flavius

Flavius Python SDK

A Python client for Flavius, a powerful graph database system.

Installation

You can install the package using pip:

pip install flavius # python 3.12
pip install flavius-py310 # for python 3.10
pip install flavius-py311 # for python 3.11
pip install flavius-py312 # for python 3.12

Quick Start

Here's a simple example to get you started with Flavius:

from flavius import GraphDatabase, DataType

# Initialize the driver
driver = GraphDatabase.driver("http://your-flavius-server:port")
driver.verify_connectivity()

# Create namespace and graph
driver.create_namespace("my_namespace")
driver.create_graph("my_graph", namespace="my_namespace")

# Create vertex table
driver.create_vertex_table(
    "User",
    [
        ("id", DataType.BIGINT, False),  # NOT NULL
        ("name", DataType.VARCHAR),
        ("active", DataType.BOOL),
        ("score", DataType.DOUBLE),
        ("created_at", DataType.TIMESTAMP),
    ],
    "id",  # primary key
    namespace="my_namespace",
    graph="my_graph",
)

# Execute queries
records, keys = driver.execute_query(
    "MATCH (n:User) WHERE n.id = $id RETURN n",
    namespace="my_namespace",
    graph="my_graph",
    parameters={"id": 1}
)

for record in records:
    for key in keys:
        print(f"{key}: {record[key]}")
    print()

# Don't forget to close the driver
driver.close()

More Examples

For more detailed examples and advanced usage scenarios, please refer to the example/example.py file in the repository. This file contains comprehensive demonstrations of various SDK features and common use cases.

Data Types

List of Data Types

Flavius support the following data types:

Data Type	Description	Storage Size	Min Value	Max Value
BOOL	Boolean	1 Byte	N/A	N/A
BIGINT	Int64	8 Bytes	-9223372036854775808	9223372036854775807
DOUBLE	Float64	8 Bytes	-1.7976931348623157E+308	1.7976931348623157E+308
VARCHAR	UTF-8 valid string	N/A	N/A	N/A
Vertex	Vertex/Node structure	N/A	N/A	N/A
Edge	Edge/Relationship structure	N/A	N/A	N/A
Date	Date in the format 'YYYY-MM-DD'	4 Bytes	'0001-01-01'	'9999-12-31'
Time	Time with nanosecond precision	8 Bytes	'00:00:00.000000000'	'23:59:59.999999999'
DateTime	Date + Time with nanosecond precision	8 Bytes	'1677-09-21 00:12:44.999999999'	'2262-04-11 23:47:16.854775807'
Timestamp	Date + Time with nanosecond precision + timezone	8 Bytes	'1677-09-21 00:12:44.999999999 UTC'	'2262-04-11 23:47:16.854775807 UTC'

Vertex data type contains __id___ field which is BIGINT type and describes the flavius internal vertex id.

Edge data type contains __srcid__ and __dstid__ fields which are BIGINT type and describe the flavius internal source vertex id and target vertex id.

Both Vertex and Edge data type contains __label__ field which is VARCHAR type and describes the associated vertex/edge label.

Data Type Conversions

Casting between different data types. Available casting (E means explicit, A means both implicit and explicit, '-' means not):

source data type / target data type	BOOL	BIGINT	DOUBLE	VARCHAR	VERTEX	EDGE	Date	Time	DateTime	Timestamp
BOOL		E	E	E	-	-	-	-	-	-
BIGINT	E		A	E	-	-	-	-	-	-
DOUBLE	E	E		E	-	-	-	-	-	-
VARCHAR	E	E	E		-	-	E	E	E	E
VERTEX	-	-	-	-		-	-	-	-	-
EDGE	-	-	-	-	-		-	-	-	-
Date	-	-	-	-	-	-		-	A	A
Time	-	-	-	-	-	-	-		-	-
DateTime	-	-	-	-	-	-	E	E		A
Timestamp	-	-	-	-	-	-	E	E	E

Cypher Statements

Namespace

Create Namespace

Create a namespace

Syntax

CREATE NAMESPACE <namespace_name>

Example

The following example create a namespace named test_ns.

CREATE NAMESPACE test_ns

Drop Namespace

Drop a namespace and all graphs inside.

Syntax

DROP NAMESPACE <namespace_name>

Example

The following example drop a namespace named test_ns.

