Example - A Recommender

We have demonstrated the basic pattern match syntax. You should have mastered the basics by this point. In this section, we show two end-to-end solutions using the pattern match syntax.

A Recommendation Application

In this example, we want to recommend some messages (comments or posts) to the person Viktor Akhiezer.

How do we do this?

One way is to find Others who likes the same messages Viktor likes. And then recommend the messages that Others like but Viktor have not seen. The pattern is roughly like below

  • Viktor - (Likes>) - Message - (<Likes) - Others

  • Others - (Likes>) - NewMessage

  • Recommend NewMessage to Viktor

However, this is too fine granularity, and we are overfitting the message level data with a collaborative filtering algorithm. The intuition is that two persons are similar to each other when their "liked" messages fall into the same category (tag). This makes more sense and common than finding two persons that "likes" the same set of messages. As a result, one way to avoid this overfitting is to go one level above. That is, instead of finding common messages as a similarity base, we find common messages' tags as a similarity base. Person A and Person B are similar if they like messages that belong to the same tag. This scheme fixes the overfitting problem. In pattern match vocabulary, we have

  • Viktor - (Likes>) - Message - (Has>) - Tag - (<Has) - Message - (<Likes) - Others

  • Others - (Likes>) - NewMessage

  • Recommend NewMessage to Viktor

GSQL. RecommendMessage Application.

This time, we create the query first, and interpret the query by calling the query name with parameters. If we are satisfied with this query, we can use "install query queryName" to get the compiled query installed which has the best performance.

GSQL Recommendation Algorithm
use graph ldbc_snb
set query_timeout=60000
DROP QUERY RecommendMessage
CREATE QUERY RecommendMessage (String fn, String ln) SYNTAX v2{
SumAccum<int> @TagInCommon;
SumAccum<float> @SimilarityScore;
SumAccum<float> @Rank;
OrAccum @Liked = false;
#1. mark messages liked by Viktor
#2. calculate log similarity score for each persons share the same
# interests at Tag level.
Others =
SELECT p
FROM Person:s-(LIKES>)-:msg - (HAS_TAG>.<HAS_TAG.<LIKES)- :p
WHERE s.firstName == fn AND s.lastName == ln
ACCUM msg.@Liked = true, p.@TagInCommon +=1
POST-ACCUM p.@SimilarityScore = log (1 + p.@TagInCommon);
# recommend new messages to Viktor that have not liked by him.
RecommendedMessage =
SELECT msg
FROM Others:o-(LIKES>) - :msg
WHERE msg.@Liked == false
ACCUM msg.@Rank +=o.@SimilarityScore
ORDER BY msg.@Rank DESC
LIMIT 2;
PRINT RecommendedMessage[RecommendedMessage.content, RecommendedMessage.@Rank];
}
INTERPRET QUERY RecommendMessage ("Viktor", "Akhiezer")
#try the second person with just parameter change.
INTERPRET QUERY RecommendMessage ("Adriaan", "Jong")

You can copy the above GSQL script to a file named app1.gsql, and invoke this script file in linux command line.

Linux Bash
gsql app1.gsql
Output of App1
Using graph 'ldbc_snb'
The query RecommendMessage is dropped.
The query RecommendMessage has been added!
{
"error": false,
"message": "",
"version": {
"schema": 0,
"edition": "enterprise",
"api": "v2"
},
"results": [{"RecommendedMessage": [
{
"v_id": "549760294602",
"attributes": {
"RecommendedMessage.@Rank": 4855.49219,
"RecommendedMessage.content": "About Indira Gandhi, Gandhi established closer relatAbout Mick Jagger, eer of the band. In 1989, he waAbout Ho Chi Minh, ce Unit and ECA International, About Ottoman Empire, After t"
},
"v_type": "Post"
},
{
"v_id": "549760292109",
"attributes": {
"RecommendedMessage.@Rank": 4828.7251,
"RecommendedMessage.content": "About Ho Chi Minh, nam, as an anti-communist state, fought against the communisAbout Shiny Happy People, sale in the U."
},
"v_type": "Post"
}
]}]
}

Install the query

When you are satisfied with your query in the GSQL interpret mode, you can now install it as a generic service which has a much faster speed. Since we have been using "CREATE QUERY .." syntax, the query is added into the catalog, we can set the syntax version and install it.

