Category: Tutorial

Day 18: Merry Christmas!


Thank you all for joining our Catmandu advent calendar this month. We hope that you enjoyed our daily posts. Catmandu is a very rich programming environment which provides command line tools and even an API. In these blogposts we provided only a short introduction into all these modules. Hopefully we will see you next year again with more examples!

The Catmandu community consists of all people involved in the project, no matter if they do programming, documentation, or drawing cats. We want to thank them all for a wonderful year!

  • Christian Pietsch
  • Dave Sherohman
  • Friedrich Summann
  • Jakob Voss
  • Johann Rolschewski
  • Jörgen Eriksson
  • Maria Hedberg
  • Mathias Lösch
  • Najko Jahn
  • Nicolas Franck
  • Nicolas Steenlant
  • Patrick Hochstenbach
  • Petra Kohorst
  • Snorri Briem
  • Vitali Peil

And a big round of applause for our contributors who kept us sending bug reports and ideas for new features. If you would like to contribute, then please take a look at the contributions section of Catmandu documentation. Don’t be shy to contact us with questions, feature requests, bug fixes, documentation and cat cartoons!

This advent calendar will stay online for you reference.

As a special gift we have still some catmandu USB sticks available that we can send to you. Please send a line to “patrick dot hochstenbach at ugent dot be”. The first 5  emailers will get a free USB!



Day 17: Exporting RDF data with Catmandu

Yesterday we learned how to import RDF data with Catmandu. Exporting RDF can be as easy as this:

catmandu convert RDF --url to RDF

By default, the RDF exporter Catmandu::Exporter::RDF emits RDF/XML, an ugly and verbose serialization format of RDF. Let’s configure catmandu to use the also verbose but less ugly NTriples. This can either by done by appending --type ntriple on command line or by adding 17_librecatthe following to config file catmandu.yml:

    package: RDF
        type: ntriples

The NTriples format illustrates the “true” nature of RDF data as a set of RDF triples or statements, each consisting of three parts (subject, predicate, object).

Catmandu can be used for converting between one RDF serialization format to another, but more specialized RDF tools, such as such rapper are more performant especially for large data sets. Catmandu can better help to process RDF data to JSON, YAML, CSV etc. and vice versa.

Let’s proceed with a more complex workflow and with what we’ve learned at day 13 about OAI-PMH and another popular repository: There is a dedicated Catmandu module Catmandu::ArXiv for searching the repository, but ArXiv also supports OAI-PMH for bulk download. We could specify all options at command line, but putting the following into catmandu.yml will simplify each call:

    package: OAI
      metadataPrefix: oai_dc
      set: cs

Now we can harvest all computer science papers (set: cs) for a selected day (e.g. 2014-12-19):

$ catmandu convert arxiv --from 2014-12-19 --to 2014-12-19 to YAML

The repository may impose a delay of 20 seconds, so be patient. For more precise data, we better use the original data format from ArXiV:

$ catmandu convert arxiv --set cs --from 2014-12-19 --to 2014-12-19 --metadataPrefix arXiv to YAML > arxiv.yaml

The resulting format is based on XML. Have a look at the original data (requires module Catmandu::XML):

$ catmandu convert YAML to XML --field _metadata --pretty 1 < arxiv.yaml
$ catmandu convert YAML --fix 'xml_simple(_metadata)' to YAML < arxiv.yaml

Now we’ll transform this XML data to RDF. This is done with the following fix script, saved in file arxiv2rdf.fix:




The following command generates one RDF triple per record, consisting of an arXiv article identifier, the property and the article title:

$ catmandu convert YAML to RDF --fix arxiv2rdf.fix < arxiv.yaml

To better understand what’s going on, convert to YAML instead of RDF, so the internal aREF data structure is shown:

