HDC Graph and Algorithm

This section introduces methods for managing HDC graph and HDC algorithms. Note that these methods require the deployment of HDC servers for the database.

HDC Graph

showHDCGraph()

Retrieves all HDC graphs created from the graph.

Parameters

• config: RequestConfig (Optional): Request configuration.

Returns

• List[HDCGraph]: The list of retrieved HDC graphs.

Python
# Retrieves all HDC graphs of the graph 'miniCircle'

requestConfig = RequestConfig(graph="miniCircle")
hdcGraphs = Conn.showHDCGraph(requestConfig)
for hdcGraph in hdcGraphs:
    print(hdcGraph.name, "on", hdcGraph.hdcServerName)

Output
miniCircle_hdc_graph on hdc-server-1
miniCircle_hdc_graph2 on hdc-server-2

createHDCGraphBySchema()

Creates an HDC graph for the graph.

Parameters

• builder: HDCBuilder: The HDC graph to be created; the attributes hdcGraphName and hdcServerName are mandatory, nodeSchema, edgeSchema, syncType, direction, loadId, and isDefault are optional.
• config: RequestConfig (Optional): Request configuration.

Returns

• JobResponse: Response of the request.

Python
# Creates an HDC graph named 'test_hdc_graph' for the graph 'miniCircle'

requestConfig = RequestConfig(graph="miniCircle")
hdcBuilder = HDCBuilder(
    hdcGraphName="test_hdc_graph",
    hdcServerName="hdc-server-1",
    nodeSchema=[{"*": ["*"]}],
    edgeSchema=[
        {"direct": ["*"]},
        {"review": ["value", "content"]}
    ],
    syncType=HDCSyncType.STATIC
)

response = Conn.createHDCGraphBySchema(hdcBuilder, requestConfig)
jobID = response.jobId

time.sleep(3)
jobs = Conn.showJob(jobID, requestConfig)
for job in jobs:
    print(job.id, "-", job.status)

Output
61 - FINISHED
61_1 - FINISHED

dropHDCGraph()

Deletes a specified HDC graph of the graph.

Parameters

• hdcGraphName: str: Name of the HDC graph.
• config: RequestConfig (Optional): Request configuration.

Returns

• Response: Response of the request.

Python
# Drops the HDC graph 'miniCircle_hdc_graph2' of the graph 'miniCircle'

requestConfig = RequestConfig(graph="miniCircle")
response = Conn.dropHDCGraph("miniCircle_hdc_graph2", requestConfig)
print(response.status.code.name)

Output
SUCCESS

HDC Algorithms

showHDCAlgo()

Retrieves all HDC algorithms installed on an HDC server.

Parameters

• hdcServerName: str: Name of the HDC server.
• config: RequestConfig (Optional): Request configuration.

Returns

• List[Algo]: The list of retrieved HDC algorithms.

Python
# Retrieves all HDC algorithms installed on the HDC server 'hdc-server-1'

algos = Conn.showHDCAlgo("hdc-server-1")
data = [[algo.name, algo.writeSupportType] for algo in algos if algo.type == "algo"]
headers = ["Name", "writeSupportType"]
print(tabulate.tabulate(data, headers=headers, tablefmt="grid"))

Output
+-----------+--------------------+
| Name      | writeSupportType   |
+===========+====================+
| fastRP    | DB,FILE            |
+-----------+--------------------+
| struc2vec | DB,FILE            |
+-----------+--------------------+

installHDCAlgo()

Installs an HDC algorithm on an HDC server.

Parameters

• files: List[str]: List of the paths of the installation files, the package file (.so) is necessary while the configuration file (.yml) is optional.
• hdcServerName: str: Name of the HDC server.
• config: RequestConfig (Optional): Request configuration.

Returns

• Response: Response of the request.

Python
# Installs the HDC algorithm LPA on the HDC server 'hdc-server-1'
# The files 'libplugin_lpa.so' and 'lpa.yml' are located in the 'algo' folder that is placed in the same directory as the file you executed

response = Conn.installHDCAlgo(["algo/libplugin_lpa.so", "algo/lpa.yml"], "hdc-server-1")
print(response.status.code.name)

Output
SUCCESS

uninstallHDCAlgo()

Uninstalls an HDC algorithm from an HDC server.

Parameters

• algoName: str: Name of the algorithm.
• hdcServerName: str: Name of the HDC server.
• config: RequestConfig (Optional): Request configuration.

Returns

• Response: Response of the request.

Python
# Uninstalls the HDC algorithm LPA from the HDC server 'hdc-server-1'

response = Conn.uninstallHDCAlgo("lpa", "hdc-server-1")
print(response.status.code.name)

Output
SUCCESS

rollbackHDCAlgo()

Rolls back a specified HDC algorithm on an HDC server.

Parameters

• algoName: str: Name of the algorithm.
• hdcServerName: str: Name of the HDC server.
• config: RequestConfig (Optional): Request configuration.

Returns

• Response: Response of the request.

Python
# Rolls back the HDC algorithms LPA on the HDC server 'hdc-server-1'

response = Conn.rollbackHDCAlgo("lpa", "hdc-server-1")
print(response.status.code.name)

Output
SUCCESS

Full Example

example.py
from ultipa import UltipaConfig, Connection

ultipaConfig = UltipaConfig()
# URI example: ultipaConfig.hosts = ["mqj4zouys.us-east-1.cloud.ultipa.com:60010"]
ultipaConfig.hosts = ["192.168.1.85:60061", "192.168.1.87:60061", "192.168.1.88:60061"]
ultipaConfig.username = "<username>"
ultipaConfig.password = "<password>"

Conn = Connection.NewConnection(defaultConfig=ultipaConfig)

# Installs the HDC algorithm LPA on the HDC server 'hdc-server-1'
# The files 'libplugin_lpa.so' and 'lpa.yml' are located in the 'algo' folder that is placed in the same directory as the file you executed

response = Conn.installHDCAlgo(["algo/libplugin_lpa.so", "algo/lpa.yml"], "hdc-server-1")
print(response.status.code.name)