DROP NAMESPACE test_ns

Describe Namespace

Desribe the meta information about namespace

Syntax

DESCRIBE NAMESPACE <namespace_name>

Example

DESCRIBE NAMESPACE test_ns

Graph

Create Graph

Create a graph

Syntax

CREATE GRAPH <graph_name>

Example

Create a graph named test_graph

CREATE GRAPH test_graph

List Graph

List graph under a namespace.

Syntax

LIST GRAPH

Example

LIST GRAPH

Drop Graph

Drop graph and all underlying vertex tables and edge tables.

Syntax

DROP GRAPH <graph_name>

Example

Drop a graph named test_graph

DROP GRAPH test_graph

Describe Graph

Show meta information about given graph.

Syntax

DESCRIBE GRAPH <graph_name>

Example

DESCRIBE GRAPH test_graph

Vertex Table

Create Vertex Table

Create vertex table with given name.

Syntax

CREATE VERTEX <vertex_table_name>
(
    <column_name> <data_type> [NOT NULL | NULL ],
    <column_name> <data_type> ...
    ...
)
PRIMARY KEY <column_name> | ( <column_name>, ... )

Example

Create a vertex table named Person, has three columns, namely col1, col2 and col3. And requires col1 has an NOT NULL constraint.

CREATE VERTEX Person
(
  col1 BIGINT NOT NULL,
  col2 VARCHAR,
  col3 VARCHAR
)
PRIMARY KEY col1

Drop Vertex Table

Drop a vertex table with given name.

Syntax

DROP VERTEX <vertex_table_name>

Example

DROP VERTEX Person

Describe Vertex Table

Describe meta information about a vertex table

Syntax

DESCRIBE VERTEX <vertex_table_name>

Example

DESCRIBE VERTEX Person

List Vertex Tables

List all vertex table names.

Syntax

LIST VERTEX

Edge Table

Create Edge Table

Create edge table with associated endpoint vertex tables.

NOTE: Currently only support create directed edges.

Syntax

CREATE DIRECTED | UNDIRECTED EDGE <edge_table_name>
(
    FROM <source_vertex_table_name>
    TO <target_vertex_table_name>,
    <column_name> <data_type> [NOT NULL | NULL ],
    <column_name> <data_type> ...
)
[ WITH REVERSE EDGE <reverse_edge_table_name> ]

Example

Create a edge table named Buy, with sourcee vertex table User and target vertex table Item. And edge table has three columns, namely col1, col2 and col3. And requires col1 has an NOT NULL constraint.

And also create an edge table named rBuy which store the reverse direction of Buy.

CREATE DIRECTED EDGE Buy
(
  FROM User
  TO Item,
  col1 BIGINT NOT NULL,
  col2 VARCHAR,
  col3 VARCHAR
)
WITH REVERSE EDGE rBuy.

Drop Edge Table

Drop an edge table with given name.

Syntax

DROP EDGE <edge_table_name>

Example

DROP EDGE Buy

Describe Edge Table

Describe meta information about an edge table

Syntax

DESCRIBE EDGE <edge_table_name>

Example

DESCRIBE EDGE Buy

List Edge Tables

List all edge table names.

Syntax

LIST EDGE

Job

Create Import Job

Import Vertex Job

Syntax

IMPORT VERTEX <vertex_table_name>
COLUMNS (<vertex_property_name> = $<file_column_index>, <vertex_property_name> = $<file_column_index>, ... )
FROM <import_file_source_uri> sourceOptions
FORMAT AS { CSV } fileFormatOptions
[ importOptions ]

For S3:

WITH (
  REGION = <region>,
  ACCESS_KEY_ID = <access_key_id>,
  SECRET_ACCESS_KEY = <secret_access_key>
)

For Oss :

WITH (
  REGION = <region>,
  ACCESS_KEY_ID = <access_key_id>,
  SECRET_ACCESS_KEY = <secret_access_key>,
  ENDPOINT = <endpoint>
)

fileFormatOptions

(
  <key> = <value>,
  <key> = <value>,
  ...
)

For csv file format, user can specify the following options:

Key	Description	Value Type	Default Value
has_header	Whether the file as header line	Boolean	false
delimiter	Delimiter to separate the columns	String	","
null_value	Recognized spellings for null values.	String	""

importOptions

PROPERTIES (
  <key> = <value>,
  <key> = <value>,
  ...
)

Key	Description	Value Type	Default Value
duplicate_vertex_handling	When importing encountered duplicated vertex, how to handle it.	"fail" : Fail the job. "overwrite" : Overwrite the duplicated vertex value. "ignore": Ignore the vertex.	"ignore"
log_problematic_lines	Whether to log problematic lines.	Boolean	false
format_error_handling	How to handle bad format lines.	"fail" : Fail the job. "ignore" : Skip the error line.	"ignore"