$ catmandu convert YAML to YAML --fix arxiv2rdf.fix < arxiv.yaml

dc_title: On Conditional Decomposability

This record looks similar to the records imported from RDF at day 13. The special field _id refers to the subject in RDF triples: a handy feature for small RDF graphs that share the same subject in all RDF triples. Nevertheless, the same RDF graph could have been encoded like this:

  dc_title: On Conditional Decomposability

To transform more parts of the original record to RDF, we only need to map field names to prefixed RDF property names. Here is a more complete version of arxiv2rdf.fix:

unless exists(dc_creator.0)
do list(path=>dc_creator)
  join_field(foaf_name,' ')

The result is one big RDF graph for all records:

$ catmandu convert YAML to RDF --fix arxiv2rdf.fix < arxiv.yaml

Have a look at the internal aREF format by using the same fix with convert to YAML and try conversion to other RDF serialization forms. The most important part of transformation to RDF is to find matching RDF properties from existing ontologies. The example above uses properties from Dublin Core, Creative Commons, Friend of a Friend,, and Bibliographic Ontology.

Continue to Day 18: Merry Christmas! >>

Day 16: Importing RDF data with Catmandu

16_librecatA common problem of data processing is the large number of data formats, dialects, and conceptions. For instance the author field in one record format may differ from a similar field another format in its meaning or name. As shown in the previous articles, Catmandu can help to bridge such differences, but it can also help to map from and to data structured in a completely different paradigm. This article will show how to process data expressed in RDF, the language of Semantic Web and Linked Open Data.

RDF differs from previous formats, such as JSON and YAML, MARC, or CSV in two important aspects:

  1. There are no records and fields: RDF data instead is a graph structure, build of nodes (“resources” or “values”) and directed links.
  2. Link types (“properties”) are identified by URI and defined in “ontologies”. In theory this removes the introductory common problem of data processing.

Because graph structures are fundamentally different to record structures, there is no obvious mapping between RDF and records in Catmandu. For this reason you better use dedicated RDF technology as long as your data is RDF. Catmandu, however, can help to process from RDF and to RDF, as shown today and tomorrow, respectively. Let’s first install the Catmandu module Catmandu::RDF for RDF processing:

$ cpanm –sudo Catmandu::RDF

If you happen to use this on a virtual machine from the Catmandu USB stick, you may first have to update another module to remove a nasty bug (the password is “catmandu”):

$ cpanm –sudo List::Util

You can now retrieve RDF data from any Linked Open Data URI like this:

$ catmandu convert RDF –url to YAML

We could also download RDF data into a file and parse the file with Catmandu afterwards:

$ curl -L -H 'Accept: application/rdf+xml' > rdf.xml
$ catmandu convert RDF --type rdfxml to YAML < rdf.xml
$ catmandu convert RDF --file rdf.xml to YAML # alternatively

Downloading RDF with Catmandu::RDF option --url, however, is shorter and adds an _url field that contains the original source. The RDF data converted to YAML with Catmandu looks like this (I removed some parts to keep it shorter). The format is called another RDF Encoding Form (aREF) because it can be transformed from and to other RDF encodings:

  dct_title: Frictional Coefficient under Banana Skin@
  - <>
  - <>
  - <>
  - <>
  dct_date:- 2012^xs_gYear
  dct_isPartOf: <>
  a: bibo_Journal
  bibo_issn: 1881-2198@
  dct_title: Tribology Online@
  a: foaf_Person
  foaf_name:Daichi Uchijima@
  foaf_name: Kensei Tanaka@
  foaf_name: Kiyoshi Mabuchi@
  foaf_name: Rina Sakai@

The sample record contains a special field _url with the original source URL and six fields with URLs (or URIs), each corresponding to an RDF resource. The field with the original source URL ( can be used as starting point. Each subfield (dct_title, dct_creator, dct_date, dct_isPartOf) corresponds to an RDF property, abbreviated with namespace prefix. To fetch data from these fields, we could use normal fix functions and JSON path expressions, as shown at day 7 but there is a better way:

Catmandu::RDF provides the fix function aref_query to map selected parts of the RDF graph to another field. Try to get the the title field with this command:

