Parse Interactive Brokers Flex XML reports and convert to Python types
Project description
ibflex is a Python library for converting brokerage statement data in Interactive Brokers’ Flex XML format into standard Python data structures, so it can be conveniently processed and analyzed with Python scripts.
N.B. This module has nothing to do with programmatic trading. It’s about reading brokerage reports.
ibflex is compatible with Python version 3.7+. The parser has no dependencies beyond the Python standard library (although the optional client for fetching Flex Statements from Interactive Brokers’ server does depend on requests ).
This module is alpha software! It works and it’s useful, but the API, data structures, etc. are likely to see major changes. Several XML schemata are missing, and a few of the more newly-introduced attributes for the existing schemata. There are probably bugs.
Pull requests are welcome.
Breaking Changes in Version 0.13
Don’t upgrade to version 0.13 if you don’t want to rewrite your code.
Version 0.13 is a near-complete rewrite, with different data structures. Instead of returning nested dictionaries, ibflex.parser.parse() now returns object instances (subclasses of ibflex.Types.FlexElement). As such, values are retrieved by attribute “dot access” rather than dictionary key lookups. See below for an example.
The more interesting/useful enumerated values from the Flex reference (e.g. trade notes, corporate action types) are now parsed into Python Enums.
All sequences are tuples instead of lists. Everything is immutable.
There’s now a hard dependency on Python 3.7; the whole library is built on dataclasses and PEP 484 type hinting.
Sorry for the disruption, but it was necessary to transition the library from a quick hack to a solid foundation that is much more maintainable. The basic data structure (instance attribute access rather than dict keys) is not expected to change again.
Installation
pip install ibflex
Flex Parser
The primary facility provided is the ibflex.parser module, which parses Flex-format XML data into a hierarchy of Python objects whose structure corresponds to that of the original Flex statements, with the data converted into appropriate Python types (datetime.datetime, decimal.Decimal, etc.)
Usage example:
In [1]: from ibflex import parser
In [2]: response = parser.parse("2017-01_ibkr.xml")
In [3]: response
Out[3]: FlexQueryResponse(queryName='SCP Everything', type='AF', len(FlexStatements)=1)
In [4]: stmt = response.FlexStatements[0]
In [5]: stmt
Out[5]: FlexStatement(accountId='U770993', fromDate=datetime.date(2017, 1, 2), toDate=datetime.date(2017, 1, 31), period=None, whenGenerated=datetime.datetime(2017, 5, 10, 11, 41, 38), len(CashReport)=3, len(EquitySummaryInBase)=23, len(StmtFunds)=344, len(ChangeInPositionValues)=2, len(OpenPositions)=2140, len(FxPositions)=1, len(Trades)=339, len(CorporateActions)=1, len(CashTransactions)=4, len(InterestAccruals)=1, len(ChangeInDividendAccruals)=5, len(OpenDividendAccruals)=2, len(SecuritiesInfo)=30, len(ConversionRates)=550)
In [6]: trade = stmt.Trades[-1]
In [7]: trade
Out[7]: Trade(transactionType=<TradeType.EXCHTRADE: 'ExchTrade'>, openCloseIndicator=<OpenClose.CLOSE: 'C'>, buySell=<BuySell.SELL: 'SELL'>, orderType=<OrderType.LIMIT: 'LMT'>, assetCategory=<AssetClass.STOCK: 'STK'>, accountId='U770993', currency='USD', fxRateToBase=Decimal('1'), symbol='WMIH', description='WMIH CORP', conid='105068604', cusip=None, isin=None, listingExchange=None, multiplier=Decimal('1'), strike=None, expiry=None, putCall=None, tradeID='1742757182', reportDate=datetime.date(2017, 1, 30), tradeDate=datetime.date(2017, 1, 30), tradeTime=datetime.time(15, 39, 36), settleDateTarget=datetime.date(2017, 2, 2), exchange='BYX', quantity=Decimal('-8'), tradePrice=Decimal('1.4'), tradeMoney=Decimal('-11.2'), taxes=Decimal('0'), ibCommission=Decimal('-0.00680792'), ibCommissionCurrency='USD', netCash=Decimal('11.19319208'), netCashInBase=None, closePrice=Decimal('1.4'), notes=(<Code.PARTIAL: 'P'>,), cost=Decimal('-10.853621'), mtmPnl=Decimal('0'), origTradePrice=Decimal('0'), origTradeDate=None, origTradeID=None, origOrderID='0', openDateTime=None, fifoPnlRealized=Decimal('0.339571'), capitalGainsPnl=None, levelOfDetail='EXECUTION', ibOrderID='865480117', orderTime=datetime.datetime(2017, 1, 30, 15, 39, 36), changeInPrice=Decimal('0'), changeInQuantity=Decimal('0'), proceeds=Decimal('11.2'), fxPnl=Decimal('0'), clearingFirmID=None, transactionID='7248583136', holdingPeriodDateTime=None, ibExecID='0001090f.588f449a.01.01', brokerageOrderID=None, orderReference=None, volatilityOrderLink=None, exchOrderId=None, extExecID='S2367553204796', traderID=None, isAPIOrder=False, acctAlias='SCP 0-0', model=None, securityID=None, securityIDType=None, principalAdjustFactor=None, dateTime=None, underlyingConid=None, underlyingSecurityID=None, underlyingSymbol=None, underlyingListingExchange=None, issuer=None, sedol=None, whenRealized=None, whenReopened=None)