Example

IMPORT VERTEX Person COLUMNS("col1" = $0, "col2" = $1, "col3" = $2)
  FROM "s3://kasma-fileio-ci/tinysoc/vPerson.csv"
  WITH (region = "xxx", access_key_id = "xxx", secret_access_key = "xxx" )
  FORMAT AS CSV (has_header = true, delimiter = ",")
  PROPERTIES (duplicate_vertex_handling = "ignore", log_problematic_lines = true, format_error_handling = "ignore")

Example of importing from oss

IMPORT VERTEX Person COLUMNS("col1" = $0, "col2" = $1, "col3" = $2)
  FROM "oss://kasma-fileio-ci/tinysoc/vPerson.csv"
  WITH (region = "xxx", access_key_id = "xxx", secret_access_ke = "xxx", endpoint = "https://oss-cn-hongkong.aliyuncs.com")
  FORMAT AS CSV (has_header = true, delimiter = "," )

Import Edge Job

Syntax

IMPORT EDGE <edge_table_name>
FROM ( <source_vertex_primary_key_name> = $<file_column_index>, <source_vertex_primary_key_name> = $<file_column_index>, ... )
TO ( <target_vertex_primary_key_name> = $<file_column_index>, <target_vertex_primary_key_name> = $<file_column_index>, ... )
COLUMNS ( <edge_property_name> = $<file_column_index>, <edge_property_name> = $<file_column_index>, ... )
FROM <import_file_source_uri> sourceOptions
FORMAT AS { CSV } fileFormatOptions
[ importOptions ]

importOptions

PROPERTIES (
  <key> = <value>,
  <key> = <value>,
  ...
)

Key	Description	Value Type	Default Value
incident_vertex_not_exists_handling	How to handle cases where the edge endpoint vertex does not exist.	"fail" : Fail the job. "ignore": Ignore the edge.	"ignore"
log_problematic_lines	Whether to log problematic lines.	Boolean	false
format_error_handling	How to handle bad format lines.	"fail" : Fail the job. "ignore" : Skip the error line.	"ignore"

Example

IMPORT EDGE Knows FROM ("col1" = $0) TO ("col1" = $1)
  COLUMNS("col1" = $2)
  FROM "s3://kasma-fileio-ci/tinysoc/eKnows.csv"
  WITH (region = "xxx", access_key_id = "xxx", secret_access_key = "xxx" )
  FORMAT AS CSV (has_header = true, delimiter = "," )
  PROPERTIES (incident_vertex_not_exists_handling = "fail", log_problematic_lines = true, format_error_handling = "ignore")

Check Import Job Staus

CHECK JOB <job_id>

Query

MATCH (n:Person) RETURN n

Match on multiple node labels

Find nodes with Person or Item labels.

MATCH (a:Person:Item) RETURN a

Match on multiple rel types

Find relationship with rBuy or Knows relationship types.

MATCH (a)<-[r:rBuy|:Knows]-(b) RETURN a

Insert to flavius

Insert one or mutiple record into a vertex/edge.

Syntax

INSERT INTO <vertex/edge>
    -- Optionally specify the insert properties
    ( PROPERTYES ( ... ) )
    -- Insertion options:
    {
        MATCH ...
    }

Properties for insert vertex:

Key	Description	Value
duplicate_vertex_handling	Indicates how to handle the case insert data has duplicated vertex.	"fail": Fail the query. "overwrite": overwrite the vertex value. "ignore": ignore this vertex(default)
log_problematic_lines	True on log the records that has problems.	"true": log the records. "false": do not log the records(Default).

Properties for insert edge:

Key	Description	Value
incident_vertex_not_exists_handling	Indicateshow to handle cases where the edge endpoint vertex does not exist.	"fail": Fail the query. "ignore": Ignore the edge(Default).
log_problematic_lines	True on log the records that has problems.	"true": log the records. "false": do not log the records(Default).

Examples

Insert vertex

INSERT INTO Person2 PROPERTIES (duplicate_vertex_handling = "ignore", log_problematic_lines = "true")
MATCH (n:Person) RETURN n.col1, n.col2, n.col3

Insert edge

INSERT INTO Knows2 PROPERTIES (incident_vertex_not_exists_handling = "fail", log_problematic_lines = "true")
MATCH (a:Person)-[r:Knows]->(b:Person) RETURN a.col1, b.col1, r.col1

Insert to object store

Insert one or mutiple record into a object store.

