Difference between revisions of "Lab: Semantic Lifting - CSV"

(Created page with "=Lab 5: Group Project Presentations= ==Topics== * Group Presentations. * SPARQL programming in python with SPARQLWrapper and Blazegraph, or alternatively RDFlib. ==Present...")
 
 
(72 intermediate revisions by 4 users not shown)
Line 1: Line 1:
=Lab 5: Group Project Presentations=
+
= Lab 6: Semantic Lifting - CSV =
  
==Topics==
+
== Topic ==
* Group Presentations.
+
Today's topic involves lifting data in CSV format into RDF. The goal is for you to learn how we can convert non-semantic data into RDF as well as getting familiar with some common vocabularies.
* SPARQL programming in python with SPARQLWrapper and Blazegraph, or alternatively RDFlib.  
 
  
 +
Fortunately, CSV is already structured in a way that makes the creation of triples relatively easy.
  
==Presentations==
+
We will also use Pandas Dataframes which will contain our CSV data in python code, and we'll do some basic data manipulation to improve our output data.
Today you will be presenting your ideas for the group project in the lab. Andreas Opdahl will be present to give you further feedback and ideas.  
 
  
 +
== Relevant Libraries - Classes, Functions and Methods and Vocabularies==
 +
=== Libraries ===
 +
* RDFlib concepts from earlier (Graph, Namespace, URIRef, Literal, BNode)
 +
* Pandas: DataFrame, apply, iterrows, astype
 +
* DBpedia Spotlight
  
==Tasks (if we have additional time)==
+
=== Semantic Vocabularies ===
 +
You do not have to use the same ones, but these should be well suited.
 +
* RDF: type
 +
* RDFS: label
 +
* Simple Event Ontology (sem): Event, eventType, Actor, hasActor, hasActorType, hasBeginTimeStamp, EndTimeStamp, hasTime, hasSubEvent
 +
* TimeLine Ontology (tl): durationInt
 +
* An example-namespace to represent terms not found elsewhere (ex): IndictmentDays, Overturned, Pardoned
 +
* DBpedia
  
After the presentations you can start on the tasks for next week.
+
== Tasks ==
These tasks are about programming SPARQL queries and inserts in a python program. Last week we added triples manually from the web interface.
+
Today we will be working with FiveThirtyEight's russia-investigation dataset. It contains special investigations conducted by the United States since the Watergate-investigation with information about them to May 2017. If you found the last weeks exercice doable, I recommend trying to write this with object-oriented programming (OOP) structure, as this tends to make for cleaner code.
  
However, sometimes we would rather want to program the insertion or updates of triples for our graphs/databases.
+
It contains the following columns:
 +
* investigation
 +
* investigation-start
 +
* investigation-end
 +
* investigation-days
 +
* name
 +
* indictment-days
 +
* type
 +
* cp-date
 +
* cp-days
 +
* overturned
 +
* pardoned
 +
* american
 +
* president
  
 +
More information about the columns and the dataset here: https://github.com/fivethirtyeight/data/tree/master/russia-investigation
  
* Redo all the SPARQL queries and updates from [https://wiki.uib.no/info216/index.php/Lab:_SPARQL Lab 4], this time writing a Python program.  
+
Our goal is to convert this non-semantic dataset into a semantic one. To do this we will go row-by-row through the dataset and extract the content of each column.
 +
An investigation may have multiple rows in the dataset if it investigates multiple people, you can choose to represent these as one or multiple entities in the graph. Each investigation may also have a sub-event representing the result of the investigation, this could for instance be indictment or guilty-plea.
  
 +
For a row we will start by creating a resource representing the investigation. In this example we handle all investigations with the same name as the samme entity, and will therefore use the name of the investigation ("investigation"-column) to create the URI:
  
==With Blazegraph==
+
<syntaxhighlight>
The most important part is that we need to import a SPARQLWrapper in order to connect to the SPARQL endpoint of Blazegraph.  
+
name = row["investigation"]
When it comes to how to do some queries I recommend scrolling down on this page for help: https://github.com/RDFLib/sparqlwrapper
+
 
 +
investigation = URIRef(ex + name)
 +
g.add((investigation, RDF.type, sem.Event))
 +
</syntaxhighlight>
 +
 
 +
Further we will create a relation between the investigation and all its associated columns. For when the investigation started we'll use the "investigation-start"-column and we can use the property sem:hasBeginTimeStamp:
 +
 
 +
<syntaxhighlight>
 +
investigation_start = row["investigation-start"]
 +
 
 +
g.add((investigation, sem.hasBeginTimeStamp, Literal(investigation_start, datatype=XSD.date)))
 +
</syntaxhighlight>
  
