and fiat amount/currency directly from the string, accommodating different display formats.
Switches to retrieving all journal entries from Fava and filtering in the application to allow filtering by account type and user.
This provides more flexibility and control over the data being presented to the user.
Also extracts and includes relevant metadata such as entry ID, fiat amounts, and references for improved frontend display.
Ensures that voided transactions are not included in the
list of pending entries. This prevents displaying
transactions that have been cancelled or reversed,
providing a more accurate view of truly pending items.
Removes redundant metadata from entries and postings.
The cost syntax already contains fiat/exchange rate information.
Metadata such as 'created-via', 'is-equity', and payer/payee
can be inferred from transaction direction, tags, and account names.
Instead of deleting pending expense entries, marks them as voided by adding a #voided tag.
This ensures an audit trail while excluding them from balances.
Updates the Fava client to use 'params' for the delete request.
Adds functionality to interact with Fava for managing
Beancount entries, including fetching, updating, and
deleting entries directly from the Beancount ledger.
This allows for approving/rejecting pending entries
via the API by modifying the source file through Fava.
The changes include:
- Adds methods to the Fava client for fetching all journal
entries, retrieving entry context (source and hash),
updating the entry source, and deleting entries.
- Updates the pending entries API to use the Fava journal
endpoint instead of querying transactions.
- Implements entry approval and rejection using the new
Fava client methods to modify the underlying Beancount file.
Modifies balance queries to exclude pending transactions (flag='!')
and only include cleared/completed transactions (flag='*'). This
ensures accurate balance calculations by reflecting only settled transactions.
Improves the handling of pending entries by extracting and deduplicating data from Fava's query results.
Adds support for displaying fiat amounts alongside entries and extracts them from the position data in Fava.
Streamlines receivables/payables/equity checks on the frontend by relying on BQL query to supply account type metadata and tags.
Inverts the sign of Beancount balances to represent the user's perspective, where liabilities are positive and receivables are negative.
This change ensures that user balances accurately reflect the amount the castle owes the user (positive) or the amount the user owes the castle (negative). It simplifies the logic by consistently negating the Beancount balance rather than using conditional checks based on account type.
Updates journal entry flags to align with Beancount's limited flag support.
Beancount only uses cleared (*) and pending (!) flags.
Removes the VOID and FLAGGED flags and recommends using tags instead
(e.g., "! + #voided" for voided entries, "! + #review" for flagged entries).
Updates the API to reflect this change, removing the ability to directly
"reject" an expense entry via the void flag. Instead, instructs users to
add the #voided tag in Fava.
Updates reconciliation summary to count entries with voided/review tags
instead of voided/flagged flags.
Replaces direct database queries for transactions with calls to the Fava API,
centralizing transaction logic and improving data consistency.
This change removes redundant code and simplifies the API by relying on Fava
for querying transactions based on account patterns and other criteria.
Specifically, the commit introduces new methods in the FavaClient class for
querying transactions, retrieving account transactions, and retrieving user
transactions. The API endpoints are updated to utilize these methods.
Migrates balance calculation and inventory tracking to
Fava/Beancount, leveraging Fava's query API for all
accounting calculations. This simplifies the core module
and centralizes accounting logic in Fava.
Removes direct journal entry creation in favor of using Fava for accounting.
This change centralizes accounting logic in Fava, improving auditability and consistency.
It replaces direct database interactions for recording payments and settlements with calls to the Fava client.
The changes also refactor balance retrieval to fetch data from Fava.
Integrates Fava/Beancount for managing journal entries.
This change introduces functions to format entries into Beancount
format and submit them to a Fava instance.
It replaces the previous direct database entry creation with Fava
submission for expense, receivable, and revenue entries. The existing
create_journal_entry function is also updated to submit generic
journal entries to Fava.
Refactors account balance retrieval to fetch data from Fava/Beancount
for improved accounting accuracy.
Updates user balance retrieval to use Fava/Beancount data source.
Updates Castle settings ledger slug name.
Refactors the payment processing logic to submit journal entries directly to
Fava/Beancount instead of storing them in the Castle database. It queries
Fava to prevent duplicate entries. The changes include extracting fiat
metadata from the invoice, formatting the data as a Beancount transaction
using a dedicated formatting function, and submitting it to the Fava API.
