HIRO Graph API

All type ids (attribute ids, entity ids, verb ids) in HIRO Graph are namespaced.

To ensure consistency and avoid ambiguity, these have a mandatory prefix http://www.purl.org/. Everything after this prefix is the shortened id, that will be by the graph, e.g. ogit/_id derives from http://www.purl.org/ogit/_id. (see http://www.purl.org/docs/index.html for further info). Replace "/" in "_id" with "%2F".

Attributes, that are not defined in OGIT have an empty namespace, e.g.: /IssueXML.

The properties id and label can never be used due to storage limitations.

Free attributes have to begin with "/", e.g. "/environmentType": "PROD",.

All attribute values must be of type string.

System attributes start with ogit/_.

ogit/_xid is an external ID which could be set on vertex creation used to map to an external system. Setting this id in a local cache will throw a 409 if another entity has the same ogit/_xid. Nevertheless multiple entities with the same ogit/_xid may exist at the same time if data is created with sync.

All attributes are indexed with the keyword analyzer, so wildcards and submatches cannot be used on them, they are also case-sensitive.

ogit/_content when written is indexed with the standard analyzer. It is also indexed with an ngram analyzer. a text like "a sample text of things" can be queried like ogit/_content:sample to find words in it, or ogit/_content.ngram:amp to find ngrams in it. This field is allowed for all entities and can be filled automatically by setting the indexed: property in the entity definition.

ogit/_content is the sum of all attribute values that are specified in OGIT under ogit:indexed-attributes in OGIT. e.g. https://github.com/arago/OGIT/blob/master/NTO/Automation/entities/MARSNode.ttl#L131

ogit/_tags when written is indexed as a list of keywords. The format when writing is "some, tag, with spaces" and internally is indexed as ["some", "tag", "with spaces"]. It can be searched like ogit/_tags:some to find full tags in the list. This field is allowed for all entities.

ogit/_owner could be set to team id to allow sharing data by policies. if not set in request default team from account profile (in case not set default for organization) would be used

ogit/_scope could be set on vertex creation to put vertex in specific data scope, if not default from context would be used

The length of attribute values must not exceed 60kB. If they are larger than 60kB, they will not be indexed. The following attributes can have an arbitrary content length, but can not be used to match in queries:

Free attributes can have any length but cannot be used in queries if over 60kB.

Limitations on data size (trying to create or update data bigger than allowed would end up with 400 error). For performance reasons it is recommended to build applications in the way that single data piece is not bigger than 1MB.

Stream responses are used to transfer possibly bigger amount of data pieces like query or timeseries values response. each item it send from server separately and could be parsed by json streaming. e. g. {"items":[{$data1},{$data2},…​,{"error":{"code":408,"messge":"internal request timeout"}}]}. Note: client shall check if data piece contain `"error"`and handle it (after "error" no more data would be send)

All requests(except authenticate) must contain a valid access token. Request a valid access token using /auth/ api.

All requests will be made against a base $url (e.g. https://core.arago.co/api/).

Most of responses are in JSON (to see exception see specific api specs)