Remember, before you can program with Blazegraph you have to make sure its running like we did in  [https://wiki.uib.no/info216/index.php/Lab:_SPARQL Lab 4].
+
To represent the result of the investigation, if it has one, We can create another entity and connect it to the investigation using the sem:hasSubEvent. If so the following columns can be attributed to the sub-event:
Now you will be able to program queries and updates.
+
* type
 +
* indictment-days
 +
* overturned
 +
* pardon
 +
* cp_date
 +
* cp_days
 +
* name (the name of the investigatee, not the name of the investigation)
  
 +
=== Code to get you started ===
 
<syntaxhighlight>
 
<syntaxhighlight>
# How to establish connection to Blazegraph endpoint.
 
  
from SPARQLWrapper import SPARQLWrapper, JSON
+
import pandas as pd
 +
import rdflib
 +
 
 +
from rdflib import Graph, Namespace, URIRef, Literal, BNode
 +
from rdflib.namespace import RDF, RDFS, XSD
 +
 
 +
ex = Namespace("http://example.org/")
 +
dbr = Namespace("http://dbpedia.org/resource/")
 +
sem = Namespace("http://semanticweb.cs.vu.nl/2009/11/sem/")
 +
tl = Namespace("http://purl.org/NET/c4dm/timeline.owl#")
  
sparql = SPARQLWrapper("("http://localhost:9999/bigdata/sparql")")
+
g = Graph()
 +
g.bind("ex", ex)
 +
g.bind("dbr", dbr)
 +
g.bind("sem", sem)
 +
g.bind("tl", tl)
  
 +
df = pd.read_csv("data/investigations.csv")
 +
# We need to correct the type of the columns in the DataFrame, as Pandas assigns an incorrect type when it reads the file (for me at least). We use .astype("str") to convert the content of the columns to a string.
 +
df["name"] = df["name"].astype("str")
 +
df["type"] = df["type"].astype("str")
 +
 +
# iterrows creates an iterable object (list of rows)
 +
for index, row in df.iterrows():
 +
# Do something here to add the content of the row to the graph
 +
pass
 +
 +
g.serialize("output.ttl", format="ttl")
 
</syntaxhighlight>
 
</syntaxhighlight>
  
 +
== If you have more time ==
 +
If you have not already you should include some checks to assure that you don't add any empty columns to your graph.
  
 +
If you have more time you can implement DBpedia Spotlight to link the people mentioned in the dataset to DBpedia resources.
 +
You can use the same code example as in the last lab, but you will need some error-handling for when DBpedia is unable to find a match. For instance:
  
==Without Blazegraph==
+
<syntaxhighlight>
If you have not been able to run Blazegraph yet, you can
+
# Parameter given to spotlight to filter out results with confidence lower than this value
instead program SPARQL queries directly with RDFlib.  
+
CONFIDENCE = 0.5
  
For help, look at the link below:  
+
def annotate_entity(entity, filters={"types":"DBpedia:Person"}):
 +
annotations = []
 +
try:
 +
annotations = spotlight.annotate(SERVER, entity, confidence=CONFIDENCE, filters=filters)
 +
    # This catches errors thrown from Spotlight, including when no resource is found in DBpedia
 +
except SpotlightException as e:
 +
print(e)
 +
# Implement some error handling here
 +
return annotations
 +
</syntaxhighlight>
  
[https://rdflib.readthedocs.io/en/4.2.0/intro_to_sparql.html Querying with Sparql]
+
Here we use the types-filter with DBpedia:Person, as we only want it to match with people. You can choose to only implement the URIs in the response, or the types as well. An issue here is that
  
==Useful Readings==
+
== Useful readings ==
*[https://github.com/RDFLib/sparqlwrapper SPARQLWrapper]
+
* [https://github.com/fivethirtyeight/data/tree/master/russia-investigation Information about the dataset]
*[https://rdflib.readthedocs.io/en/4.2.0/intro_to_sparql.html RDFlib - Querying with Sparql]
+
* [https://towardsdatascience.com/pandas-dataframe-playing-with-csv-files-944225d19ff Article about working with pandas.DataFrames and CSV]
 +
* [https://pandas.pydata.org/pandas-docs/stable/reference/frame.html Pandas DataFrame documentation]
 +
* [https://semanticweb.cs.vu.nl/2009/11/sem/#sem:eventType Simple Event Ontology Descripiton]
 +
* [http://motools.sourceforge.net/timeline/timeline.html The TimeLine Ontology Description]
 +
* [https://www.dbpedia-spotlight.org/api Spotlight Documentation]

Latest revision as of 16:16, 1 March 2022

Lab 6: Semantic Lifting - CSV

Topic

Today's topic involves lifting data in CSV format into RDF. The goal is for you to learn how we can convert non-semantic data into RDF as well as getting familiar with some common vocabularies.

Fortunately, CSV is already structured in a way that makes the creation of triples relatively easy.