Implements a new handler to process Castle invoice payments by submitting them to Fava, which in turn writes them to a Beancount file.
This approach avoids storing payment data directly in the Castle database. The handler formats the payment as a Beancount transaction, includes fiat currency if available, and queries Fava to prevent duplicate entries.
The commit also updates documentation to reflect the changes to the invoice processing workflow.
Initializes the Fava client with default settings when the Castle extension starts.
This ensures the client is ready to interact with Fava immediately and provides feedback if Fava is not configured correctly.
The client is re-initialized if the admin updates settings later.
Introduces utilities to format Castle data models into Beancount
transactions for Fava API compatibility.
Provides functions to format transactions, postings with cost basis,
expense entries, receivable entries, and payment entries.
These functions ensure data is correctly formatted for Fava's
add_entries API, including cost basis, flags, and metadata.
Implements an asynchronous HTTP client to interact with the Fava accounting API.
This client provides methods for adding journal entries, retrieving account balances,
and querying user balances, allowing the application to delegate all accounting
logic to Fava and Beancount.
Adds settings to the Castle extension for integration
with a Fava/Beancount accounting system. This enables
all accounting operations to be managed through Fava.
It includes settings for the Fava URL, ledger slug,
and request timeout.
Improvements to equity account handling across the Castle extension:
Transaction Categorization (views_api.py):
- Prioritize equity accounts when enriching transaction entries
- Use two-pass lookup: first search for equity accounts, then fall back to liability/asset accounts
- Ensures transactions with Equity:User-<id> accounts are correctly categorized as equity
UI Enhancements (index.html, index.js):
- Add 'Equity' filter option to Recent Transactions table
- Display blue "Equity" badge for equity entries (before receivable/payable badges)
- Add isEquity() helper function to identify equity account entries
Beancount Import (import_beancount.py):
- Support importing Beancount Equity:<name> accounts
- Map Beancount "Equity:Pat" to Castle "Equity:User-<id>" accounts
- Update extract_user_from_user_account() to handle Equity: prefix
- Improve error messages to include equity account examples
- Add equity account lookup in get_account_id() with helpful error if equity not enabled
These changes ensure equity accounts (representing user capital contributions) are properly distinguished from payables and receivables throughout the system.
Add account type filtering to Recent Transactions table and fix pagination issue where filters were applied after fetching results, causing incomplete data display.
Database layer (crud.py):
- Add get_journal_entries_by_user_and_account_type() to filter entries by
both user_id and account_type at SQL query level
- Add count_journal_entries_by_user_and_account_type() for accurate counts
- Filters apply before pagination, ensuring all matching records are fetched
API layer (views_api.py):
- Add filter_account_type parameter ('asset' for receivable, 'liability' for payable)
- Refactor filtering logic to use new database-level filter functions
- Support filter combinations: user only, account_type only, user+account_type, or all
- Enrich entries with account_type metadata for UI display
Frontend (index.js):
- Add account_type to transactionFilter state
- Add accountTypeOptions computed property with receivable/payable choices
- Reorder table columns to show User before Date
- Update loadTransactions to send account_type filter parameter
- Update clearTransactionFilter to clear both user and account_type filters
UI (index.html):
- Add second filter dropdown for account type (Receivable/Payable)
- Show clear button when either filter is active
- Update button label from "Clear Filter" to "Clear Filters"
This fixes the critical bug where filtering for receivables would only show a subset of results (e.g., 2 out of 20 entries fetched) instead of all matching receivables. Now filters are applied at the database level before pagination, ensuring users see all relevant transactions.
Convert the Recent Transactions card from a list view to a paginated table
with enhanced filtering capabilities for super users.
Frontend changes:
- Replace q-list with q-table for better data presentation
- Add pagination with configurable page size (default: 20 items)
- Add transaction filter dropdown for super users to filter by username
- Define table columns: Status, Date, Description, User, Amount, Fiat, Reference
- Implement prev/next page navigation with page info display
- Add filter controls with clear filter button
Backend changes (views_api.py):
- Add pagination support with limit/offset parameters
- Add filter_user_id parameter for filtering by user (super user only)
- Enrich transaction entries with user_id and username from account lookups
- Return paginated response with total count and pagination metadata
Database changes (crud.py):
- Update get_all_journal_entries() to support offset parameter
- Update get_journal_entries_by_user() to support offset parameter
- Add count_all_journal_entries() for total count
- Add count_journal_entries_by_user() for user-specific count
This improves the Recent Transactions UX by providing better organization, easier navigation through large transaction lists, and the ability for admins to filter transactions by user.
