Up next in 10
FTX API Python Tutorial (Crypto Exchange)
317 views
Dec 11, 2024
Learn how to connect to the FTX exchange using the REST API in Python. This will get you up to speed on how to use a REST API; however, I do recommend CCXT for crypto algos: https://github.com/ccxt/ccxt āļø FTX US link: https://bit.ly/3ESQbNr š Subscribe for more: https://bit.ly/3lLybeP š Follow along: https://analyzingalpha.com/ftx-rest-api-python #crypto #api #algotrading 00:00 Intro 00:16 Import FTX API 01:25 Use the FTX Market API 02:20 Get FTX Markets 05:00 Get Individual FTX Market 06:50 Get Historical FTX Data 10:30 Get FTX Order Book 14:00 Get FTX Trades
View Video Transcript
0:00
today you're going to learn how to use
0:01
the ftx api specifically the rest api in
0:05
python to grab market data and trade
0:08
data asset data and potentially even
0:11
trade data but first things first what
0:13
we'll do is we'll go ahead and get our
0:15
imports i'll go you know i'll type in
0:17
each one and cover why i'm importing it
0:20
as we go so import date time pretty
0:22
self-explanatory but we need to be able
0:24
to manipulate
0:25
date time data
0:27
import requests
0:29
we need to be able to make api requests
0:31
to the ht you know via https to the api
0:35
import pandas as pd so we can manipulate
0:38
the data as a data frame from client
0:40
import ft
0:42
x client now this is something that was
0:44
downloaded from github it isn't in pip
0:48
so you just have to be careful because
0:49
if someone says oh yeah download my
0:51
client from my repository they
0:53
theoretically could put something not so
0:55
nice or some malicious code in the
0:57
client and you never want that so
0:59
anytime you're downloading connectivity
1:01
stuff like this
1:02
just do a quick glance and make sure
1:05
that it's from a legitimate source and
1:06
there's nothing funky going on in there
1:09
and then from local settings import
1:12
ftx
1:14
us as settings this is just my api key
1:17
and api secret i've covered the local
1:19
settings file multiple times
1:21
but i'll put i'll put an example in
1:23
github
1:24
all right perfect so with the imports
1:27
out of the way the next thing we want to
1:28
do is use the market api so using
1:32
the ftx market
1:34
api
1:43
so there's a lot of information that we
1:44
can get
1:46
from
1:47
you know the market api i'm going to go
1:49
ahead and paste this
1:52
make it a little easier for you whenever
1:53
you're
1:55
you know browsing the api but basically
1:58
here's you know what we'll be greeted
1:59
with after making an api request
2:01
you can easily
2:03
see this stuff on the ftx us docs so go
2:06
to the rest api you can see markets
2:09
and we can see your response format you
2:12
know order book get a single market
2:14
um historical prices there's all sorts
2:16
of stuff and by the time that
2:19
you're done with this tutorial you'll be
2:20
able to read this too right it's pretty
2:23
easy to do so you'll you'll see what i
2:25
mean so let's go ahead and make our
2:27
first request so we'll do a get request
2:30
to the markets
2:31
api
2:32
create an api url
2:35
https
2:37
ftx us api
2:40
we'll do the api will be markets
2:44
and the url
2:45
equals
2:46
api plus
2:48
api
2:50
nope i've got that backwards api
2:54
url plus api
2:56
and then if i type that out we should
2:58
see
2:59
the proper url and that looks correct
3:03
okay so now what we'll do is we'll use
3:05
the requests to make a you know to get
3:07
the market response data okay so we'll
3:10
say markets
3:13
equal requests
3:15
dot get and our url that we just created
3:18
and we'll get that in json
3:20
and then what we'll do is we'll say data
3:22
equals markets result well i'll show you
3:25
how this might be helpful to see how
3:27
this data is returned first so you can
3:29
see that it's
3:30
returned as json
3:32
scroll down here
3:34
and that's essentially what you get
