Data layer provider for OrientDB. (ORM/Repository/helpers)

orientdb orient pyorient
pip install orientdb-data-layer==0.4.0



Current state: Alpha, in development (everything works - but I will be adding new features)

by Anton Dziavitsyn 2017



orientdb_data_layer is the easy to use framework for data layer organisation for working with OrientDB

It uses PyOrient module

And provides some additional functionality (look at next section)

New features history by version


  • The ability to organize code more structurally in an architectural sense
  • Repository pattern realisation for encapsulation of the data layer logic
  • The ability to expand data-logic by adding custom methods to repositories. (in addition to base 'add', 'get', 'update', 'delete') - which will give flexibility and will keep all data logic at one place
  • The abitlity to work with model objects or abstract dictionaries (to get fully hashable results)
  • JSON results output with including linked objects


  • get_by_tree(query_dict) method to filter objects by Link


pip install orientdb_data_layer

Usage example:

from orientdb_data_layer import data_connection
from orientdb_data_layer.data import RepositoryBase

# using pyorient.ogm (ORM realisation - part of pyorient module )
# look at last version description here https://orientdb.com/docs/last/PyOrient.html
from pyorient.ogm import property

# Connecting OrientDB database
initial_drop = True  # initial_drop used to DROP DATABASE when connected (to rebuild data schema)
data_connection.connect_database('plocal://<ip_address>:2424/<database_name>', '<username>', '<password>', initial_drop)

# Describe your database schema (in database will be created correspondent object types)

# Base class for Graph vertices (V) (Nodes or Documents)
NodeBase = data_connection.NodeBase

# Base class for Graph edges (E) (Relations)
RelationshipBase = data_connection.RelationshipBase

class CustomNode(NodeBase):
    element_plural = 'nodes_collection'  # plural name for collection of nodes is required for repository
    id = property.Integer(default=0)
    name = property.String()

class CustomSubNode(NodeBase):
    element_plural = 'subnodes_collection'

    id = property.Integer(default=0)
    parent_node = property.Link(mandatory=True, nullable=False, linked_to=CustomNode)
    name = property.String()

#  create-update/register database schema
#  refresh_models - will refresh/update database schema
#  attach_models - will attach current model to existing database schema

# Now database was refreshed and it contains 2 Nodes (by our declaration)

# Working with custom repositories
# Every repository works with one node type, and supports (add, get, update, delete operations & direct sql_command)
# in additional Repository may return results in JSON with including all linked objects (by documental links)

class CustomNodeRepository(RepositoryBase):

    def __init__(self):
        super().__init__(CustomNode)  # here should be passed Node type for repository

class CustomSubNodeRepository(RepositoryBase):

    def __init__(self):

# And now we may use our repositories (Or add some additional functionality in them - if needed)

_nodeRep = CustomNodeRepository()
_subNodeRep = CustomSubNodeRepository()

parent_record = _nodeRep.add({
    'id': 1,
    'name': 'our first parent record'

sub_record = _subNodeRep.add({
    'id': 21,
    'parent_node': parent_record,
    'name': 'child'

# and now we may obtain the records by filtering:
# this will get all records of type CustomSubNode with 'id' = 1
rec = _subNodeRep.get({
    'id': 21

# From ver[0.4.0] you may use get_by_tree(query_dict) method to filter objects by Link
# This works with unlimited count of levels
rec = _subNodeRep.get_by_tree({
    'parent_node': {
        'id': 1
# rec = list of CustomSubNode when parent_node.id == 1

# rec is list of CustomSubNode objects (look at OGM description in pyorient for details)
# or we may return result as JSON (with linked parent record by our schema)
rec = _subNodeRep.get({
    'id': 21
}, result_JSON = True)


      "@rid": "#45:0",
      "@version": 1,
      "id": 21,
      "parent_node": {
        "@rid": "#33:0",
        "@version": 1,
        "id": 1,
        "name": "our first parent record"
      "name": "child"
# Also, you may use direct calls to current graph object's methods. Graph can be accessed by:
_graph = data_connection.get_graph()

Repository base methods description:

add(prop_dict, result_JSON= False):
    add record by properties dict
    :param prop_dict: dictionary of values for properties updating (OUT OF TYPE will be ignored)
    :param result_JSON: return result as JSON text (default = False)
    :return: created MODEL OBJECT or JSON

get(query_dict, result_JSON= False):
    get records by query dict
    :param query_dict: dictionary of values for records searching
    :return: list of MODEL OBJECTS

update(query_dict, prop_dict):
    update records in database
    update(dict(name='test2', id=1), dict(name= 'test3')
    will update all records with name = 'test2' and id =1, and set value of name = 'test3'

    :param query_dict: dictionary of values for records searching
    :param prop_dict: dictionary of values for properties updating (OUT OF TYPE will be ignored)
    :return: list of updated MODEL OBJECTS

    delete records in database
    :param query_dict: dictionary of values for records searching
    :return: Count of deleted records

sql_command(sqlcommand, result_as_dict=False, result_JSON= False):
    Call direct SQL query
    :param sqlcommand: query string
    :param result_JSON: return result as JSON text (default = False)
    :param result_as_dict: return result as 'list of dict' but NOT orientRecord (default = False)
    :return: list of orient records [oRecordData] or list of Json