alibaba · lidongze0629 · Aug 6, 2021 · Aug 6, 2021
diff --git a/docs/analytics_engine.rst b/docs/analytics_engine.rst
@@ -335,36 +335,45 @@ Later, you can load your own algorithm from the gar package.
     # run my algorithm over a graph and get the result.
     ret = my_app(g, src="0")
 
+**Publications**
+
+- Wenfei Fan, Jingbo Xu, Wenyuan Yu, Jingren Zhou, Xiaojian Luo, Ping Lu, Qiang Yin, Yang Cao, and Ruiqi Xu. `Parallelizing Sequential Graph Computations. <https://dl.acm.org/doi/10.1145/3282488>`_, ACM Transactions on Database Systems (TODS) 43(4): 18:1-18:39.
+
+- Wenfei Fan, Jingbo Xu, Yinghui Wu, Wenyuan Yu, Jiaxin Jiang. `GRAPE: Parallelizing Sequential Graph Computations. <http://www.vldb.org/pvldb/vol10/p1889-fan.pdf>`_, The 43rd International Conference on Very Large Data Bases (VLDB), demo, 2017 (the Best Demo Award).
+
+- Wenfei Fan, Jingbo Xu, Yinghui Wu, Wenyuan Yu, Jiaxin Jiang, Zeyu Zheng, Bohan Zhang, Yang Cao, and Chao Tian. `Parallelizing Sequential Graph Computations. <https://dl.acm.org/doi/10.1145/3035918.3035942>`_, ACM SIG Conference on Management of Data (SIGMOD), 2017 (the Best Paper Award).
+
+
 Glossory
 --------
 
 LoadStrategy
 ^^^^^^^^^^^^
 
-There are three ways to maintain the nodes crossing different fragments in GRAPE.
+There are three ways to maintain the nodes crossing different fragments in GraphScope analytical engine.
 
 OnlyOut
 """""""
 
 Each fragment Fi maintains “local” nodes v and a set of “mirrors” for nodes v' in other fragments such that there exists an edge (v, v'). For instance, in addition to local nodes {4, 5, 6, 7}, F1 in graph G also stores a “mirror” node 3 and the edge (5, 3) when using OnlyOut strategy, as shown below.
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/onlyout.png
+.. image:: images/onlyout.png
   :alt: OnlyOut
 
 OnlyIn 
 """"""
 
 Under this case, each fragment Fi maintains “local” nodes v and a set of “mirrors” for nodes v' in other fragments such that there exists an edge (v', v). In graph G, F1 maintains “mirror” nodes {1, 9, 12} besides its local nodes.
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/onlyin.png
+.. image:: images/onlyin.png
   :alt: OnlyIn
 
 BothInOut
 """""""""
 
 Each fragment Fi maintains “local” nodes v and a set of “mirrors” for nodes v' in other fragments such that there exists an edge (v, v') or (v', v). Hence, in graph G, “mirror” nodes {1, 3, 9, 12} are stored in F1 when BothInOut is applied.
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/inandout.png
+.. image:: images/inandout.png
   :alt: BothInOut
 
 PartitionStrategy
@@ -375,7 +384,7 @@ Edge Cut
 
 An *edge cut* partitioning splits vertices of a graph into roughly equal size clusters. The edges are stored in the same cluster as one or both of its endpoints. Edges with endpoints distributed across different clusters are *crossing edges*.
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/ecut.png
+.. image:: images/ecut.png
   :alt: Edge Cut
 
 
@@ -384,12 +393,12 @@ Vertex Cut
 
 A *vertex-cut* partitioning divides edges of a graph into roughly equal size fragments. The vertices that hold the endpoints of an edge are also placed in the same fragment as the edge itself. A vertex has to be replicated when its adjacent edges are distributed across different fragments.
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/vcut.png
+.. image:: images/vcut.png
   :alt: Vertex Cut
 
 
-Vertices on GRAPE
-^^^^^^^^^^^^^^^^^
+Vertices on GraphScope analytical engine 
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 A node v is referred to as an
 
@@ -398,91 +407,84 @@ OuterVertex
 
 OuterVertex of fragment Fi if it resides at another fragment Fj and there exists a node v' in Fi such that (v, v') or (v', v) is an edge; e.g., nodes {1, 3, 9, 12} are OuterVertex of fragment F1 in graph G;
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/outvertex.png
+.. image:: images/outvertex.png
   :alt: OuterVertex
 
 InnerVertex
 """""""""""
 
 InnerVertex of fragment Fi if it is distributed to Fi; e.g. nodes {4, 5, 6, 7} are InnerVertex of fragment F1 in G;
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/invertex.png
+.. image:: images/invertex.png
   :alt: InnerVertex 
 
 InnerVertexWithOutgoingEdge
 """""""""""""""""""""""""""
 
 InnerVertexWithOutgoingEdge of fragment Fi if it is stored in Fi and has an adjacent edge (v, v') outcoming to a node v' in another fragment Fj; e.g., node 5 is an InnerVertexWithOutgoingEdge of F1 in G with the outgoing edge (5, 3);
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/invertexout.png
+.. image:: images/invertexout.png
   :alt: InnerVertexWithOutgoingEdge
 
 InnerVertexWithIncomingEdge
 """""""""""""""""""""""""""
 
 InnerVertexWithIncomingEdge of fragment Fi if it is maintained in Fi has an adjacent edge (v', v) incoming from a node v' in another fragment Fj; e.g., nodes {4, 5, 7} are InnerVertexWithIncomingEdge of F1 in G, and (1, 4), (9, 5), and (12, 7) are corresponding incoming edges.
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/invertexin.png
+.. image:: images/invertexin.png
   :alt: InnerVertexWithIncomingEdge
 
 MessageStrategy
 ^^^^^^^^^^^^^^^
 
-Below are some message passing and synchronization strategies adopted by GRAPE.  
+Below are some message passing and synchronization strategies adopted by GraphScope analytical engine.  
 