3:36
so success is true and then we have this
3:39
you know results dictionary so what
3:40
we'll do is we'll just say data equals
3:43
markets
3:44
result and then data and you can see now
3:49
we've got just the results
3:52
in a manner that we can use
3:54
for pandas right so let's scroll down
3:57
here
3:58
create a new column i'm sorry row we'll
4:01
say df equal pd
4:02
data frame
4:04
data
4:05
df equals
4:07
df.set index i'm just cleaning it well
4:11
again i'll just show you
4:13
the nitty gritty so you can see that
4:15
it's not in the best of format we want
4:18
to make it more readable so what we can
4:20
do is we'll just do df dot set index
4:24
we'll set the index to the name which is
4:25
just the ticker here and then just to
4:29
make it a little more legible right here
4:31
i'm just going to slice it just to print
4:33
it out
4:37
okay
4:42
all right and you can see let's see what
4:44
i do there
4:46
oh i messed that up
4:48
there we go
4:49
all right so you can see here um
4:51
what it looks like we just have a data
4:53
frame with some data here
4:56
pretty simple right we just made a
4:58
request
4:59
got json back convert it to a data frame
5:01
so let's do let's do that again well
5:03
this time we'll get
5:06
individual
5:08
market data
5:12
this time we'll create a get request to
5:14
markets and then it'll be the market
5:17
name so we'll do
5:22
uh ethereum why don't we do it here and
5:24
so market name equal
5:26
f usd
5:29
do path
5:30
equals we'll do the string oops
5:33
it says markets
5:34
and then market name
5:41
url equals api url plus path and then
5:44
url and you can see that you know we now
5:48
have the
5:49
url for the single market we can check
5:52
this because we can go to
5:55
markets
5:57
single market and you can see right here
6:01
you know get markets market name that's
6:03
why you know essentially that comment
6:05
there is exactly what we need so
6:08
we've also got
6:10
you know the response which we saw
6:12
previously
6:13
and you can even do order book depth and
6:15
stuff like that so basically
6:17
pretty simple that's what we need so
6:20
we've got the url formatted correctly
6:23
now we just go ahead and
6:26
make that same request we'll do response
6:28
equal request dot get url json
6:32
df equals pd data frame
6:35
result uh response sorry and then we'll
6:37
do result right because remember
6:39
that dictionary had multiple results
6:41
we'll do df here
6:43
right so now we have
6:46
that data frame with the
6:48
response for that single market so you
6:50
can see the ask
6:52
the base currency bid all of that type
6:54
of stuff
6:55
so pretty simple so you might say you
6:57
know what leo i actually want to get the
6:59
historical data well we can do that too
7:01
get
7:02
historical data
7:05
okay so let's see what we need here
7:09
get historical prices we can see this is
7:12
the
7:14
essentially the requests we need to make
7:16
so let's go ahead and
7:19
get
7:22
okay
7:26
all right and then perfect so now what
7:28
we need to do is just construct that so
7:30
we need
7:31
a resolution
7:33
equals
7:35
60 seconds times 60
7:38
minutes times 24 so that will give us
7:41
you know how many minutes because the
7:45
historical date is in minutes
7:47
how many minutes in a day
7:49
and that's
7:51
86 400. so we've got the resolution down
7:55
now we need to do the start and
7:59
we'll go ahead and do start equals date
8:01
time date time 2022 1-1
8:06
time stamp
8:07
actually let's make sure because i think
8:09
we can only get a month so it's february
8:12
so we'll just do 2-1 yep
8:15
okay
8:17
and then we'll get start
8:20
okay so that's our start date
8:23
okay now let's go ahead and construct
8:25
that string so
8:26
path equals
8:28
we'll do an f string
8:31
markets
8:32
market name which we have above right
8:36
candles
8:37
equal candles resolution
8:40
equals
8:41
resolution
8:44
and
8:45