$ catmandu convert RDF –url –fix ‘aref_query(dct_title,title)’ to YAML

More complex transformations should better be put into a fix file, so create file rdf.fix with the following content:


If you apply the fix, there are four additional fields with data extracted from the RDF graph:

$ catmandu convert RDF –url –fix rdf.fix to YAML

The aref_query function also accepts a language, similar to JSON path, but the path is applied to an RDF graph instead of a simple hierarchy. Moreover one can limit results to plain strings or to URIs. For instance the author URIs can be accessed with aref_query(dct_creator.,author). This feature is useful especially if RDF data contains a property with multiple types of objects, literal strings, and other resources. We can aggregate both with the following fixes:

aref_query(dct_creator@, authors)
aref_query(dct_creator.foaf_name@, authors)

Before proceeding you should add the following option to config file catmandu.yaml:

    package: RDF
      ns: 2014091

This makes sure that RDF properties are always abbreviated with the same prefixes, for instance dct for

Continue to Day 17: Exporting RDF data with Catmandu >>

Day 15 : MARC to Dublin Core

13_librecatToday we will look a bit further into MARC processing with Catmandu. By now you should already know how to startup the Virtual Catmandu (hint: see day 1) and start up the UNIX command prompt (hint: see day 2). We already saw a bit of MARC processing in day 9 and today we will show you how to transform MARC records into Dublin Core. This as a preparation to create RDF and Linked Data in the later posts.

First I’m going to teach you how to process different types of MARC files. On the Virtual Catmandu system we provided five  example MARC files. You can find them in your Documents folder:

  • Documents/camel.mrk
  • Documents/camel.usmarc
  • Documents/marc.xml
  • Documents/rug01.aleph
  • Documents/rug01.sample

When you examine these files with the UNIX less command you will see that all the files have a bit different format:

$ less Documents/camel.mrk
$ less Documents/camel.usmarc
$ less Documents/marc.xml
$ less Documents/rug01.sample

There are many ways in which MARC data can be written into a file. Every vendor likes to use its own format. You can compare this with the different ways a text document can be stored: as Word, as Open Office, as PDF and plain text. If we are going to process these files with catmandu, then we need to tell the system what the exact format is.

We will work today with the last example rug01.sample which is a small export out of the Aleph catalog from Ghent University Library. Ex Libris uses a special MARC format to structure their data which is called Aleph sequential. We need to tell catmandu not only that our input file is in MARC but also in this special Aleph format. Let’s try to create YAML to see what it gives:

$ catmandu convert MARC --type ALEPHSEQ to YAML < Documents/rug01.sample

To transform this MARC file into Dublin Core we need to create a fix file. You can use the UNIX command nano for this (hint: see day 5 how to create files with nano). Create a file dublin.fix:

$ nano dublin.fix

And type into nano the following fixes:







Every MARC record contains in the 245-field the title of a record. In the first line we map the MARC-245 field to new field in the record called title:


In the second and third line we map authors to a field creator. In the rug01.sample file the authors are stored in the MARC-100 and MARC-700 field. Because there is usually more than one author in a record, we need to $append them to create an array (a list) of one or more creator-s.

In line 4 and line 5 we do the same trick to filter out the ISBN and ISSN number out of the record which we store in separate fields isbn and issn (indeed these are not Dublin Core fields, we will process them later).

In line 6 and line 7 we read the MARC-260 field which contains publisher and date information. Here we don’t need the $append trick because there is usually only one 260-field in a MARC record.

