YT YouTube2Text

YouTube2Text → Transcripts → Graph Database

Graph Database — Transcript

by Dr Gowtham Ramesh · 3,220 words · 534 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:05a graph database is designed to store
0:08and manage data in the form of graphs
0:11rather than tables as in relational
0:15databases in a graph data is represented
0:18as nodes edges and properties nodes
0:22represent entities or objects such as
0:25people places or
0:28products Ed just represent the
0:31relationship between nodes such as
0:33friend owns or friend
0:36of properties or attributes or details
0:41that describe nodes and edges such as
0:44person's age or the date of
0:47purchase another example shown on the
0:50screen represents a graph that describes
0:54a scenario within an academic
0:56environment the graph highlights two
1:00different parts that illustrate the
1:03semantics and instance of a specific
1:06moment more details on this will be
1:09discussed later in this
1:12video how graph databases differ from
1:16relational
1:18databases graph databases are highly
1:20effective in applications where
1:23relationships are
1:25crucial in a relational database
1:28modeling a network with many need to
1:30many relationships across various
1:32entities involves numerous joints
1:36table which can be complex and slow with
1:40large data sets in contrast graph
1:44databases store relationships directly
1:46in the graph structure allowing for
1:49efficient
1:50[Music]
1:55traversal the image displayed on the
1:57screen highlights a sample graph
2:00to store this graph in the database it
2:03first needs to be converted into code
2:07which then be stored in the graph
2:23DB the graph consists of two parts
2:27clearly separated in a image by a
2:29horizontal line the section below this
2:32line is referred to as resource
2:35description framework rdf part which
2:39resembles two PS in the relational
2:42database the section above the line is
2:45referred to as rdf schema which is a
2:48similar to a database
2:50schema primarily describing the graphs
2:54semantics typically domain experts
2:57design the schema section
3:00a well-designed schema enhances the
3:03Precision of the graph
3:07semantics the graph includes various
3:10types of elements or nodes such as
3:14individuals classes and
3:16relationship individuals represent
3:19instance of Real World
3:21objects for example discrete mathematics
3:25and Joe are individuals in this graph
3:39entities are represented by
3:41classes which correspond to specific
3:45instances in this graph course associate
3:49professor academic staff members are
3:51examples of
3:53classes is taught by and invokes are
3:57examples of
3:58relationships that primarily connects
4:00two
4:02entities generally there are two types
4:04of
4:06relationship object relationship and a
4:08data
4:11relationship object relationship connect
4:13two different
4:25objects while data relationship link an
4:28object to a literal
4:31literals are General forms that can be
4:35specifically represented as numbers
4:37strings integer dates or times with many
4:41data types borrowed from xml's basic
4:44data
4:46type now let us start converting the
4:49graph shown in the figure into
4:51corresponding pood let us take the first
4:53statement thep mathematics is taught
4:57by Joe Smith
5:00this is called a
5:01triples so we are going to write three
5:03different words and we make sure we have
5:06a colon in front of all three words let
5:10us take another statement from the graph
5:11which is nothing but Joe Smith is of
5:14type associate
5:16professor so write the statement in the
5:18form of a triple that is Jo Smith is
5:22type of associate professor make sure
5:26you are having colon in front of every
5:31resources let us take one more example
5:34discrete mathematics is of type
5:37course any statement we take from the
5:40graph that will be written in the form
5:42of a
5:43triples with a prefix as a
5:47coding now let us take an property is
5:51start by now first of all we have to
5:53Define is start by is of type
5:57property so we have to Define is is
6:01taught by is of type
6:14property as I already told you there are
6:17two different properties are there
6:19object property and data type
6:22property from this particular screen we
6:24are able to understand is start by is an
6:27object property
6:30similarly whenever you talk about an
6:32object property the object property will
6:34have some resources on the left side and
6:37some resources on the right
6:39side so he T by is having a domain on
6:44one side having a
6:48course
6:51similarly is Tau by is having an other
6:54side that's called a range side as an
6:57academic sta member
6:59is taught
7:02by rdfs
7:05range academic staff
7:10M and with respect to this particular
7:13graph we can also see is thought by is
