Create dataset french_treebank

- uid: french_treebank
- type: processed
- description:
- name: French Treebank
- description: A reference treebank for French starting with a tagged newspaper corpus of 1 Million words.
- homepage: http://ftb.linguist.univ-paris-diderot.fr/index.php?langue=en
- validated: True
- languages:
- language_names:
- French
- language_comments:
- language_locations:
- Europe
- Western Europe
- France
- validated: False
- custodian:
- name: Laboratoire de linguistique formelle, Université de Paris.
- in_catalogue:
- type: A university or research institution
- location: France
- contact_name: Anne Abeillé
- contact_email: anne.abeille@linguist.univ-paris-diderot.fr
- contact_submitter: False
- additional: http://ftb.linguist.univ-paris-diderot.fr
- validated: False
- availability:
- procurement:
- for_download: No - but the current owners/custodians have contact information for data queries
- download_url:
- download_email: anne.abeille@linguist.univ-paris-diderot.fr
- licensing:
- has_licenses: Yes
- license_text: Terms and conditions of use for internal research purposes of the French Treebank

In order to get an access to the CORPUS, the User agrees to these terms and conditions of use.

Article 1 – Definitions

1.1 CORPUS refers to: "Le corpus arboré pour le français" French Treebank described in annex and crafted by the Laboratoire de linguistique formelle and owned by the université de Paris.

1.2 Supplier refers to: Université Paris-Diderot – Paris 7 (Laboratoire de linguistique formelle)

Article 2 – Purpose of the agreement

This agreement describes the conditions in which Université Paris-Diderot – Paris 7 (the Supplier) provides the User a nonexclusive, nontransferable and personal right to use the CORPUS, whithout any cost, for research purposes only, valid worldwide. Specifications are described on the website of the Supplier.

Article 3 – Provision of the CORPUS

There is no such thing as a recording medium of the CORPUS. The provision is made possible by the delivery of a personal identification number (P-ID). This P-ID is delivered after having agreed to these terms and conditions of use by the User.

Since the User is given a P-ID by the Supplier, he can download the COPURS within thirty (30) days, and only three (3) times.

Article 4 – Intellectual property

The Supplier is the only owner of the rights attached to the CORPUS, guidelines included. This agreement does not imply the transfer of intellectual property to the User.

The User has a nonexclusive, nontransferable and personal right to use the CORPUS for his own research purposes only.

The User commits himself to never delete or amend the tags, copyright notes or any other distinct mark appended on the CORPUS.

Article 5 – User obligations

The User commits himself to use the CORPUS within the scope of a clear and identified project, and only for internal research purposes.

The User is not allowed to make any copy of the CORPUS except a safety one.

The User is not allowed to translate, adapt, dismantle or uncompile the CORPUS.

The User commits himself to not give up to anyone, even if the transaction is free of charge, the right to use granted by the present agreement.

The User commits himself to not use the CORPUS for any commercial or advertising purposes.

The User commits himself to redistribute to the Supplier any improved or enriched version he could make of the CORPUS. The User will do his best to share his experience of the use of the CORPUS with the Supplier (errors, mistakes…).

Article 6 – Related services

The present agreement does not include any maintenance contract or assistance with the use of the CORPUS.

Yet, the Supplier reachable through the contact form.

When a new version of the CORPUS is distributed, the Supplier will inform the User by email in order to provide him a chance to download it.

Article 7 – Guarantees and limitation of liability

The ressource is distributed with no guarantee and the Supplier is not responsible for the uses that will be made of it.

In no case the Supplier will be held responsible for any damages to the User caused by the use of the CORPUS.

Article 8 – Publications

In all the publications about the results related to the use of the CORPUS, the User has to mention that his studies were produced thanks to the CORPUS. He also commits himself to mention the name of the Supplier as the owner ot the intellectual property rights and make a reference to this publication: « Abeillé, A., L. Clément, and F. Toussenel. 2003. "Building a treebank for French", in A. Abeillé (ed) Treebanks, Kluwer, Dordrecht. (p. 165-187) ».

Article 9 – Duration

This agreement comes into effect at the time the personal identification number is given by the Supplier and is still valid during the whole term of the intellectual property rights that protect the CORPUS.

Article 10 – Disputes – Applicable Law

This agreement is subject to French laws and regulations.

In case of dispute over the reading or the execution of this agreement, the parties will do their best to find an amicable solution.

If this dispute longs more than two (2) months, the parties will have recourse at their initiative to the Court of Paris.

Annex: list of the resources of the CORPUS

Corpus over 21.550 sentences (664.533 tokens) from the newspaper Le Monde (1990-1993) tokenized, tagged and corrected for the compounds, categories and inflectional morphology, lemmas, constituants and grammatical functions; in three formats: original XML, Tiger-XML (searchable with Tiger-search), PTB (depleted);
Corpus of texts segmented into articles with TEI annotations (article, date, author, coverage) related to the files above;
Documentation about the annotated corpus with annotation choices, divided into three guidelines: Morphosyntactic annotations (simple words, compounds), Constituants annotations, Functions annotations.
- license_properties:
- research use
- license_list:
- other: Other license
- pii:
- has_pii: Yes
- generic_pii_likely: very likely
- generic_pii_list:
- names
- physical addresses
- dates (birth, death, etc.)
- numeric_pii_likely: very likely
- numeric_pii_list:
- telephone numbers
- sensitive_pii_likely: very likely
- sensitive_pii_list:
- racial or ethnic origin
- political opinions
- data concerning a person's sex life or sexual orientation
- trade-union membership
- religious or philosophical beliefs
- health-related data
- no_pii_justification_class:
- no_pii_justification_text:
- validated: False
- processed_from_primary:
- from_primary: Taken from primary source
- primary_availability: No - the dataset curators describe the primary sources but they are fully private
- primary_license:
- primary_types:
- validated: False
- from_primary_entries:
- le_monde_newspaper
- media:
- category:
- text
- text_format:
- .TXT
- audiovisual_format:
- image_format:
- database_format:
- .CSV
- text_is_transcribed: No
- instance_type: article
- instance_count: 100