In line 8 the subjects are extracted from the 260-field using the same $append trick as above. Notice that we only extracted the $a subfields? If you want to add more subfields you can list them as in marc_map(650abcdefgh,subject.$append)

Given the dublin.txt file above we can execute the filtering command like this:

$ catmandu convert MARC --type ALEPHSEQ to YAML --fix dublin.fix < Documents/rug01.sample

As always you can type | less at the end of this command to slow down the screen output, or store the results into a file with > results.txt. Hint:

$ catmandu convert MARC --type ALEPHSEQ to YAML --fix dublin.fix < Documents/rug01.sample | less
$ catmandu convert MARC --type ALEPHSEQ to YAML --fix dublin.fix < Documents/rug01.sample > results.txt

The results should look like this:

_id: '000000002'
- Katz, Jerrold J.
date: '1977.'
- '0855275103 :'
publisher: Harvester press,
- Semantics.
- Proposition (Logic)
- Speech acts (Linguistics)
- Generative grammar.
- Competence and performance (Linguistics)
title: Propositional structure and illocutionary force :a study of the contribution of sentence meaning to speech acts /Jerrold J. Katz.

Congratulations, you’ve created your first mapping file to transform library data from MARC to Dublin Core! We need to add a bit more cleaning to delete some periods and commas here and there but as is we already have our first mapping.

Below you’ll find a complete example. You can read more about our Fix language online.

marc_map(245,title, -join => " ")



replace_all(isbn.," .","")
replace_all(issn.," .","")




Continue to Day 16: Importing RDF data with Catmandu >>

Day 14: Set up your own OAI data service

14_librecatIn the last days you have learned how to store data with Catmandu. Storing data is a cool thing, but sharing data is awesome. Interoperability is important as other people may use your data (and you will profit from other people’s interoperable data)

In the day 13 tutorial we’ve learned the basic principle of metadata harvesting via OAI-PMH.

We will set up our OAI service with the Perl Dancer framework and an easy-to-use plugin called Dancer::Plugin::Catmandu::OAI. To install the required modules run:

$ cpanm Dancer

$ cpanm Dancer::Plugin::Catmandu::OAI

and you also might need

$ cpanm Template

Let’s start and index some data with Elasticsearch as learned in the previous post:

$ catmandu import OAI --url --metadataPrefix oai_dc --set flandrica --handler oai_dc to Elasticsearch --index_name oai --bag publication

After this, you should have some data in your Elasticsearch index. Run the following command to check this:

$ catmandu export Elasticsearch --index_name oai --bag publication

Everything is fine, so let’s create a simple webservice which exposes to collected data via OAI-PMH. The following code can be downloaded from this gist.

What’s going on here? Well, the script defines a route /oai via the plugin Dancer::Plugin::Catmandu::OAI.
The template defines the xml output of the records. And finally the configuration file catmandu.yml handles the settings for the Dancer plugin as well as for the Elasticsearch indexing and querying.

Run the following command to start a local webserver

$ perl

and point your browser to https://localhost:3000/oai?verb=Identify. To get some records go to http://localhost:3000/oai?verb=ListRecords&metadataPrefix=oai_dc.

Yes, it’s that easy. You can extend this simple example by adding fixes to transform the data as you need it.

Continue to Day 15: MARC to Dublin Core >>

Day 13: Harvest data with OAI-PMH

14_librecatThe Open Archives Initiative Protocol for Metadata Harvesting (OAI-PMH) is a protocol to harvest metadata records from OAI compliant repositories. It was developed by the Open Archives Initiative as a low-barrier mechanism for repository interoperability. The Open Archives Initiative maintains a registry of OAI data providers.

Every OAI server must provide metadata records in Dublin Core, other (bibliographic) formats like MARC may be supported additionally. Available metadata formats can be detected with “ListMetadataFormats“. You can set the metadata format for the Catmandu OAI client via the --metadataPrefix parameter.

The OAI server may support selective harvesting, so OAI clients can get only subsets of records from a repository. The client requests could be limited via datestamps (--from, --until) or set membership (--set).

