Skip to content

Latest commit

 

History

History
161 lines (99 loc) · 4.76 KB

features.rst

File metadata and controls

161 lines (99 loc) · 4.76 KB
Raw

GrapheekDB Features :

GrapheekDB lets you visualize your data :

For further information, you can read : Tutorial part 1 : Adding and removing nodes and edges

GrapheekDB has lookups :

(similar to Django lookups)

  • exact
  • iexact
  • contains
  • icontains
  • in
  • gt
  • gte
  • lt
  • lte
  • startswith
  • istartswith
  • endswith
  • iendswith
  • isnull

and lookups can be used on both nodes and edges

For further information, you can read : Tutorial part 2 : Lookup nodes and edges

GrapheekDB has methods for path traversal :

(similar to Gremlin path traversal methods)

  • outV, inV, bothV (for outer, inner or both vertices)
  • outE, inE, bothE (for outer, inner or both edges)
  • out_, in_, both_ (for the common pattern : "find a edge of a given kind, then find related outter, inner or both vertices")

For further information, you can read : Tutorial part 3 : Path traversal, collecting and aggregating data

GrapheekDB has aliasing et collecting methods

An example will be better :

from grapheekdb.backends.data.localmem import LocalMemoryGraph
g = LocalMemoryGraph()

# Aliasing and collecting methods
print g.V(kind__exact='document').aka('X').outV().aka('Y').collect('X', 'Y')
  • aka ("also known as") allows to give a node (or edge) an alias for further use
  • collect allows to get a collection given aliases

In the previous example, the last line will :

  • find every vertex (node) whose property 'kind' is equal to 'document' (the __exact lookup) - note, this could be written : g.V(kind='document')
  • give the name (alias) 'X' to each starting vertex iteratively
  • find every related outer vertex - which will be aliased 'Y'
  • then it will create then return a collection (a Python list) of pair of nodes

For further information, you can read : Tutorial part 3 : Path traversal, collecting and aggregating data

GrapheekDB has aggregation methods

Example 1 : count

g.V(kind='document').outV().outV().outV().count()

This line will :

  • find every vertex (node) whose property 'kind' is equal to 'document'
  • do 3 consecutive "outer vertex" traversal
  • count the number of final vertices

Example 2 : sum method

g.V(kind='document').outV().outV().outV().sum()

This line will :

  • find every vertex (node) whose property 'kind' is equal to 'document'
  • do 3 consecutive "outer vertex" traversal
  • create a dictionnary of vertex -> number of occurence

Because, as it stands, the same vertex can be found multiple times

For further information, you can read : Tutorial part 3 : Path traversal, collecting and aggregating data

GrapheekDB has some syntactic sugars

The previous example can be written :

g.V(kind='document').outV(3).sum()

"outV(3)" means : "traverse the outer vertices, 3 times"

For further information, you can read : Tutorial part 3 : Path traversal, collecting and aggregating data

GrapheekDB support nodes and edges indexes :

Example of a IPython session :

In [1]: from grapheekdb.backends.data.localmem import LocalMemoryGraph

In [2]: g = LocalMemoryGraph()

In [3]: for i in range(100000):
   ...:     g.add_node(my_id=i)
   ...:

In [4]: %timeit g.V(my_id=54321).count()
1 loops, best of 3: 197 ms per loop

In [5]: g.add_node_index('my_id')

In [6]: %timeit g.V(my_id=54321).count()
10000 loops, best of 3: 14 µs per loop

In [7]: g.remove_node_index('my_id')

In [8]: %timeit g.V(my_id=54321).count()
1 loops, best of 3: 194 ms per loop

Indexes are currently only exact match indexes

For further information, you can read : Tutorial part 4 : Scaling our app : indexes, performance tips

GrapheekDB can be used stand-alone or in client/server configuration

  • stand alone <=> plug directly to a backend such as LocalMemory or KyotoCabinet
  • In a client/server way - a server is provided and a client api is available

For further information, you can read : Tutorial part 5 : Production use : client/server configuration

GrapheekDB server supports concurrency

You can see the benchmarks