GSQL Prepare Install Query
#before install the query, need to set the syntax version
SET syntax_version="v2"
USE GRAPH ldbc_snb
#install query
INSTALL QUERY RecommendMessage
GSQL Run the Installed Query
GSQL > install query RecommendMessage
Start installing queries, about 1 minute ...
RecommendMessage query: curl -X GET 'http://127.0.0.1:9000/query/ldbc_snb/RecommendMessage?fn=VALUE&ln=VALUE'. Add -H "Authorization: Bearer TOKEN" if authentication is enabled.
[========================================================================================================] 100% (1/1)
GSQL > run query RecommendMessage("Viktor", "Akhiezer")
{
"error": false,
"message": "",
"version": {
"schema": 0,
"edition": "enterprise",
"api": "v2"
},
"results": [{"RecommendedMessage": [
{
"v_id": "549760294602",
"attributes": {
"RecommendedMessage.@Rank": 4855.49219,
"RecommendedMessage.content": "About Indira Gandhi, Gandhi established closer relatAbout Mick Jagger, eer of the band. In 1989, he waAbout Ho Chi Minh, ce Unit and ECA International, About Ottoman Empire, After t"
},
"v_type": "Post"
},
{
"v_id": "549760292109",
"attributes": {
"RecommendedMessage.@Rank": 4828.7251,
"RecommendedMessage.content": "About Ho Chi Minh, nam, as an anti-communist state, fought against the communisAbout Shiny Happy People, sale in the U."
},
"v_type": "Post"
}
]}]
}
Linux Bash: Shutdown The System
#when you are not using the TigerGraph System on your laptop,
# to save resource, you can stop it by
gadmin stop
#when you need to start it again, use
gadmin start

The above use log-cosine as a similarity measurement. We can also use cosine similarity by using two persons liked messages.

GSQL Recommendation Algorithm 2
use graph ldbc_snb
set query_timeout=60000
DROP QUERY RecommendMessage
CREATE QUERY RecommendMessage (String fn, String ln) SYNTAX v2{
SumAccum<int> @MsgInCommon = 0;
SumAccum<int> @MsgCnt = 0 ;
SumAccum<int> @@InputPersonMsgCnt = 0;
SumAccum<float> @SimilarityScore;
SumAccum<float> @Rank;
SumAccum<float> @TagCnt = 0;
OrAccum @Liked = false;
float sqrtOfInputPersonMsgCnt;
#1. mark messages liked by input user
#2. find common msg between input user and other persons
Others =
SELECT p
FROM Person:s-(LIKES>)-:msg -(<LIKES)-:p
WHERE s.firstName == fn AND s.lastName == ln
ACCUM msg.@Liked = true, @@InputPersonMsgCnt += 1,
p.@MsgInCommon += 1;
sqrtOfInputPersonMsgCnt = sqrt(@@InputPersonMsgCnt);
#calculate cosine similarity score.
#|AxB|/(sqrt(Sum(A_i^2)) * sqrt(Sum(B_i^2)))
Others =
SELECT o
FROM Others:o-(LIKES>)-:msg
ACCUM o.@MsgCnt += 1
POST-ACCUM o.@SimilarityScore = o.@MsgInCommon/(sqrtOfInputPersonMsgCnt * sqrt(o.@MsgCnt));
#recommend new messages to input user that have not liked by him.
RecommendedMessage =
SELECT msg
FROM Others:o-(LIKES>) - :msg
WHERE msg.@Liked == false
ACCUM msg.@Rank +=o.@SimilarityScore
ORDER BY msg.@Rank DESC
LIMIT 3;
PRINT RecommendedMessage[RecommendedMessage.content, RecommendedMessage.@Rank];
}
INTERPRET QUERY RecommendMessage ("Viktor", "Akhiezer")
#try the second person with just parameter change.
INTERPRET QUERY RecommendMessage ("Adriaan", "Jong")