Syntax

INSERT INTO FILES(
    -- Optionally specify the insert properties
  <key> = <value>,
  ...
)
    {
        MATCH ...
    }

For local files, the valid key and values are :

Key	Description	Value
PATH	path to export to, should starts with file://	String, e.g. "file://a/b/c
FORMAT	file format	String, avaialbe values: "PARQUET".

For s3 files, the valid key and values are :

Key	Description	Value
PATH	path to export to, should starts with s3://	String, e.g. "s3://a/b/c
REGION	s3 region to export to	String
ACCESS_KEY_ID	s3 access key id	String
SECRET_ACCESS_KEY	s3 secret access key	String
FORMAT	file format	String, avaialbe values: "PARQUET".

For oss files, the valid key and values are :

Key	Description	Value
PATH	path to export to, should starts with oss://	String, e.g. "oss://a/b/c
REGION	oss region to export to	String
ACCESS_KEY_ID	oss access key id	String
SECRET_ACCESS_KEY	oss secret access key	String
ENDPOINT	oss endpoint	String, e.g. "https://oss-cn-hongkong.aliyuncs.com"
FORMAT	file format	String, avaialbe values: "PARQUET".

Explain

EXPLAIN VERBOSE|HIR|REL <stmt>

Functions

This section provides a detailed overview of aggregation and scalar functions in the database, including parameter descriptions, return types, and usage examples to help users apply them effectively.

Aggregation Functions

• count(*)

  Description: Returns the number of input rows. Applicable to all types.

  Return Type: BIGINT

  Example:

  MATCH (n:Person) RETURN count(*) AS total_people;
  

• count(x)

  Description: Returns the count of non-null input values.

  Parameter:

  • x: any type

  Return Type: BIGINT

  Example:

  MATCH (n:Person) RETURN count(n.age) AS known_ages;
  

• sum(x)

  Description: Returns the sum of all input values.

  Parameter:

  • x: BIGINT or DOUBLE

  Return Type: same as input type

  Example:

  MATCH (n:Transaction) RETURN sum(n.amount) AS total_amount;
  

• min(x)

  Description: Returns the minimum value among all input values, ignoring nulls. x must not contain nulls if it is a complex type.

  Parameter:

  • x: orderable type BIGINT or DOUBLE

  Return Type: same as input type

  Example:

  MATCH (n:Product) RETURN min(n.price) AS lowest_price;
  

• max(x)

  Description: Returns the maximum value among all input values, ignoring nulls. x must not contain nulls if it is a complex type.

  Parameter:

  • x: orderable type BIGINT or DOUBLE

  Return Type: same as input type

  Example:

  MATCH (n:Product) RETURN max(n.price) AS highest_price;
  

• avg(x)

  Description: Returns the average (arithmetic mean) of all non-null input values.

  Parameter:

  • x: BIGINT or DOUBLE

  Return Type: DOUBLE

  Example:

  MATCH (n:Student) RETURN avg(n.grade) AS average_grade;
  

• variance(x)

  Description: Returns the sample variance of all input values.

  Parameter:

  • x: BIGINT or DOUBLE

  Return Type: DOUBLE

  Example:

  MATCH (n:Employee) RETURN variance(n.salary) AS salary_variance;
  

• stddev(x)

  Description: Returns the sample standard deviation of all input values.

  Parameter:

  • x: BIGINT or DOUBLE

  Return Type: DOUBLE

  Example:

  MATCH (n:Employee) RETURN stddev(n.salary) AS salary_stddev;
  

• count_if(x)

  Description: Returns the count of TRUE input values, equivalent to count(CASE WHEN x THEN 1 END).

  Parameter:

  • x: boolean

  Return Type: BIGINT

  Example:

  MATCH (n:Person) RETURN count_if(n.active) AS active_count;
  

• set_agg(x)

  Description: Returns an list created from distinct input x elements. For complex types, x must not contain nulls.

  Parameter:

  • x: any type

  Return Type: LIST<[same as x]>

  Example:

  MATCH (n:Person) RETURN set_agg(n.city) AS unique_cities;
  

• array_agg(x)

  Description: Returns an list created from the input x elements. Ignores null inputs if the setting presto.array_agg.ignore_nulls is false.

  Parameter:

  • x: any type

  Return Type: LIST<[same as x]>

  Example:

  MATCH (n:Person) RETURN array_agg(n.name) AS names;
  

Scalar Functions

• logical AND OR NOT XOR

  Description: Logical operators for combining boolean expressions.