We will also use Pandas Dataframes which will contain our CSV data in python code, and we'll do some basic data manipulation to improve our output data.

Relevant Libraries - Classes, Functions and Methods and Vocabularies

Libraries

  • RDFlib concepts from earlier (Graph, Namespace, URIRef, Literal, BNode)
  • Pandas: DataFrame, apply, iterrows, astype
  • DBpedia Spotlight

Semantic Vocabularies

You do not have to use the same ones, but these should be well suited.

  • RDF: type
  • RDFS: label
  • Simple Event Ontology (sem): Event, eventType, Actor, hasActor, hasActorType, hasBeginTimeStamp, EndTimeStamp, hasTime, hasSubEvent
  • TimeLine Ontology (tl): durationInt
  • An example-namespace to represent terms not found elsewhere (ex): IndictmentDays, Overturned, Pardoned
  • DBpedia

Tasks

Today we will be working with FiveThirtyEight's russia-investigation dataset. It contains special investigations conducted by the United States since the Watergate-investigation with information about them to May 2017. If you found the last weeks exercice doable, I recommend trying to write this with object-oriented programming (OOP) structure, as this tends to make for cleaner code.

It contains the following columns:

  • investigation
  • investigation-start
  • investigation-end
  • investigation-days
  • name
  • indictment-days
  • type
  • cp-date
  • cp-days
  • overturned
  • pardoned
  • american
  • president

More information about the columns and the dataset here: https://github.com/fivethirtyeight/data/tree/master/russia-investigation

Our goal is to convert this non-semantic dataset into a semantic one. To do this we will go row-by-row through the dataset and extract the content of each column. An investigation may have multiple rows in the dataset if it investigates multiple people, you can choose to represent these as one or multiple entities in the graph. Each investigation may also have a sub-event representing the result of the investigation, this could for instance be indictment or guilty-plea.

For a row we will start by creating a resource representing the investigation. In this example we handle all investigations with the same name as the samme entity, and will therefore use the name of the investigation ("investigation"-column) to create the URI:

name = row["investigation"]

investigation = URIRef(ex + name)
g.add((investigation, RDF.type, sem.Event))

Further we will create a relation between the investigation and all its associated columns. For when the investigation started we'll use the "investigation-start"-column and we can use the property sem:hasBeginTimeStamp:

investigation_start = row["investigation-start"]

g.add((investigation, sem.hasBeginTimeStamp, Literal(investigation_start, datatype=XSD.date)))

To represent the result of the investigation, if it has one, We can create another entity and connect it to the investigation using the sem:hasSubEvent. If so the following columns can be attributed to the sub-event:

  • type
  • indictment-days
  • overturned
  • pardon
  • cp_date
  • cp_days
  • name (the name of the investigatee, not the name of the investigation)

Code to get you started

import pandas as pd
import rdflib

from rdflib import Graph, Namespace, URIRef, Literal, BNode
from rdflib.namespace import RDF, RDFS, XSD

ex = Namespace("http://example.org/")
dbr = Namespace("http://dbpedia.org/resource/")
sem = Namespace("http://semanticweb.cs.vu.nl/2009/11/sem/")
tl = Namespace("http://purl.org/NET/c4dm/timeline.owl#")

g = Graph()
g.bind("ex", ex)
g.bind("dbr", dbr)
g.bind("sem", sem)
g.bind("tl", tl)

df = pd.read_csv("data/investigations.csv")
# We need to correct the type of the columns in the DataFrame, as Pandas assigns an incorrect type when it reads the file (for me at least). We use .astype("str") to convert the content of the columns to a string.
df["name"] = df["name"].astype("str")
df["type"] = df["type"].astype("str")

# iterrows creates an iterable object (list of rows)
for index, row in df.iterrows():
	# Do something here to add the content of the row to the graph 
	pass

g.serialize("output.ttl", format="ttl")

If you have more time

If you have not already you should include some checks to assure that you don't add any empty columns to your graph.

If you have more time you can implement DBpedia Spotlight to link the people mentioned in the dataset to DBpedia resources. You can use the same code example as in the last lab, but you will need some error-handling for when DBpedia is unable to find a match. For instance:

# Parameter given to spotlight to filter out results with confidence lower than this value
CONFIDENCE = 0.5

def annotate_entity(entity, filters={"types":"DBpedia:Person"}):
	annotations = []
	try:
		annotations = spotlight.annotate(SERVER, entity, confidence=CONFIDENCE, filters=filters)
    # This catches errors thrown from Spotlight, including when no resource is found in DBpedia
	except SpotlightException as e:
		print(e)
		# Implement some error handling here
	return annotations

Here we use the types-filter with DBpedia:Person, as we only want it to match with people. You can choose to only implement the URIs in the response, or the types as well. An issue here is that

Useful readings