shopify_graphql
is published on Hex.
Add it to your list of dependencies in mix.exs
:
def deps do
[
{:shopify_graphql, "~> 2.1.0"}
]
end
You are also required to specify an HTTP client and JSON codec as dependencies.
shopify_graphql
supports hackney
and jason
out of the box.
You can make a request to the Shopify GraphQL admin API by passing a query to
the Shopify.GraphQL.send/2
function.
query =
"""
{
shop {
name
}
}
"""
Shopify.GraphQL.send(query, access_token: "...", shop: "myshop"))
You can manage variables using the Shopify.GraphQL.put_variable/3
and
Shopify.GraphQL.put_variables/2
functions.
query =
"""
{
query GetCustomer($customerId: ID!) {
customer(id:$customerId)
}
}
"""
query
|> Shopify.GraphQL.put_variable(:customerId, "gid://shopify/Customer/12195007594552")
|> Shopify.GraphQL.send(access_token: "...", shop: "myshop")
query
|> Shopify.GraphQL.put_variables(%{customerId: "gid://shopify/Customer/12195007594552"})
|> Shopify.GraphQL.send(access_token: "...", shop: "myshop")
All configuration must be provided on a per-request basis as a keyword list to
the second argument of Shopify.GraphQL.send/2
.
:access_token
- Shopify access token for making authenticated requests:endpoint
- endpoint for making GraphQL requests. Defaults tographql.json
.:http_client
- the HTTP client used for making requests. Defaults toShopify.GraphQL.Client.Hackney
.:http_client_opts
- additional options passed to:http_client
. Defaults to[]
.:http_headers
- a list of additional headers to send when making a request. Example:[{"x-graphql-cost-include-fields", "true"}]
. Defaults to[]
.:http_host
- HTTP host to make requests to. Defaults tomyshopify.com
. Note that using:host
rather than a combination of:host
and:shop
may be more convenient when working with public apps.:http_path
- path to the admin API. Defaults toadmin/api
.:http_port
- the HTTP port used when making requests:http_protocol
- the HTTP protocol when making requests. Defaults tohttps
.:json_codec
- codec for encoding and decoding JSON payloads:limiter
- whether to use the limiter to manage Shopify rate limiting. May betrue
,false
or an atom. Iffalse
the limiter will not be used. Iftrue
the limiter will be used and the default nameShopify.GraphQL.Limiter
will be used to interact with the limiter process. If an atom is used the limiter will be used and the atom will be used to interact with the limiter process. Defaults tofalse
.:limiter_opts
- additional options used with:limiter
. Defaults to[]
.:max_requests
- the maximum number of concurrent requests per shop. Defaults to 3.:monitor
- whether to monitor a limiter. When set totrue
the limiter process will be stopped after a certain period of time of inactivity in order to keep limiter process size to a minimum. When set tofalse
the limiter process will not stop and will stay alive indefinitely. Defaulttrue
.:monitor_timeout
- number of miliseconds to check for inactivity before stopping a partition:restore_to
- the minimum cost to begin making requests again after being throttled. Possible values are:half
,:max
or an integer. Defaults to:half
.
:retry
- module implementing a strategy for retrying requests. Disabled when set tofalse
. Defaults tofalse
:retry_opts
- options for configuring retry behavior. Defaults to[]
.:max_attempts
- the maximum number of retries. Defaults to3
.
:shop
- name of the shop that a request is being made to:version
- version of the API to use. Defaults tonil
. According to Shopify, when not specifying a version Shopify will use the oldest stable version of its API.
shopify_graphql
provides the ability to automatically manage the rate limiting
of Shopify's GraphQL admin API. We do this using what's called a limiter. The
limiter will automatically detect when queries are being rate limited and begin
managing the traffic sent to Shopify to ensure queries get executed.
The limiter is an optional feature of shopify_graphql
. To use it you will
need to add gen_stage
as a dependency to your application.
You will then need to add Shopify.GraphQL.Limiter
to your supervision tree.
When starting the limiter you may optionally pass a :name
argument. If the
:name
argument is used the process will use that value as it's name.
To send queries through the limiter you will need to pass the limiter: true
config value to Shopify.GraphQL.send/2
.
Shopify.GraphQL.send(query, access_token: "...", limiter: true, shop: "myshop")
If you named your process something other than Shopify.GraphQL.Limiter
you
will need to pass the name of the process to the :limiter
config option
instead of true
.
shopify_graphql
has a built-in mechanism for retrying requests that either
return an HTTP status code of 500 or a client error. You can enabled retries
by providing a module that implements the Shopify.GraphQL.Retry
behaviour to the
:retry
option when calling Shopify.GraphQL.send/2
.
Currently, shopify_graphql
provides a Shopify.GraphQL.Retry.Linear
strategy for
retrying requests. This strategy will automatically retry a request on a set
interval. You can configure the interval by adding :retry_in
with the number
of milliseconds to wait before sending another request to the :retry_opts
option.
Example
Shopify.GraphQL.send("{ shop { name } }", access_token: "...", retry: Shopify.GraphQL.Retry.Linear, retry_opts: [retry_in: 250], shop: "myshop")
The example above would retry a failed request after 250 milliseconds. By
default Shopify.GraphQL.Retry.Linear
will retry a request immediately if
:retry_in
has no value