7:17defined as a subproperty of
7:20in so can be converted into
7:22corresponding statement as is T
7:27by subproperty of
7:45in now in this particular
7:48code you can see there are few
7:53keywords having prefixes like rdfs and
7:57rdf and out
8:00there is a
8:02reason why we have chosen rdf rdfs and
8:06all before that let us take one more
8:08statement in the graph that is academic
8:10staff member which is nothing but
8:12associate professor is a subass of
8:14academic staff
8:17member so there are few places we are
8:20using the keywords like rdfs there are
8:23few places we are using out those are
8:26syntaxes of
8:27generating this graph
8:30those are syntactic sugar of
8:34this semantic web
8:37Concepts so wherever there is a type we
8:40include the prefix as
8:43rdf these are
8:46keywords the terms which are underlined
8:49are keywords
9:01similarly if you see the very nature of
9:04this graph comes with so many user
9:07defined terminologies so for example
9:09mathematics
9:10Jo so these terminologies are non
9:15keywords so this non keywords should be
9:18prefixed with some user defined
9:22prefixes so for example consider we are
9:25defining IO
9:26string so we Define uh
9:30um what is the purpose of having an IO
9:32Stream So IO stream will Define the name
9:34space called STD so whenever we use C
9:37out or a c in statement how do we use it
9:40in C++ STD scope resolution C out and
9:44STD scope resolution seeing so the name
9:48space name scope resolution and the
9:52class name similarly with respect to
9:54semantic web we use the Nam space name
9:59colon the resource name ABC
10:03colon course ABC colon assistant
10:06professor ABC called associate professor
10:09so in total if you see this particular
10:12sample code is having the following
10:14prefixes ABC rdf rdfs and R few prefixes
10:20are comes as part of language construct
10:23and few prefixes are user defined
10:26prefixes say for example ABC
10:30so now let us include these prefixes as
10:33part of our previous code any prefixes
10:36to be included as part of the code we
10:37have to use a statement like at prefix
10:41space the prefix name colum followed by
10:45some
10:46Ur
10:48similarly we use prefix for rdf and rdfs
10:53and
10:54out so if it is in default keywords like
10:57rdf rdfs and out
11:00the prefixes will be given by the
11:03language construct itself so the Uris we
11:07WR here all are default or specified by
11:10the language Constructor we cannot
11:12change that but if you take the user
11:15defined prefixes like ABC we can give
11:17our own uis it is not necessary that
11:20that Uris should exist in the internet
11:23all it expects is the
11:27Uris supposed to be in the proper format
11:30that's
11:32all so now you can see that a complete a
11:36code snippet of a graph the graph shown
11:40in the image is converted into
11:42corresponding
11:45code writing such complex code manually
11:49is difficult and is timec
11:51consuming instead of writing such
11:54complex code we can alternatively use
11:57the visual tools like prot to construct
12:00the similar graphs at
12:02e the tools facilitate creating such
12:05complex graphs using simple drag and
12:09drop and G based creation of
12:12graphs further the tools facilitate
12:16converting the constructed graph into a
12:18code just in a
12:21click as part of this tutorial let us
12:24discuss creating the sample example
12:26graph using prot tool
12:30let us start creating schema part first
12:34the schema consists of classes like
12:36courses Professor associate professor
12:40academic staff member and a staff
12:43member the schema also contains object
12:46relationships like involves and is Tau
12:51by let us start creating schema of this
12:54ontology first the schema consist of
12:58classes course Professor associate
13:02professor and assistant professor let us
13:05create these classes using prot
13:12tool here the professor associate
13:15professor and assistant professor are
13:18subass of the class called academic
13:21staff member let us create a class
13:25academic staff member first
13:29then let us make sure assistant
13:32professor associate professor and
13:33Professor are the classes subass of
13:37academic staff member for creating this
13:40subass relationship all we need to do is
13:43just drag and drop these classes into
13:46academic staff member so you can see
13:49moment you drag and drop the professor
13:52assant professor associate professor to
13:54an academic staff member the subass
13:57relationship will be automatically
13:58established
14:02in the same line the image shows the
14:06academic staff member is subass of staff
14:10member so let us create the staff member
14:14and drag and drop academic staff member
14:18into staff member so
14:21automatically the academic staff member
