Menu
HubRise

General Concepts

This chapter takes a close look at the API. If you're looking for a brief introduction, check our Quick Start tutorial.

1. Endpoints

HubRise API is based on REST. It uses POST, GET, PATCH, PUT and DELETE HTTP methods to create, retrieve, list, update and delete resources. Data is transmitted in the JSON format.

An endpoint is an API operation. It comprises a URL and HTTP method. Endpoints URLs are rooted at https://api.hubrise.com/v1.

Versions are included in the endpoints URLs for compatibility purposes. No breaking change will be made without changing the version, and old versions will be supported for a while.

Every API request must include an access token, which uniquely identifies the connection. The token is passed in the X-Access-Token header:

GET https://api.hubrise.com/v1/location/ordersX-Access-Token: [your_access_token]

Access tokens are created via OAuth 2.0. See Authentication.

Note: further in this documentation, the root part of the request URLs will be omitted. For example, we will use GET /location/orders in lieu of GET https://api.hubrise.com/v1/location/orders.

2. Pagination

Index endpoints return a collection of results. For example, GET /location/orders is an index endpoint.

Index endpoints paginate results. A maximum of 100 results are returned in a response. If more results exist, the request returns the first set of results along with a X-Cursor-Next response header.

To get the next set of results, send a new request and include the previously returned cursor value in the cursor URL query parameter. Repeat until the request returns no cursor value, which indicates the last set has been returned.

Index endpoints accept 2 optional parameters:

  • count: the maximum number of results to return per request. The default (and maximum) value is 100. Decrease this value if needed.

  • cursor: the next subset of results to return. Must be set to the value received in the previous X-Cursor-Next response header to iterate through the results. If this parameter is omitted, the first set of results is returned.

Example for a request returning 150 results:

First request:

GET /location/orders->Headers: X-Cursor-Next: cjkl44Body: [ // Orders 1-100 returned ... ]

Next request:

GET /location/orders?cursor=cjkl44->Headers: NoneBody: [ // Orders 101-150 returned ... ]

3. Rate Limiting

If a connection makes too many requests over a defined time window, HubRise will return a 429 (Too Many Requests) HTTP status code. This keeps HubRise performance consistent for all users.

A connection should not exceed any of the following limits:

  • 500 requests over a 1-minute window
  • 2,500 requests over a 1-hour window
  • 10,000 requests over a 1-day window

There is an additional limit of 10 requests per minute that applies to high load queries such as GET /catalogs/:id, GET /orders or GET /customer_lists/:id/customers. Make sure to throttle your request speed if you need to retrieve many orders or customers in a row.

Time windows start at a round minute, hour, and day respectively. For example, 1-hour windows begin at the start of an hour. So if a connection has used its hourly limit by 10:35, it will remain throttled until the start of the next hour, ie 11:00.

If you need to send more requests, you may be doing something wrong! Contact us at integration@hubrise.com for advice.

4. Overriding HTTP Method

Some HTTP clients can only send GET and POST requests.

To increase accessibility to these limited clients, the HTTP method can be overridden by setting the X-Http-Method-Override header in a POST request:

POST /location/orders/sd89mmHeaders: X-Http-Method-Override=PUT

This parameter is not accepted in a GET request, since a GET request should not change the state of a resource.

5. Common Data Types

Monetary Values

A number with 2 decimal digits, followed by a space and the ISO 4217 currency name. Can be preceded by a - sign for negative amounts.

Examples:

  • 8.90 EUR
  • -0.05 GBP

Decimal Values

HubRise represent decimal values as strings to eliminate any ambiguity and loss of precision during the parsing.

Most JSON implementations parse decimal numbers (eg 1.5) as floating point numbers, primarily because of the lack of decimal native type in their programming language. However in HubRise, decimal numbers are always precise numeric values and using floating point numbers would result in abnormalities such as quantities being handled as 1.000000001.

In order to enforce the use of a "precise" decimal type and to avoid the default floating point number conversion by JSON parsers, decimal values are encoded as strings in HubRise. These strings should be handled using either a decimal built-in type or a fixed point number library (such as bigdecimal in Ruby) and never converted to floating point numbers.

Examples:

  • "1"
  • "-2.5"

Dates and Times

Dates and times are encoded in ISO 8601 format.

Locations have a default timezone, which can be configured from the back office. HubRise converts times to the location default timezone, for all resources attached to a location (eg orders).

Examples:

  • Date: 2020-08-20
  • Time: 2020-08-20T06:42:46+02:00

Days of the Week

A DOW (= "Days of the Week") value designates specific days of the week.

It is a string of 7 characters, where each character represents a day of the week, from Monday (1) to Sunday (7). When the digit is replaced by -, the particular day is excluded from the list.

Examples:

  • 1234567: every day of the week
  • 12----7: Monday, Tuesday and Sunday

6. HTTP Status Codes

The API returns appropriate HTTP status codes for every request.

CodeNameDescription
200OKAll Good!
304Not ModifiedThere was no new data to return. This is not an error.
400Bad RequestThe request was invalid or cannot be otherwise served. An accompanying error message will generally explain further.
401UnauthorizedAuthentication credentials were missing or incorrect.
403ForbiddenThe request is understood, but it has been refused or access is not allowed.
404Not FoundThe requested URL is invalid or the requested resource does not exist.
415Unsupported Media TypeThe Content-Type header of the request is not supported.
422Unprocessable EntityThe request syntax is correct but it could not be successfully completed. The requests needs to be modified before being attempted again.
429Too many requestsThe rate limit of the user or organization has been reached.
500Internal ErrorThe server encountered an unexpected error. The support team generally needs to be involved to investigate the error.

7. Errors

HubRise returns comprehensive error messages when the request cannot be processed.

An error response looks like this:

{ "error_type": "routing_error"}

The possible error_types are:

  • unauthorized
  • forbidden
  • not_found
  • unprocessable_entity
  • unsupported_media_type
  • too_many_requests
  • internal_error
  • routing_error

The response may also include a field breakdown, like this:

{ "message": "Validation failed", "errors": [ { "field": "price", "message": "'abc' is not a valid monetary amount" }, ... ], "error_type": "unprocessable_entity"}

8. Private Refs

HubRise allows API clients to attach their own internal references to various objects, such as orders, order items, customers, and a few others. This can be convenient when clients need to link HubRise objects to their internal objects, but they cannot store HubRise ids.

A private ref is only visible to the client that set it. For example, let's assume client A assigns a private ref to an order:

CLIENT A: PATCH /location/orders/sd89mm

{ "private_ref": "6986"}

When client A later retrieves the order, the private_ref is included in the response:

CLIENT A: GET /location/orders/sd89mm

Response:

{ "private_ref": "6986", "status": "new", ...}

However, when client B retrieves that same order, no private_ref is included in the response:

CLIENT B: GET /location/orders/sd89mm

Response:

{ "private_ref": null, "status": "new", ...}

Furthermore, client B can set its own private ref on this order, without affecting client A's private ref. They are indeed private!

HubRise indexes private refs efficiently, which allows clients to use private refs in lieu of ids in some endpoints:

CLIENT A: GET /location/orders?private_ref=6986

Response:

[ { "private_ref": "6986", "status": "new", ... }]
Feedback on this page