National Corpus of Polish

The National Corpus of Polish is a shared initiative of four institutions: Institute of Computer Science at the Polish Academy of Sciences (coordinator), Institute of Polish Language at the Polish Academy of Sciences, Polish Scientific Publishers PWN, and the Department of Computational and Corpus Linguistics at the University of Łódź. It has been carried out as a research-development project of the Ministry of Science and Higher Education.

These four institutions have started cooperation to build a reference corpus of Polish language containing over fifteen hundred millions of words. The list of sources for the corpora contains classic literature, daily newspapers, specialist periodicals and journals, transcripts of conversations, and a variety of short-lived and internet texts. For a corpus to be reliable, not only it is necessary to contain a high number of words, but it also needs a diversity of texts with respect to the subject and genre. The conversations ought to represent both male and female speakers, in various age groups, coming from various regions in Poland.

Official project website

Manually annotated subcorpus

The manually annotated 1-million word subcorpus of the National Corpus of Polish, available on CC-BY licence (NKJP tagset):

• NKJP-PodkorpusMilionowy-1.2.tar.gz — GZip compressed tar archive,

• NKJP-PodkorpusMilionowy-1.2.tar.bz2 — BZip2 compressed tar archive.

XCES encoded version

The manually annotated subcorpus, encoded using the XCES format (NKJP tagset).

• nkjp1m-1.2-xces-xml.bz2 — XCES encoded corpus; here you may find the order of the source directories in the combined file,

• nkjp1m-1.2-xces-reanalysed-sgjp-xml.bz2 — XCES encoded corpus, reanalysed according to the procedure described here, using Morfeusz SGJP,

• nkjp1m-1.2-xces-reanalysed-polimorf-xml.bz2 — XCES encoded corpus, reanalysed according to the procedure described here, using Morfeusz Polimorf.

SGJP tagset version

The manually annotated subcorpus, converted to the newest version of the SGJP tagset.

• Newest version is automatically created after any changes to the SGJP tagset are made. This version is released at the SGJP website.

Tagger models, trained on the latest version of the subcorpus

• Model for Concraft 2.0 is available at its official GitHub page.

• concraft-model-nkjp1m-1.2.gz — model for Concraft 0.7.1 tagger,

• pantera-model-nkjp1m-1.2.bz2 — model for PANTERA 0.9.1 tagger,

• wcrft-model-nkjp-1.2.tar.gz — model for WCRFT1 and WCRFT2 taggers,

• wmbt-model-nkjp1m-1.2.tar.bz2 — model for WMBT tagger.

The training data for all models above has been preprocessed using Morfeusz SGJP morphological analyzer.

Additional resources

• Distributable versions of NKJP,

• N-grams extracted from balanced National Corpus of Polish,

• Economy-related subcorpus of the National Corpus of Polish, with manually created sense annotation layer,

• A java library for parsing NKJP-compatible TEI P5 files,

• NKJP-based semantic space,

• Description of the TEI P5 encoding and schemata,

• Newly (2020) tagged 300M subcorpus.

Reporting errors

Please use the linked form to report any errors found in the corpus.

Older files

For reference and comparison purposes, older versions of the manually annotated subcorpus are available:

• NKJP-PodkorpusMilionowy-1.1.tgz,

• NKJP-PodkorpusMilionowy-1.0.tgz.