Stonky is a Telegram bot that provides access to financial informations. It is backed by the publicly available Yahoo Finance HTTP API, it is written in C and is released under the BSD license.
- Create your bot using the Telegram @BotFather.
- After obtaining your bot API key, store it into a file called
apikey.txtinside the bot working directory. Alternatively you can use the--apikeycommand line argument to provide your Telegram API key. - Build Stonky: you need libcurl and libsqlite installed. Just type
make. - Run Stonky with
./stonky. There is also a debug mode if you run it using the--debugoption (add --debug multiple times for even more verbose messages). For a more moderate output use--verbose. Trystonky --helpfor the full list of command line options. - Add the bot to your Telegram channel.
- IMPORTANT: The bot must be an administrator of the channel in order to read all the messages that are sent in such channel.
By default the bot will create an SQLite database in the working directory.
If you want to specify another path for your SQLite db, use the --dbfile
command line option.
The above steps should work, but you may also need this before compiling:
apt install libatomic-ops-dev
To use the bot, send messages in a Telegram channel where the bot is an administrator, otherwise directly address the bot via a private message. All the Stonky commands start with the "$" character.
$AAPL -- Reply with an update about AAPL price.
$AAPL 1y -- Reply with an ASCII art graph of AAPL price during
the past year. Valid intervals are: 1d|5d|1m|6m|1y|5y
$AAPL mc -- Reply with the result of a Montecarlo simulation on
the 52 weeks price of the specified stock. You can
say "montecarlo" instead of "mc" if you want.
The $AAPL mc command supports a period and range option, in case you
want to perform experiments with a fixed amount of days of distance between
buying and selling, or if you want to change the default 365 days range of
analysis.
$AAPL mc period 10 range 200
$AAPL vol -- Perform a volatility analysis on the stock. The
command analyzes the number of days the stock gained
and lost separately, reporting the maximum and
average profit and loss in a single day.
$AAPL trend -- Perform a Montecarlo analysis in different ranges
and all with fixed give days period between selling
and buying, to show how the stock is doing over
time.
The bot supports the concept of "list of stocks", you can add stocks to a list, then query the list to have all the prices with a single message:
$mylist: +VMW +AAPL +T -KO -- Modify the list adding/removing stocks.
$mylist: -- Ask day change of stocks in a given list.
$mylist:: -- Ask details of stocks in a given list.
$mylist: ? -- Just show the stocks in the list without
any additional info. This works without
contacting the Yahoo API at all.
$$ ls -- Show all the defined lists
Every list can be used as a portfolio for paper trading (or just to reflect your real portfolio of stocks). For every stock it is possible to buy or sell stocks, optionally specifying the amount and price. For instance to add 50 AAPL stocks bought at an average price of 130.57 write:
$myportfolio: buy AAPL 50@130.57
If just the quantity is specified, the current stock price will be used:
$myportfolio: buy AAPL 50
If only the stock name is specified, just a single stock is bought, using the current market price:
$myportfolio: buy AAPL
The bot tracks the average price of your purchases, this is an example of a chat with the bot:
me> $myportfolio: buy AAPL 50@120
bot> Now you have 50 AAPL stocks at an average price of 120.00
me> $myportfolio: buy AAPL 30@75.40
bot> Now you have 80 AAPL stocks at an average price of 103.28
You can see the list of stocks in your portfolio writing $listname?.
For example:
me> $myportfolio?
bot> AAPL | 80 | +2567.60 (+31.08%) 💚💚💚
The bot will put a green <3 for every 10% gain, a red one for every 10% loss.
Finally you can sell stocks:
me> $myportfolio: sell AAPL 10
bot> You are left with 70 AAPL stocks at an average price of 103.28
me> $myportfolio: sell AAPL
bot> You no longer own AAPL stocks
As you can see, selling without specifying the quantity will just sell all the stocks.
