Riso-Tree

How to deploy the experiment for maxPNSize

0. Rename the empty graph db folder to neo4j-community-3.4.12_Gleenes_1.0_-1_new_version

1. Generate 0-1 hop pn for spatial nodes

required files:

1. graph.txt
2. graph_label.txt
3. entity.txt
4. entity_string_label.txt

I will output spatialNodesZeroOneHopPN_-1.txt

2. Load graph nodes and edges

required files:

1. Graph database file folder. (graph.db folder should be removed at the beginning)
2. graph.txt

3. Construct RTree

required files:

1. Graph database file folder.
2. entity.txt
3. spatialNodesZeroOneHopPN_-1.txt

4. Generate containID.txt

1-4 are run by using constructTree_single.sh If fail the load step (either Step 2 or 3, need to delete this folder: data/databases/graph.db).

5. Rename the generated graph database folder to from neo4j-community-3.4.12_Gleenes_1.0_-1_new_version to four different folder names. It can done manually or using copy_maxPNSize.sh (path modification required).

1. neo4j-community-3.4.12_Gleenes_1.0_10_new_version
2. neo4j-community-3.4.12_Gleenes_1.0_40_new_version
3. neo4j-community-3.4.12_Gleenes_1.0_160_new_version
4. neo4j-community-3.4.12_Gleenes_1.0_640_new_version

6. Generate the PN files (done already. Skip this time)

Run maxPNSize_construct_PN_single.sh

7. Load PN files into graph db. (maxPNSize_PNLoad_single.sh)

required files

1. PN files
2. Graph db folder
3. containID.txt

If this step fails after the graph db is accessed, need to re-run from scratch.

8. Execute the query and output the result to disk. (maxPNSize_query_multiple_queries_single.sh)

Add Experiment

Prepare--Run maintenance_add_prepare.sh

1. Copy the node_edges db to add data directory.
2. sampleFile: Sampling requires the edge format graph file. Need to generate for datasets (except for wikidata).
3. removeEdgesFromDb:
4. removeEdgesFromGraphFile
5. Backup the current db. Rename it with suffix.
6. Construct tree, containId file, and PN files.
7. Backup suffix db.
8. Load the PN.