14:23Becomes subass of staff
14:27member the graph consists of set of
14:31object properties is Tau by and
14:35invols let us create these object
14:38properties using object properties
14:43tab let me create is thought by as an
14:46object property and then involves as an
14:49object property similar to a class and
14:53subclass creation just drag and drop is
14:56Tau by into
14:58invols it automatically
15:00establishes the subproperty
15:06relationship the object property of the
15:09graph having two sides domain and range
15:13side let us create this domain and range
15:16for involves so click on the domain and
15:21navigate to the corresponding class and
15:23select for instance course is the class
15:26on the domain side similarly click on
15:29the Range side and select academic staff
15:32member as a
15:34range similarly let us create domain and
15:38range of is by so for is by the domain
15:44will be course and the range will be
15:47academic staff member of course this
15:50domain and range will be automatically
15:52inherited from invols
15:56property the graph consist of two data
16:00property ID and a phone number let us
16:03create these data properties in prot
16:08using data properties
16:18tab let us create the ID and phone
16:21number using data properties tab the ID
16:26and phone number are created as a child
16:29class of top data
16:32property so the ID will have domain as
16:37well as
16:39range the domain of ID is staff
16:43member which is selected from the
16:46following
16:49tab similarly the ID having a range as a
16:55literal now you can see the literal tab
16:59not only having the type as literal you
17:02can select any one of the literal type
17:04which includes date time everything is
17:07displayed in the screen so as part of
17:10this we have successfully created the
17:12schema part of this graph let us start
17:16creating the individual section of the
17:19graph let us go to individuals by class
17:23Tab and select the class for which you
17:27want to create an instance say for
17:29example I want to create an instance of
17:31class course so click on the diamond
17:34symbol and specify discrete mathematics
17:37now the instance has been created
17:39similarly I want to create an instance
17:41as Joe Smith of type associate professor
17:45click on associate professor and then
17:48diamond symbol then specify
17:51Joe so I have created instance of
17:54associate
17:56professor let us connect the disc
17:58discret mathematics to Joe Smith by
18:01using the property called is
18:09by now navigate to course member and
18:13discrete
18:15mathematics and click on object property
18:18assertion specify the name of the object
18:21property which is nothing but is start
18:22by and specify the range which is
18:25nothing but
18:27Joe so
18:29we have created an object property is T
18:32by which connects dm with
18:35Joe now the entire graph has been
18:38successfully created using the tool prot
18:43let us save this graph into any one of
18:46the synthatic
18:49style now we can store the graph as an
18:52XML or trle or any different format
18:57though the formats are different the
18:59graph it looks
19:00similar Let Us store the graph in the
19:03form of trle specify the graph name I
19:07mentioned as a course
19:09here so extension of this graph is O WL
19:14that is web onology language Let Us open
19:17this file using visual studio
19:23code now what you see here is set of
19:27triples the entire f file is set of
19:29triples so for example involves object
19:34property okay domain is having a course
19:38now whatever the triples which we have
19:40created manually here the same set of
19:43triples has been
19:44created using the tool called
19:47prot The Tool successfully converts the
19:51graph as shown in the image to
19:53corresponding
19:56code after success converting the graph
20:00into corresponding code that is course.
20:04file let us load that all file into a
20:08graph DB what you see here is an
20:11interface of graph DB
20:14database now if you want to load an ow
20:18file which you created then you have to
20:20create a new
20:22repository specify name of an
20:25repository and create a repository
20:29moment you create a repository now you
20:32can see there is a small symbol called
20:34connect repository click on the connect
20:36repository so that the repository become
20:39active now whatever the owl file which
20:42you created you can import the owl file
20:46using upload rdf
20:49file once you successfully import the
20:52owl file into this graph DB environment
20:56now you can operate with the code from
21:00the graph
21:02DB now if you want to see what the
21:06exactly the graph which is loaded from
21:08the course. file you can click on graph
21:13overview and you can see what are all
21:16the triples has been loaded from the co.