To get some Dublin Core records from the collection of Ghent University Library and convert it to JSON (default) run the following catmandu command:

$ catmandu convert OAI --url --metadataPrefix oai_dc --set flandrica --handler oai_dc

You can also harvest MARC data and store it in a file:

$ catmandu convert OAI --url --metadataPrefix marcxml --set flandrica --handler marcxml to MARC --type USMARC > ugent.mrc

Instead of harvesting the whole metadata you can get the record identifiers (--listIdentifiers) only:

$ catmandu convert OAI --url --metadataPrefix marcxml --set flandrica --listIdentifiers 1 to YAML

You can also transform incoming data and immediately store/index it with MongoDB or Elasticsearch. For the transformation you need to create a fix (see Day 6):

$ nano simple.fix

Add the following fixes to the file:


Now you can run an ETL process (extract, transform, load) with one command:

$ catmandu import OAI --url --metadataPrefix marcxml --set flandrica --handler marcxml --fix simple.fix to Elasticsearch --index_name oai --bag ugent
$ catmandu import OAI ---url --metadataPrefix marcxml --set flandrica --handler marcxml --fix simple.fix to MongoDB --database_name oai --bag ugent

The Catmandu OAI client provides special handler (--handler) for Dublin Core (oai_dc) and MARC (marcxml). For other metadata formats use the default handler (raw) or implement your own. Read our documentation for further details.

Continue to Day 14: Set up your own OAI data service >>

Day 12: Index your data with ElasticSearch

12_librecatElasticSearch is a flexible and powerful open source, distributed, real-time search and analytics engine. You can store structured JSON documents and by default ElasticSearch will try to detect the data structure and index the data. ElasticSearch uses Lucene to provide full text search capabilities with a powerful query language. Install guides for various platforms are available at the ElasticSearch reference. To install the corresponding Catmandu module run:

$ cpanm Catmandu::Store::ElasticSearch
$ cpanm Search::Elasticsearch::Client::5_0::Direct

[For those of you running the Catmandu VirtualBox this installation is not required. ElasticSearch is by default installed]
Now get some JSON data to work with:

$ wget -O banned_books.json

First index the data with ElasticSearch. You have to specify an index (--index_name) and type (--bag), client version (--client) for Elasticsearch versions >= 2.0 you also have to add a prefix to your IDs (--key_prefix):

$ catmandu import -v JSON --multiline 1 to ElasticSearch --index_name books --bag banned --key_prefix my_ --client '5_0::Direct' < banned_books.json

Now you can export all items from an index to different formats, like XLSX, YAML and XML:

$ catmandu export ElasticSearch --index_name books --bag banned --client '5_0::Direct' to YAML
$ catmandu export ElasticSearch --index_name books --bag banned --client '5_0::Direct' to XML
$ catmandu export -v ElasticSearch --index_name books --bag banned --client '5_0::Direct' to XLSX --file banned_books.xlsx

You can count all indexed items or those which match a query:

$ catmandu count ElasticSearch --index_name books --bag banned --client '5_0::Direct'
$ catmandu count ElasticSearch --index_name books --bag banned --client '5_0::Direct' --query 'firstEditionPublicationYear: "1937"'
$ catmandu count ElasticSearch --index_name books --bag banned --client '5_0::Direct' --query 'firstEditionPublicationPlace: "Berlin"'

You can search an index for a specific value and export all matching items:

$ catmandu export ElasticSearch --index_name books --bag banned --client '5_0::Direct' --query 'firstEditionPublicationYear: "1937"' to JSON
$ catmandu export ElasticSearch --index_name books --bag banned --client '5_0::Direct' --query 'firstEditionPublicationPlace: "Berlin"' to CSV --fields 'my_id,authorFirstname,authorLastname,title,firstEditionPublicationPlace'

You can delete whole collections from a database or just items which match a query:

$ catmandu delete ElasticSearch --index_name books --bag banned --client '5_0::Direct' --query 'firstEditionPublicationPlace: "Berlin"'
$ catmandu delete ElasticSearch --index_name books --bag banned --client '5_0::Direct'

Catmandu::Store::ElasticSearch supports CQL as query language. For setup and usage see documentation.

Continue to Day 13: Harvest data with OAI-PMH >>