Addresses an issue where the transaction offset could be non-numeric, causing errors in pagination.
Adds validation and parsing to ensure the offset is always an integer, falling back to 0 if necessary. Also ensures that limit is parsed into an Int.
Implements pagination for the transaction history, enabling users
to navigate through their transactions in manageable chunks. This
improves performance and user experience, especially for users
with a large number of transactions. It also introduces total entry counts.
Implements account lookup logic for user-specific accounts,
specifically Liabilities:Payable and Assets:Receivable.
This allows the system to automatically map Beancount accounts
to corresponding accounts in the Castle system based on user ID.
Improves error messages when user accounts are not properly configured.
Implements a script to import Beancount ledger transactions into the Castle accounting extension.
The script fetches BTC/EUR rates, retrieves accounts from the Castle API, maps users, parses Beancount transactions, converts EUR to sats, and uploads the data to Castle.
Adds error handling, dry-run mode, and detailed logging for improved usability.
Displays equity account status and validates the existence of user equity accounts.
Simplifies the representation of journal entry lines by replacing separate debit and credit fields with a single 'amount' field.
Positive amounts represent debits, while negative amounts represent credits, aligning with Beancount's approach. This change improves code readability and simplifies calculations for balancing entries.
Modifies receivable and payable checks to align with Beancount's accounting principles.
This adjustment ensures that the system correctly identifies receivables and payables based on the sign of the amounts, rather than just debit/credit.
Also, calculates transaction sizes using the absolute value of amounts.
Refactors the data model to use a single 'amount' field for journal entry lines, aligning with the Beancount approach.
This simplifies the model, enhances compatibility, and eliminates invalid states.
Includes a database migration to convert existing debit/credit columns to the new 'amount' field.
Updates balance calculation logic to utilize the new amount field for improved accuracy and efficiency.
Removes generic equity accounts that are no longer necessary
due to the user-specific equity model introduced by the
castle extension. Specifically, it removes
"Equity:MemberEquity" and "Equity:RetainedEarnings" accounts.
Allows superusers to grant and revoke equity eligibility for users.
Adds UI components for managing equity eligibility.
Equity-eligible users can then contribute expenses as equity.
Removes parent accounts from the database to simplify account management.
Since the application exports to Beancount and doesn't directly interface with it, parent accounts for organizational hierarchy aren't necessary. The hierarchy is implicitly derived from the colon-separated account names.
This change cleans the database and prevents accidental postings to parent accounts. Specifically removes "Assets:Bitcoin" and "Equity" accounts.
Simplifies the API endpoint for retrieving Castle users.
Instead of gathering users from various sources (accounts,
permissions, equity), it now focuses on users who have configured
their wallet settings, streamlining the process and aligning with
the intended use case.
Replaces the user ID input field with a user selection dropdown,
allowing administrators to search and select users for permission
management. This simplifies the process of assigning permissions
and improves user experience.
Fetches Castle users via a new API endpoint and filters them
based on search input. Only users with Castle accounts
(receivables, payables, equity, or permissions) are listed.
Adds a comment to the `equity_account_name` field in the
`CreateUserEquityStatus` model, clarifying that it is
auto-generated if not provided.
This improves clarity for developers using the API.
Automatically creates a user-specific equity account when a user is granted equity eligibility.
This simplifies the process of granting equity and ensures that each eligible user has a dedicated equity account.
Expands the default chart of accounts with a more
detailed hierarchical structure. This includes new
accounts for fixed assets, livestock, equity
contributions, and detailed expense categories.
The migration script only adds accounts that don't
already exist, ensuring a smooth update process.
Implements an admin permissions management page.
This change allows superusers to manage permissions directly from the castle interface, providing a more streamlined experience for administrative tasks.
Implements a Vue-based UI for managing user permissions, allowing administrators to grant and revoke access to expense accounts.
Provides views for managing permissions by user and by account, along with dialogs for granting and revoking permissions.
Integrates with the LNbits API to load accounts and permissions and to persist changes.