21:20all to the graph environment let us
21:22click on the academic staff member and
21:25visual graph you can visually interact
21:28with this graph now the color code
21:31represents different resources in the
21:34graph so so for example here the DM and
21:38Joe are individuals in the graph just
21:41click on the Arrow Mark you you'll get
21:44the corresponding type information
21:46property information class information
21:49okay all this information whatever you
21:52created as part of the course. you can
21:55visually interact this is one advantage
21:58of this
21:59tool once you created and loaded a graph
22:03into this
22:04environment you can write a simple
22:07sparql queries which is much similar to
22:10SQL queries to get information from this
22:16graph now what you see in the screen is
22:19an example sparkel query which is
22:22equivalent to select star from a table
22:25name the select star from table name
22:27retrives all the tuples from the table
22:32similarly this particular quy consist of
22:35a pattern the pattern is of the form
22:40question mark s question mark p and
22:42question mark o what are these question
22:45mark S P stands for all are variables so
22:51s p are selected randomly and can be
22:55used your own variables as well so so
22:58what happens here the graph tried to
23:01match the corresponding patterns from
23:04the existing code so if you look at this
23:07the discrete mathematics is taught by
23:09Joe where the discrete mathematic can be
23:11an example of question mark s is taught
23:14by can be considered as question mark p
23:17and Joe Smith can be question mark o
23:19these triple patterns will be matched
23:21against all the throughout the graph so
23:25you can see there are so many triples
23:27can be matched from the existing graph
23:30all these triples matched against this
23:33particular pattern will be the result
23:35but we prefer only the random 10 results
23:39from the spark W query so these 10
23:42results will be displayed in the screen
23:45now let us replace the property question
23:48mark p by rdf type now we have changed
23:52the pattern into question mark s that is
23:54variable rdf type and question mark o so
23:58subject and object can be anything but
24:02we are very particular about the
24:04property rdf type so the result carries
24:08only two columns that is question mark s
24:11and question mark o these two resources
24:14that is subject and object connected by
24:17a property rdf type will be wred as a
24:21result we can
24:23further uh extend this particular
24:26pattern like the same query is extended
24:30as the question mark o is rdf type will
24:33be all class so it means that it's like
24:36an
24:37uncondition I want to have a statements
24:41with question mark s and question mark o
24:44connected by rdf type as well as the
24:47question mark W type must be an owl
24:50class the corresponding pattern will be
24:53matched against the graph the results
24:56will be written according to the pattern
24:59match now let us discuss how do we
25:02restrict our pattern for a specific
25:06subject I have taken a a complete URI of
25:10a subject course and mentioned as a
25:15subject followed by two variables it
25:17means that I need to find a pattern with
25:21a subject as course and object and um
25:26predicate can be anything so when
25:28whenever this particular pattern has
25:29been matched in the graph the
25:31corresponding set of pup will be
25:33returned now just to change I want to
25:36keep the course in the object part
25:39rather than the subject part so whenever
25:41you change that okay any subject
25:44followed by any predicate and followed
25:47by the object as in course will be
25:50returned as
25:53result the owl file loaded as part of
25:56graph DB can also been accessed from any
26:00of the higher level programming language
26:01so for example you can access this graph
26:04DB from python from java from nodejs any
26:07programming languages so the server
26:10programs written on any of these
26:11languages can interact with the graph TB
26:14and you can retry whatever the results
26:16it needs so for example as part of this
26:19particular uh presentation we have some
26:23two demo programs written in
26:25Python where we have three variables uh
26:29the Python program consists of three
26:31variables uh these three variables are
26:35loaded with the parameters retrived from
26:37the graph
26:39TV the parameters include query endpoint
26:42address header or copied from the graph
26:47DBS interface so for example the query
26:49can be any query which we discussed
26:50earlier in addition to that the endpoint
26:53address which can be obtained from graph
26:56TB as well as the accepted head retrive
26:59so all this information will be passed
27:01as an request from the application
27:03program to retrive necessary information
27:06from the graph DB the graph DB Returns
27:10the result in the form of a Json object
27:12so the Json object will be further
27:15passed to get the necessary results from
27:18the graph DB from the application uh
27:21program point of view from the
27:23literature point of view there are few
27:25terminologies has been used
27:27interchangeably such as semantic web
27:30Knowledge
27:31Graph so whenever we have a data in the
27:34form of a semantic web or a Knowledge
27:36Graph how do we store the data for
27:39storing the data we use graph TB what
27:43are the applications of these
27:45areas we can assume applications of
27:48either Knowledge Graph or semantic web
27:52or graph DB if there is a complex on the
27:55interconnected data then the best way to
27:59represent is Knowledge Graph based
28:01things highlighted in this particular
28:03screen so we have used Knowledge Graph
28:06for representing food related domain
28:09building information model
28:12iot and uh computer security
28:14recommendation system natural language
28:17processing building
28:19ontology so these domains are not
28:22limited with few applications thank you

About this transcript

This page contains the full transcript of Graph Database by Dr Gowtham Ramesh, generated from the public captions YouTube serves with the video. The transcript has 3,220 words across 534 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.