About taln2x configuration#

Configurating a taln2x project amounts to specifying two types of information, namely:

  • information about the event which underlies the collection of papers (e.g. title, chairs, dates, etc.)

  • information about taln2x options (e.g. which formats to use)

Event information#

Event information is stored in the event.yml file located at the root of the project, and which uses the yaml format. The example event.yml file created when invoking taln2x new illustrates what event information can be used by taln2x:1

abbrev: MYCONF
title: 19e REncontres MYCONF
url: https://gitlab.inria.fr/talnarchives/taln2x
booktitle: Actes des 19e REncontres MYCONF
short_booktitle: Actes de MYCONF
begin: 2017-06-26
end: 2017-06-30
location: Orléans, France
publisher: ATALA
  - UNETELLE, Machine
  - TRUC, Machin
    fullname: "Volume 1 : conférence principale"
    startpage: 1
    order: title
    sessions: long;short;demo
    long: "Papiers longs"
    short: "Papiers courts"
    demo: "Démonstrations"
    accept: ACCEPT_long;ACCEPT_short;ACCEPT_demo

Here are the definition (and uses) of these fields:


string used as a prefix in article file names (and output directories)


string used to refer to the event in output formats


string appearing on PDF proceedings' frontpage


string appearing on PDF proceedings' frontpage


string appearing on PDF proceedings' frontpage

begin, end

dates in ISO-format used in various output formats


string used in various output formats


string appearing on the PDF proceedings


list of strings of the form "Lastname, Firstname" used in various output formats


list of dictionaries mapping a track name with attribute-value pairs. At list a track fullname must be given. Information about paper ordering, page numbering or sessions (cf infra) can be given as well.


Tracks corresponds to separate volumes in the proceedings. They can have specific chairs (optional field), their page numbering can start from a given integer (optional field). They must have a title (which will be appended to the event title on the PDF proceedings front page). Their papers can be ordered by title, authors or input (which means manually, i.e. following the paper order in the articles.csv metadata file).


Tracks can also be devided into so-called sessions. In that case, papers will be grouped by session, and each session will correspond to a separate part in the PDF proceedings. This feature is optional. To use it, one needs:

  • to declare which session a given paper belongs to by setting a specific acceptance keyword in articles.csv, a common practice is to suffix the acceptance keyword with the id of the session (as in the example above, where the session ids are long, short and demo).

  • to activate taln2x's sessions by setting the sessions flag to true in config.toml.


One can further customize the compiled full proceedings (pdf format) by providing the following files:


image located in the project root and used in the frontpage of the PDF proceedings (the default picture is a photo of the glasswork "Roses et Mouettes" located in the Maison Bergeret in Nancy, France ; the picture has been taken by Alexandre Prevot, and is available on Flickr under a CC-BY-SA 2.0 license)


image of the license used for each paper (the default license is CC-BY 4.0), located in the project root


image used in the frontpage of the PDF proceedings (the default logo is that of the ATALA, the French Association for Natural Language Processing), located in the project root


an image containing the logos of the sponsoring institutions (optional, located in the track subdirectory)


a LaTeX file containing credits (optional, located in the track subdirectory)


a LaTeX file containing a short message to appear before the preface (optional, located in the track subdirectory)


a LaTeX file containing the preface (optional, located in the track subdirectory)


a LaTeX file containing the organizing / reviewing / etc. committees (optional, located in the track subdirectory)


a LaTeX file containing a description of the invited talks if any (optional, located in the track subdirectory)


a LaTeX file containing the ISBN reference of the proceedings if any (optional, located in the track subdirectory)

taln2x options#

taln2x options are set by default by using the config.toml file located at the root of the project. This file uses the toml human-readable format providing mainstream data types (e.g. attribute-value pairs, lists, tables, etc.).

taln2x's options can be overwritten by using the command-line interface. For a full list of options, one can invoke:

taln2x --help

If you look at config.toml, you will see the full list of taln2x options(along with their default values):

name         = "MYCONF"         # Project name

in_format    = "csv"            # Input format
out_format   = "pdf"            # Output format
verbose      = 0                # Verbosity level (0..3)
stopwords    = false            # Ignoring stopwords 
ignore_pdf   = true             # Do not use pdf for metadata
sessions     = true             # Use sessions
latex_encode = false            # Latex encode diacritics
dump         = false            # Dump input config

[project.csv]                   # Input data in CSV
id_col       = 0                # Index of paper id
authors_col  = 1                # Index of authors
title_col    = 2                # Index of title
idhal_col    = 3                # Index of idhal (if any)
file_col     = 4                # Index of file name
accept_col   = 7                # Index of status
keywords_col = 8                # Index of keywords

[project.xml]                   # Input data in XML
xml_file     = ""               

[project.zip]                   # Input data in Zip
zip_file     = ""

[project.pdf]                   # PDF output
tex_log      = false            # Keep latex logs
img          = ""               # User's image directory
language     = "fr"             # Language (cf babel)
halid        = false            # Insert HAL id in paper

[project.acl]                   # ACL output
anthology_id = "myconf"         # Anthology ID 
bilingual    = false            # Bilingual bibtex fields 

[project.hal]                   # HAL output
x_onbehalf   = ""               # Declare deposit owners
stamp        = ""               # Declare HAL collections
include_pdf  = false            # Include PDF 
guess        = false            # Let HAL guess affiliations
dry_run      = false            # Use pre-prod server
update       = false            # Update existing references
national     = false            # Declare national event
no_meta      = false            # Do not update meta-data

Note that in the example configuration above, the default csv structure is the one used by easychair. In case you collected metadata using another abstract management system, or manually, you can set the column numbers according to your own csv file.

  1. The structure of this event.yml file was inspired by the meta file used by the ACL anthology

Documentation built with MkDocs and Terminal for Mkdocs.