 AlongOutgoingEdgeToOuterVertex
 """"""""""""""""""""""""""""""
 
 Here the message is passed along crossing edges from InnerVertexWithOutgoingEdge to OuterVertex. For instance, the message is passed from node 5 to 3 in graph G.   
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/intoout.png
+.. image:: images/intoout.png
   :alt: AlongOutgoingEdgeToOuterVertex
 
 AlongIncomingEdgeToOuterVertex
 """"""""""""""""""""""""""""""
 
 Under this case, the message is passed along crossing edges from InnerVertexWithIncomingEdge to OuterVertex. For example, the message is passed from node 5 to 9 in graph G.   
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/intoout2.png
+.. image:: images/intoout2.png
   :alt: AlongIncomingEdgeToOuterVertex
 
 AlongEdgeToOuterVertex
 """"""""""""""""""""""
 
 Each message is passed along crossing edges from nodes that are both  InnerVertexWithIncomingEdge and InnerVertexWithOutgoingEdge to OuterVertex, e.g., messages are passed from node 5 to 3 and 9 and vice versa in graph G.
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/intoout3.png
+.. image:: images/intoout3.png
   :alt: AlongEdgeToOuterVertex
 
 SyncOnOuterVertexAsTarget
 """""""""""""""""""""""""
 
 It is applied in company with the OnlyOut loading strategy. Here each fragment Fi sends the states of its “mirror” node of OuterVertex v to Fj that v resides, if there exists edge (v', v) and v' is “local” node of Fi, for synchronizing different states of v. For instance, the state of “mirror” node 3 is sent from F1 to F0 for synchronization at F0. 
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/sync1.png
+.. image:: images/sync1.png
   :alt: SyncOnOuterVertexAsTarget
 
 SyncOnOuterVertexAsSource
 """""""""""""""""""""""""
 
 It is applied together with the OnlyIn loading strategy. Similar to **SyncStateOnOuterVertexAsTarget**, each fragment Fi sends the states of its “mirror” nodes v to the corresponding fragments for synchronization. The difference is that for each such “mirror”, there exists outgoing edge (v, v') to certain “local” node v' of Fi. For example, the states of “mirror” nodes 1, 9, and 12 are sent from F1 to F0 and F2 for synchronization with other states.
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/sync2.png
+.. image:: images/sync2.png
   :alt: SyncOnOuterVertexAsSource
 
 SyncOnOuterVertex
 """""""""""""""""
 
 This is applied together with the BothInOut loading strategy. Under this case, each fragment Fi sends the states of all its “mirror” nodes v to the corresponding fragments for synchronization, regardless of the directions of edges adjacent to v, e.g., the states of “mirror” nodes 1, 3, 9 and 12 are sent from F1 to F0 and F2 for further synchronization. 
 
-.. image:: https://github.com/hibee12/grapewiki/blob/master/sync3.png
+.. image:: images/sync3.png
   :alt: SyncOnOuterVertex
 
 
-**Publications**
-
-- Wenfei Fan, Jingbo Xu, Wenyuan Yu, Jingren Zhou, Xiaojian Luo, Ping Lu, Qiang Yin, Yang Cao, and Ruiqi Xu. `Parallelizing Sequential Graph Computations. <https://dl.acm.org/doi/10.1145/3282488>`_, ACM Transactions on Database Systems (TODS) 43(4): 18:1-18:39.
-
-- Wenfei Fan, Jingbo Xu, Yinghui Wu, Wenyuan Yu, Jiaxin Jiang. `GRAPE: Parallelizing Sequential Graph Computations. <http://www.vldb.org/pvldb/vol10/p1889-fan.pdf>`_, The 43rd International Conference on Very Large Data Bases (VLDB), demo, 2017 (the Best Demo Award).
-
-- Wenfei Fan, Jingbo Xu, Yinghui Wu, Wenyuan Yu, Jiaxin Jiang, Zeyu Zheng, Bohan Zhang, Yang Cao, and Chao Tian. `Parallelizing Sequential Graph Computations. <https://dl.acm.org/doi/10.1145/3035918.3035942>`_, ACM SIG Conference on Management of Data (SIGMOD), 2017 (the Best Paper Award).
diff --git a/docs/images/ecut.png b/docs/images/ecut.png
diff --git a/docs/images/inandout.png b/docs/images/inandout.png
diff --git a/docs/images/intoout.png b/docs/images/intoout.png
diff --git a/docs/images/intoout2.png b/docs/images/intoout2.png
diff --git a/docs/images/intoout3.png b/docs/images/intoout3.png
diff --git a/docs/images/invertex.png b/docs/images/invertex.png
diff --git a/docs/images/invertexin.png b/docs/images/invertexin.png
diff --git a/docs/images/invertexout.png b/docs/images/invertexout.png
diff --git a/docs/images/onlyin.png b/docs/images/onlyin.png
diff --git a/docs/images/onlyout.png b/docs/images/onlyout.png
diff --git a/docs/images/outvertex.png b/docs/images/outvertex.png
diff --git a/docs/images/sync1.png b/docs/images/sync1.png
diff --git a/docs/images/sync2.png b/docs/images/sync2.png
diff --git a/docs/images/sync3.png b/docs/images/sync3.png
diff --git a/docs/images/vcut.png b/docs/images/vcut.png