Stonky remembers all the sell operations performed for a given portfolio. When you sell stocks, you can tell stonky the price at which you sold. This way it will calculate the profit & loss. Example:
me> $newportfolio: buy AAPL 10
bot> Now you have 10 AAPL stocks at an average price of 133.19
me> $newportfolio: buy TSLA 10
bot> Now you have 10 TSLA stocks at an average price of 796.22
me> $newportfolio: sell AAPL 5@142.11
bot> You are left with 5 AAPL stocks at an average price of 133.19
Now with the "??" suffix, you can see the history of your sells:
me> $newportfolio??
bot> AAPL : 5 sold at 710.55$ (P/L +44.60 +6.70% 🍀), 20 seconds ago
Total portfolio performance: +44.60 USD
Both with the $list? and $list?? forms, you can specify a glob-style pattern in order to show just a subset of matching stocks, like in:
$portfolio?? AAPL
or
$portfolio?? [ab]*
To get a recap of all the available commands, just send an $HELP message
to the bot.
Not all the features may be listed in the readme. Send $HELP to the bot for more information about what it is able to do.
If the file marketdata/symbols.txt exists in the bot working directory,
and contains a list of stock symbols (it is included in the default
distribution, so the file exists unless you remove it intentionally), the
bot starts continuously analyzing the listed stocks, one per second, looking
for stocks that performed in specific ways.
Such stocks are put into lists that are handled by the bot itself. You can see the stocks in this list by writing simply:
$tothemoon:
In the channels where the bot is admin or directly to the bot itself as a private message.
The bot handles the following lists:
- The
tothemoonlist contains stocks that performed poorly in the past, but now for some reason are showing to have a massive momentum. - The
pennylist, liketothemoonbut for stocks worth less than 15$. - The
evenbetterlist contains stocks that are very strong performers in the past, and now are performing even better. - The
unstoppablelist contains stocks that have generally good parameters and gained value in every of the last 10 days of trading.
$AAPL 1y
AAPL 1y | min 57.31 max 136.91
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣾⠀⠀⡀⡀⣀⣴⣿⣾⣿
⠀⠀⠀⠀⠀⠀⠀⠀⠀⢰⣾⣿⣶⣿⣿⣿⣿⣿⣿⣿⣿
⠀⠀⠀⠀⠀⠀⠀⣠⣶⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿
⣦⣄⠀⢀⣠⣶⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿
⣿⣿⣶⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿
$AAPL mc
Buying and selling 'AAPL' at random days during
the past year would result in 35.53% average gain/loss.
1000 experiments with an average interval of 83.76 days.
$AAPL
Apple Inc. (AAPL) price is 136.91$ (+0.11%) |
pre-market: 136.66$ (-0.18%)
$AAPL vol
AAPL volatility report:
Reported profits 134 times (53.17%)
Reported loss 118 times (46.83%)
Average profit 2.18%
Average loss -1.97%
Max profit 11.98%
Max loss -12.86%
Data from last 253 days (adjusted) range.
If you are developing this bot adding features, like I'm doing right now,
every time a new feature is added to ssh to the remote server, recompile it
and execute it again can be boring. You also need to do that if you are
just a normal user and want to update the bot with the new features from
time to time. For this reason the bot has a $$ quit <password> feature
that let the administrator of the bot to quit it remotely.
To set the admin password use --adminpass option, otherwise the command
is not recognized at all. Once you do that, you can start the bot, remotely,
under a screen or tmux session, with:
while [ 1 ]; do make; ./stonky; sleep 5; done
Every time the bot quits, it recompiles itself and restarts. The sleep is important since if there is a problem, you don't want to burn all your server CPU in a busy loop.
Using the --cache command line option, it is possible to tell the bot to
store Yahoo queries about historical data (so not Yahoo queries about
the current price of the stock, which must be always updated) in the
SQLite database. This allows the "auto lists" feature of the bot to run
much faster, and is useful for development purposes. This way you can change
the parameters of the tothemoon and other autolists, and check what is
the result on the selected stocks. Note that this feature will take
more or les 650 MB of disk space, used by your SQLite database.
You can also use --noevict in addition to --cache if you want to use stale data you have in your cache. This is useful for testing.
The cache can be flushed using:
$$ flush-cache