ipython-ngql 0.10.0

Jupyter extension for NebulaGraph

https://github.com/wey-gu/jupyter_nebulagraph/assets/1651790/10135264-77b5-4d3c-b68f-c5810257feeb

jupyter_nebulagraph (previously known as ipython-ngql) is a Python package designed to facilitate connections to NebulaGraph directly from Jupyter Notebooks or iPython environments. It streamlines the process of creating, debugging, and sharing Jupyter Notebooks that interact with NebulaGraph, enhancing collaborative efforts.

Inspired by ipython-sql by Catherine Devlin.

Get Started

Try it out in Google Colab.

Or check Docs: Get Started.

Installation

jupyter_nebulagraph could be installed either via pip or from this git repo itself.

Install via pip

pip install jupyter_nebulagraph

Install inside the repo

git clone [e-mail]:wey-gu/jupyter_nebulagraph.git
cd jupyter_nebulagraph
python setup.py install

Load it in Jupyter Notebook or iPython

%load_ext ngql

Connect to NebulaGraph

Arguments as below are needed to connect a NebulaGraph DB instance:

Argument	Description
--address or -addr	IP address of the NebulaGraph Instance
--port or -P	Port number of the NebulaGraph Instance
--user or -u	User name
--password or -p	Password

Below is an exmple on connecting to 127.0.0.1:9669 with username: "user" and password: "password".

%ngql --address 127.0.0.1 --port 9669 --user user --password password

Make Queries

Now two kind of iPtython Magics are supported:

Option 1: The one line stype with %ngql:

%ngql USE basketballplayer;
%ngql MATCH (v:player{name:"Tim Duncan"})-->(v2:player) RETURN v2.player.name AS Name;

Option 2: The multiple lines stype with %%ngql

%%ngql
SHOW TAGS;
SHOW HOSTS;

Query String with Variables

jupyter_nebulagraph supports taking variables from the local namespace, with the help of Jinja2 template framework, it's supported to have queries like the below example.

The actual query string should be GO FROM "Sue" OVER owns_pokemon ..., and "{{ trainer }}" was renderred as "Sue" by consuming the local variable trainer:

In [8]: vid = "player100"

In [9]: %%ngql
   ...: MATCH (v)<-[e:follow]- (v2)-[e2:serve]->(v3)
   ...:   WHERE id(v) == "{{ vid }}"
   ...: RETURN v2.player.name AS FriendOf, v3.team.name AS Team LIMIT 3;
Out[9]:   RETURN v2.player.name AS FriendOf, v3.team.name AS Team LIMIT 3;
FriendOf	Team
0	LaMarcus Aldridge	Trail Blazers
1	LaMarcus Aldridge	Spurs
2	Marco Belinelli	Warriors

Draw query results

Just call %ng_draw after queries with graph data.

# one query
%ngql GET SUBGRAPH 2 STEPS FROM "player101" YIELD VERTICES AS nodes, EDGES AS relationships;
%ng_draw

# another query
%ngql match p=(:player)-[]->() return p LIMIT 5
%ng_draw

Draw Graph Schema

%ng_draw_schema

Load Data from CSV

It's supported to load data from a CSV file into NebulaGraph with the help of ng_load_csv magic.

For example, to load data from a CSV file actor.csv into a space basketballplayer with tag player and vid in column 0, and props in column 1 and 2:

"player999","Tom Hanks",30
"player1000","Tom Cruise",40
"player1001","Jimmy X",33

Just run the below line:

%ng_load --source actor.csv --tag player --vid 0 --props 1:name,2:age --space basketballplayer

Some other examples:

# load CSV from a URL
%ng_load --source https://github.com/wey-gu/jupyter_nebulagraph/raw/main/examples/actor.csv --tag player --vid 0 --props 1:name,2:age --space demo_basketballplayer
# with rank column
%ng_load --source follow_with_rank.csv --edge follow --src 0 --dst 1 --props 2:degree --rank 3 --space basketballplayer
# without rank column
%ng_load --source follow.csv --edge follow --src 0 --dst 1 --props 2:degree --space basketballplayer

Configure ngql_result_style

By default, jupyter_nebulagraph will use pandas dataframe as output style to enable more human-readable output, while it's supported to use the raw thrift data format that comes from the nebula3-python itself.

This can be done ad-hoc with below one line:

%config IPythonNGQL.ngql_result_style="raw"

After the above line is executed, the output will be like this:

ResultSet(ExecutionResponse(
    error_code=0,
    latency_in_us=2844,
    data=DataSet(
        column_names=[b'Trainer_Name'],
        rows=[Row(
            values=[Value(
                sVal=b'Tom')]),
...
        Row(
            values=[Value(
                sVal=b'Wey')])]),
    space_name=b'pokemon_club'))

The result are always stored in variable _ in Jupyter Notebook, thus, to tweak the result, just refer a new var to it like:

In [1] : %config IPythonNGQL.ngql_result_style="raw"

In [2] : %%ngql USE pokemon_club;
    ...: GO FROM "Tom" OVER owns_pokemon YIELD owns_pokemon._dst as pokemon_id
    ...: | GO FROM $-.pokemon_id OVER owns_pokemon REVERSELY YIELD owns_pokemon._dst AS Trainer_Name;
    ...:
    ...:
Out[3]:
ResultSet(ExecutionResponse(
    error_code=0,
    latency_in_us=3270,
    data=DataSet(
        column_names=[b'Trainer_Name'],
        rows=[Row(
            values=[Value(
                sVal=b'Tom')]),
...
        Row(
            values=[Value(
                sVal=b'Wey')])]),
    space_name=b'pokemon_club'))

In [4]: r = _

In [5]: r.column_values(key='Trainer_Name')[0].cast()
Out[5]: 'Tom'

Get Help

Don't remember anything or even relying on the cheatsheet here, oen takeaway for you: the help!

In [1]: %ngql help

Examples

Jupyter Notebook

Please refer here: get_started.ipynb

iPython

In [1]: %load_ext ngql

In [2]: %ngql --address 192.168.8.128 --port 9669 --user root --password nebula
Connection Pool Created
Out[2]: 
                        Name
0           basketballplayer
1  demo_movie_recommendation
2                        k8s
3                       test

In [3]: %ngql USE basketballplayer;
   ...: %ngql MATCH (v:player{name:"Tim Duncan"})-->(v2:player) RETURN v2.player.name AS Name;
Out[3]: 
            Name
0    Tony Parker
1  Manu Ginobili