• Author: Llanero
• Twitter: @iamllanero

This is not ready for prime time yet! Although this project does not do anything destructive on the blockchain, it does consume Debank credits and uses the DefiLlama and CoinGecko API.

Use at your own peril.

2024-04-10: This project is undergoing major changes right now for the 2023 tax season. I'm adjusting the code to be more chain aware and am currently moving on from flatten.py to txns.py (which I expect will be a pretty hefty lift). So I haven't verified any output beyond flatten.csv.

Reckon is a series of Python scripts to perform various calculations on cryptocurrency. Rather than one monolithic app, reckon uses a series of different, focused scripts with intermediate outputs to provide transparency into how calculations are performed. The overall data pipeline is as follows:

build.py -> flatten.py -> txns.py -> price.py -> tax_hifo.py

Here are the main scripts (in order of intended use):

build.py

• Responsible for obtaining onchain wallet data
• Uses config/wallets.toml for wallet addresses
• Depends on debank module for working with the API and object model
• Builds and refreshes wallets using Debank API
• Stores data as JSON files for each wallet in the output/wallets dir

flatten.py

• Creates a denormalized CSV of the JSON wallets in the output/flatten dir
• Augments data with spam detection, address tags, and transaction URLs
• Consolidates to output/flatten.csv
• Provides a number of work outputs in output/work
• The most important output is output/work/flatten_receive_tokens.toml which shows how the spam detection is working (or not working)
• Uses config/wallets.toml for wallet addresses
• Uses config/tags.toml for address tagging
• Depends on debank module for working with the API and object model

txns.py

• Responsible for compilling meaningful transactions from:
  • output/flatten.csv
  • Reports as listed in config/txns.toml
• Output is a single txns.csv file of all transactions
• Uses config/txns.toml to get reports and some configuration options
• Uses config/txns_manual.toml to manually override transactions from flatten.csv

price.py

• Responsible for adding USD value to transactions from txns.csv through a combination of DefiLlama, CoinGecko API, inference, and overrides.
• Uses config/price.toml to determine pricing rules.
• Uses config/price_manual.csv allowing final overrides. Note that the price column should be UNIT price (not total USD value).
• Output is output/price.csv and output/price_worksheet.csv if there are any missing prices
• Price rules defaults to defillama, but uses coingecko for non-EVM chains as well as problematic tokens. Refer to config/price.toml for more info on configuring the price rules.
• Depends on defillama and coingecko modules to handle API calls and caching.
• Price generates a number of intermediate working files to help debug and reconcile data. One such file is output/work/price_manual_used.csv which shows a list of all of the entries in output/price_manual.csv that have been used. This can be a drop-in replacement to keep the csv file clean or used to do a side-by-side comparison.

tax_hifo.py

• Responsible for generating data for IRS 8949-like reporting
• Uses the HIFO (highest in, first out) strategy which will take the least gains (or highest losses)
• Output is a combination of the following:
  • output/tax_hifo.csv that provides capital gains and income on a per-year, per-token basis
  • output/tax_hifo_pivot_gainloss.txt that contains a summary table of capital gains and losses
  • output/tax_hifo_pivot_income.txt that contains a summary table of all income transactions
  • output/tax_hifo/ contains detailed reporting for every taxable token

pf.py

• Creates py.csv a view portfolio style view that shows for every token:
  • Name
  • Current Qty
  • Gain Loss
  • Unit Cost
  • Buy Qty
  • Buy USD
  • Buy Unit Price
  • Sell Qty
  • Sell USD
  • Sell Unit Price
  • Income Qty
  • Income USD

You will need an access key and credits from Debank Cloud. You can get them at https://cloud.debank.com/.

Be sure to configure the following file in the config directory:

• config/wallets.toml (see sample_wallet.toml)

Other files with configuration type items that don't need to be changed:

• config/coingecko.toml
• config/debank.toml
• config/stablecoins.toml
• config/tags.toml - list of address tags
• config/txns.toml - reports from CEX and trade equivalents

Consider getting an updated version of the coingecko_coins_list.json from https://www.coingecko.com/en/api/documentation under /coins/list. You can just click on "Try it out" and cut/paste the resulting content to replace the current data/coingecko_coins_list.json.

# Be sure you have set DEBANK_ACCESSKEY in your environment
export DEBANK_ACCESSKEY="mykey"

# First build the local cache of wallets.
python3 reckon/build.py

# Flatten the wallets into CSV files and consolidates into a single file.
python3 reckon/flatten.py

# Calculates taxable transactions
python3 reckon/txns.py

# Gets the pricing data
python3 reckon/price.py

# Create a close to an IRS 8949 list based on the HIFO cost basis
# WARNING: This will not run as long as there are missing prices in price.csv
python3 reckon/taxes_hifo.py

Issues can typically be found by looking at the following files:

• output/flatten.csv
• output/txns.csv
• output/price.csv

However, Reckon also provides a lot of visibility in intermediate work files to help make track down potential issues. Look in the output/work directory.

There is also a flag in the reckon/config/__init__.py called TAX_HIFO_WARN_THRESHOLD that can be adjusted to help show high gain/loss values in the console when running tax_hifo.py.

Reckon tries to use defillama API by default and infers prices when possible. However, if it can't figure out the price, you have a few options:

• Create price rules in config/price.toml to address systemic issues such as remapping a token symbol to another (i.e. eth to weth). Options include:
  • Changing the action to "ignore"
  • Changing the action to use "coingecko"
  • Remapping the chain, symbol, or token_id
• Create an entry in config/price_manual.csv to address problems in a single transaction. Options include:
  • Providing a manual price
  • Changing the transaction type

date,symbol,chain,token_id,timestamp,price,txn_type,comment
2021-11-29 01:03:21,fBEETS,ftm,0xfcef8a994209d6916eb2c86cdd2afd60aa6f54b1,1638147801,0.54,,debank

Reckon provides a handy output/price_worksheet.csv to use to figure out the what's missing.

• If a token is consistently showing the incorrect price, most likely solution is to create a price rule.
• If a single transaction is showing the incorrect price, look at creating a manual price entry in config/price_manual.csv in the same format as for missing prices.

• Some transactions show up incorrectly due to the way the protocol mechanism works. For example:
  • A purchase that requires a subsequent claim may show up as income
  • You may have purchased a token OTC or directly from someone
• Bridging is usually excluded but could also show up as income.
• The fix will likely require making an entry in config/price_manual.csv to remap the entry to a "buy" txn_type.
• An example entry is:

date,symbol,chain,token_id,timestamp,price,txn_type,comment
2023-09-28 13:08:23,FXN,eth,0x365accfca291e7d3914637abf1f7635db165bb09,1695906503,9.30,buy,0.005626406602 ETH each

• Modify the flatten script to generate a file showing all tokens detected in the form of: chain, token_id, symbol, name, project.id, project.name, project.site_url, frequency
• Make it easy to copy items from the above list to a block list
• Modify the spam detection to use the block list instead of the current spam list