In [8]: print(f"{trade.tradeDate} {trade.buySell.name} {abs(trade.quantity)} {trade.symbol} @ {trade.tradePrice} {trade.currency}")
2017-01-30 SELL 8 WMIH @ 1.4 USD
In [9]: pos = stmt.OpenPositions[-1]
In [10]: pos
Out[10]: OpenPosition(side=<LongShort.SHORT: 'Short'>, assetCategory=<AssetClass.STOCK: 'STK'>, accountId='U770993', currency='USD', fxRateToBase=Decimal('1'), reportDate=datetime.date(2017, 1, 31), symbol='VXX', description='IPATH S&P 500 VIX S/T FU ETN', conid='242500577', securityID=None, cusip=None, isin=None, multiplier=Decimal('1'), position=Decimal('-75'), markPrice=Decimal('19.42'), positionValue=Decimal('-1456.5'), openPrice=Decimal('109.210703693'), costBasisPrice=Decimal('109.210703693'), costBasisMoney=Decimal('-8190.802777'), fifoPnlUnrealized=Decimal('6734.302777'), levelOfDetail='LOT', openDateTime=datetime.datetime(2015, 8, 24, 9, 28, 9), holdingPeriodDateTime=datetime.datetime(2015, 8, 24, 9, 28, 9), securityIDType=None, issuer=None, underlyingConid=None, underlyingSymbol=None, code=(), originatingOrderID='699501861', originatingTransactionID='5634129129', accruedInt=None, acctAlias='SCP 0-0', model=None, sedol=None, percentOfNAV=None, strike=None, expiry=None, putCall=None, principalAdjustFactor=None, listingExchange=None, underlyingSecurityID=None, underlyingListingExchange=None, positionValueInBase=None, unrealizedCapitalGainsPnl=None, unrealizedlFxPnl=None)
In [11]: print(f"{trade.tradeDate} {trade.buySell.name} {abs(trade.quantity)} {trade.symbol} @ {trade.tradePrice} {trade.currency}")
2017-01-30 SELL 8 WMIH @ 1.4 USD
In [12]: [sec for sec in stmt.SecuritiesInfo if sec.conid == trade.conid][0]
Out[12]: SecurityInfo(assetCategory=<AssetClass.STOCK: 'STK'>, symbol='WMIH', description='WMIH CORP', conid='105068604', securityID=None, cusip=None, isin=None, listingExchange=None, underlyingSecurityID=None, underlyingListingExchange=None, underlyingConid=None, underlyingCategory=None, subCategory=None, multiplier=Decimal('1'), strike=None, expiry=None, maturity=None, issueDate=None, type=None, sedol=None, securityIDType=None, underlyingSymbol=None, issuer=None, putCall=None, principalAdjustFactor=Decimal('1'), code=())
Flex Query Report Configuration
Configure Flex statements through Interactive Brokers account management . Reports > Flex Queries > Custom Flex Queries > Configure
You can configure whatever you like and ibflex should parse it, with these exceptions:
You can’t use European-style date formats (dd/MM/yy or dd/MM/yyyy). Just accept the default (yyyyMMdd) or get with the program and use ISO-8601 (yyyy-MM-dd).
You should use some delimiter between dates & times. The default delimiter (semicolon) is fastest to process.
For the Trades section of the statement, you can’t select the options at the top for “Symbol Summary”, “Asset Class”, or “Orders”. These will blow up the parser. It’s fine to check the box for “Asset Class” down below, along with the other selections for XML attributes.
Flex Client
Once you’ve defined various Flex queries, you can generate an access token that will allow you to generate statements and download them through the web API, instead of logging in to get them.
Reports > Settings > FlexWeb Service
Once you’ve got that set up - armed with the token, and the ID# of the desired Flex query - ibflex.client contains the facilities necessary to retrieve them:
In [1]: from ibflex import client
In [2]: token = '111111111111111111111111'
In [3]: query_id = '111111'
In [4]: response = client.download(token, query_id)
In [5]: response[:215]
Out[5]: b'<FlexQueryResponse queryName="Get Everything" type="AF">\n<FlexStatements count="1">\n<FlexStatement accountId="U111111" fromDate="2018-01-01" toDate="2018-01-31" period="LastMonth" whenGenerated="2018-02-01;211353">\n'
You can also just execute client.main() as a script:
$ python client.py -t 111111111111111111111111 -q 111111 > 2018-01_ibkr.xml
Finally, setup.py installs a script at ~/.local/bin/flexget… cron-tastic!
$ flexget -t 111111111111111111111111 -q 111111 > 2018-01_ibkr.xml
Resources
Interactive Brokers Activity Flex Query Reference
Interactive Brokers FlexWeb Service Reference
capgains - package that uses ibflex (inter alia) to calculate realized gains
ib-flex-analyzer - Analyze your Interactive Brokers Flex XML reports with pandas
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.