start
8:46
equals start
8:48
okay
8:49
then we'll do url equals api url plus
8:52
path and url let's see what we got there
8:56
so okay so we've got our api then we've
8:59
got our market
9:01
then we have candles resolution
9:04
and the start okay so hopefully we got
9:06
that correct and then what we can do is
9:08
the same exact request we did before so
9:10
respond sql request
9:12
get url and grab json or convert it to
9:16
json router
9:17
df equals
9:19
pd data frame
9:22
result
9:24
i'm sorry response result
9:26
df
9:27
date
9:29
uh well let's see here
9:31
we'll just see what this looks like okay
9:33
so we can see we've got this data frame
9:35
now we can make it a little prettier
9:39
df date
9:41
equals pd
9:43
to date time
9:45
df and then that's that start time
9:48
that'll convert that to a date
9:51
and then what else do we need we'll set
9:53
the index to date so df equals df set
9:56
index
9:57
date
9:59
and we can drop that start time and time
10:03
so df equals drop sorry df drop
10:08
columns equal
10:10
start time
10:12
and time
10:14
let's see if we got that correct
10:18
okay so that looks correct
10:21
uh let's see volume
10:23
okay so it looks pretty
10:26
i guess it is by minute so maybe there
10:27
wasn't much volume there okay
10:30
all right so that's how to get
10:32
historical data obviously you can loop
10:34
through it if you want to get multiple
10:36
assets but
10:38
pretty simple now the next thing we're
10:40
going to do is go ahead and get
10:42
order book data
10:45
and you'll notice
10:46
let's go to the api
10:48
order book
10:49
you know here's the api requests we need
10:52
to make and here's the
10:54
parameters and here's a response format
10:56
okay and i'll go ahead and put that in
10:58
the jupyter notebook when i'm done just
11:00
to make things easier so let's go ahead
11:03
and well i should have copied that
11:06
copy this
11:09
make a comment
11:12
all right so we'll do depth
11:15
equals 20
11:16
path equals
11:19
market
11:20
markets
11:24
market name
11:27
order book
11:29
and then
11:30
depth we'll pass that depth
11:33
okay and we'll do url equals api url
11:36
plus path you know we've done this
11:38
previously let's make sure our
11:41
url is constructed correctly that looks
11:44
good and now let's make the request so
11:47
response equals request dot get url
11:50
json
11:51
do bids equals pd
11:54
data
11:55
frame
11:57
and again i already know how this is
12:00
formatted so we get the results and then
12:02
from there we just take the bids
12:04
okay and then we'll do ask equal pd data
12:08
frame
12:10
results
12:11
and
12:11
[Music]
12:14
type
12:15
asks
12:17
and now
12:19
bid
12:20
columns equal
12:21
well let's go ahead i'll show you what
12:24
what happens here so we'll just say bits
12:28
oh let's see what i do there
12:30
um
12:36
no results
12:39
okay
12:40
path api url
12:43
that looks good
12:45
request dot get url json
12:49
bids pd data frame
12:52
let's see let's see what we got with
12:54
that response this is always fun to do
12:57
live on youtube right
12:59
okay so we did get that bids
13:02
and asks okay so result and then
13:05
okay
13:06
so bids equals pd.data frame
13:10
results
13:12
and bids
13:19
result
13:21
that's what i did
13:23
okay so
13:29
there we go
13:30
okay so we can see now that we've got
13:31
the bids in the ass but zero and ones
13:34
don't really help us that much so we'll
13:36
do is we'll do bids
13:38
columns equal
13:40
bid price
13:41
and
13:42
bid amount
13:45
and
13:46
asks dot columns equal
13:50
you guessed it ask price
13:54
ask amount
13:57
so now we can
14:00
we'll just do bids the head
14:03
all right so that that makes things a
14:05
lot more easier to see now what we want
14:07
to do is we'll just merge these because
14:09
it'll just be better to have them you
14:11
know one table at least i think so so df
14:13
pd merge
14:15