  Parameter:

  • x: boolean

  Return Type: BOOLEAN

  Example:

  MATCH (n:Person) WHERE n.age > 18 AND n.active RETURN n;
  

• compare > >= <> < <=

  Description: Comparison operators for comparing values.

  Parameter:

  • x: any comparable type

  Return Type: BOOLEAN

  Example:

  MATCH (n:Person) WHERE n.age > 30 RETURN n;
  

• math + - * /

  Description: Arithmetic operators for performing mathematical operations.

  Parameter:

  • x: numeric type (BIGINT or DOUBLE)

  Return Type: same as input type

  Example:

  MATCH (n:Transaction) RETURN n.amount * 1.1 AS increased_amount;
  

• ARRAY[Expression (, Expression)*]

  Description: Create an array.

  Parameter:

  • Expression: array item

  Return Type: LIST

  Example:

  MATCH (n:Transaction) RETURN ARRAY[n.name, "6"];
  

• array_sort(LIST(E))

  Description: Returns an list with the sorted order of the input array. E must be an orderable type. Null elements are placed at the end of the returned list. May throw an error if E is an LIST or ROW type and input values contain nested nulls.

  Parameter:

  • LIST(E): an list to be sorted; E must be an orderable type

  Return Type: LIST(E)

  Example:

  MATCH (n:Person) RETURN array_sort(n.friends) AS sorted_friends;
  

• array_sort(LIST(T), FUNCTION(T, U))

  Description: Returns the array sorted by values computed using specified lambda in ascending order. U must be an orderable type. Null elements will be placed at the end of the returned array. May throw if E is and ARRAY or ROW type and input values contain nested nulls. Throws if deciding the order of elements would require comparing nested null values.

  Parameter:

  • LIST(T): a list to be sorted
  • FUNCTION(T, U): a lambda function transform T to U
  • U must be an orderable type

  Return Type: LIST(T)

  Example:

  MATCH (n:Person) RETURN array_sort(n.friends, x -> x.age) AS sorted_friends;
  

• array_sort_desc(LIST(E))

  Description: Returns the array sorted in the descending order. E must be an orderable type. Null elements will be placed at the end of the returned array. May throw if E is and ARRAY or ROW type and input values contain nested nulls. Throws if deciding the order of elements would require comparing nested null values.

  Parameter:

  • LIST(E): an list to be sorted; E must be an orderable type

  Return Type: LIST(E)

  Example:

  MATCH (n:Person) RETURN array_sort_desc(n.friends) AS sorted_friends;
  

• array_sort_desc(LIST(T), FUNCTION(T, U))

  Description: Returns the array sorted by values computed using specified lambda in descending order. U must be an orderable type. Null elements will be placed at the end of the returned array. May throw if E is and ARRAY or ROW type and input values contain nested nulls. Throws if deciding the order of elements would require comparing nested null values.

  Parameter:

  • LIST(T): a list to be sorted
  • FUNCTION(T, U): a lambda function transform T to U
  • U must be an orderable type

  Return Type: LIST(T)

  Example:

  MATCH (n:Person) RETURN array_sort_desc(n.friends, x -> x.age) AS sorted_friends;
  

• contains(x, element)

  Description: Returns true if the array x contains the element. When element is of complex type, throws if x or element contains nested nulls and these need to be compared to produce a result. For REAL and DOUBLE, NANs (Not-a-Number) are considered equal.

  Parameter:

  • x: a list
  • element: any type

  Return Type: Boolean

  Example:

  MATCH (n:Person) RETURN array_sort_desc(n.friends, x -> x.age) AS sorted_friends;
  

• slice(array(E), start, length)

  Description: Returns a subarray starting from index start(or starting from the end if start is negative) with a length of length.

  Parameter:

  • array(E): a list
  • start: start index of subarray. start != 0.
  • length: length of subarray. length >= 0.

  Return Type: array(E)

  Example:

  MATCH (n:Person) RETURN slice(n.friends, 1, 10) AS sorted_friends;
  

Cast

CAST(expression AS datatype)

CAST(3.14 AS BIGINT)

Casting between different data types. Available casting (Y means yes, N means no):

source data type / target data type	Bool	String	Integer	Float
Bool	Y	Y	Y	Y
String	Y	Y	Y	Y
Integer	Y	Y	Y	Y
Float	Y	Y	Y	Y

Neo4j compability note: Neo4j use functions to casting expression values. Here we align with GQL standard which uses CAST expression which is more general.