and then it's just bids asks and what do
14:17
we want to merge on we want to merge on
14:19
the left and right indices so we'll do
14:22
and when i say left and right that means
14:25
the index of both data frames right so
14:27
left index
14:28
which is bids uh we'll do true
14:31
and then the right index
14:35
which is asks is true
14:38
and that should merge them hopefully
14:39
this won't make a liar out of me
14:41
uh let's
14:42
see
14:44
bit i keep doing that bids and asks all
14:47
right there we go perfect so now we've
14:49
got our bids and asks
14:52
and if we really want to we can also get
14:54
some summary statistics on this
14:57
df
14:58
describe
15:00
right so you can see you know mean
15:03
standard deviation and all of that so
15:06
perfect okay so now what we'll do is
15:09
we'll go ahead and get the trades okay
15:14
get trades
15:20
so you're starting to see a pattern
15:21
right it's pretty easy once you start
15:24
using apis um you know it's pretty
15:27
pretty trivial trivial so here's what we
15:29
here's the request
15:32
okay
15:33
make sure i got that right get trades
15:36
yep
15:44
okay
15:45
now what we want to do is
15:50
put that there and create the url so
15:53
path equals
15:55
f
15:56
to markets
16:00
market name
16:03
which we have as you know from previous
16:05
uh
16:06
previously set as fusd
16:09
or eth i guess it's ethereum
16:12
so url equals api url plus
16:15
path
16:16
and url let's see if that gives us the
16:19
correct url it looks like it's the it is
16:22
so what we'll do is now we'll get the
16:24
trades again
16:25
response equals request
16:30
dot get url json
16:33
df equals pd data frame and you know the
16:37
drill at this point result
16:40
i cannot results and df
16:42
head
16:44
oh boy
16:46
maybe i should quit today
16:49
okay
16:52
that of that frame yeah
16:54
there we go okay
17:15
i guess i'm gonna have to check okay
17:17
let's see here all right
17:18
look at the response
17:21
okay so that's the problem right so it
17:23
looks like something
17:25
goofed up
17:26
and that's
17:28
that's what it is i messed up the api so
17:31
um what we do now is now that we've
17:33
corrected that we can get the respon
17:35
proper response this is a problem about
17:37
live coding on youtube but that's okay
17:39
i'll keep the mistakes in there because
17:41
no one's perfect
17:42
and then
17:45
okay under here okay perfect so now what
17:47
we do now we've got
17:49
uh the trades right
17:51
so pretty easy so
17:53
now we can do this with futures too so
17:56
if you're interested in futures for
17:58
working with ftx features
18:02
and it's really more of the same right
18:04
we just get the url
18:06
so we'll say url equals
18:09
https ftx.com
18:13
futures
18:14
to different ul
18:16
equal
18:18
requests
18:19
dot get
18:21
url
18:22
json
18:24
response
18:26
let's see what you can see here where i
18:28
go to
18:29
futures right so
18:45
and search for oops
18:48
futures
19:02
and now you can also work with futures
19:05
if you want to it's pretty easy so we'll
19:06
just do url
19:33
now you can also grab futures too and
19:36
it's pretty easy it's the same thing
19:38
right url https only this time it'll be
19:41
a dot com
19:44
features
19:55
and that's it i mean if you wanted to
19:56
work with futures or any other markets
19:58
it's the same process but pretty simple
20:01
to use the ftx api
20:04
and
20:05
for the most part almost most
20:08
and that's it it's pretty easy to use uh
20:11
you know the ftx api or almost any api
20:14
for that instance once you get the hang
20:15
of it right just make a request the api
20:18
and you know manipulate the data in the
20:20
format that you want and then analyze
20:22
the results right so hopefully this was
20:24
helpful if it was
20:26
please go ahead and hit the thumbs up
20:28
button and i will see you in the next
20:30
one thanks bye
#Business & Industrial
#Finance