input = "<bibitem type='...'>....</bibitem>"
r = Relaton::Render::General.new(template: ..., nametemplate: ..., seriestemplate: ..., language: "en", script: "Latn")
r.render(input) == "Author1 and Author2 2000. Title &c."
bibliography = "<references><bibitem id=''>...</bibitem>...<bibitem id=''>...</bibitem></references>"
r.render_all(bibliography, type: "author-cite") ==
{ bibitem["id"] :
{ author: "Author1 and Author2",
date: "2000a",
citation: "Author1 and Author2 2000a",
formattedref: "Author1 and Author2 2000a. Title &c."
}
}
The gem processes either Relaton XML, or native Relaton classes.
The gem is intended to be inherited from by Metanorma flavours, which may specialise it with their own code. The templates are however intended to determine much of the rendering.
Gems can provide their configurations in YAML files, and parse them before passing them to the call to Iso690Render.
The built-in values for Iso690Render are given in /lib/iso690render/config.yml
, and can be overridden by
the parameters of initialising Iso690Render.
The parameters are:
language
-
in ISO-639
script
-
in ISO-15124
template
-
templates for rendering different bibliographic types
nametemplate
-
templates for rendering personal names
authorcitetemplate
-
templates for rendering names for the purpose of author-date citations
seriestemplate
-
template for rendering series
journaltemplate
-
template for rendering journals in article citations
extenttemplate
-
templates for rendering extents
sizetemplate
-
templates for rendering sizes
edition_ordinal
-
override formatting of edition with ordinal
date
-
default date format (from Twitter CLDR)
i18nhash
-
Metanorma internationalisation hash
Given an Relaton bibitem object (or equivalent Relaton XML), render(input)
will output a formatted reference for that
bibitem, following the rules laid out in the configuration. That output can be inserted into the <formattedref>
element
of the bibitem by the calling code, and used as the authoritative rendering of the citation.
This method does not address how the bibliographic item is to be cited, and does not differentiate citations with the same author and date. It should not be used for rendering of references aligned to the author-date citation system.
Given a collection of Relaton bibitem objects (or equivalent Relaton XML, wrapped in <references>
),
render_all(bibliography, type: type)
will output a hash of objects, with information to be used both for rendering references, and for generating citations.
This method addresses the disambiguation of citiations, and changes to rendering of references reflecting that disambiguation.
For that reason, this method should be used for references using the author-date citation system.
The type
argument of render_all
reflects the citation system to be used.
-
If it is set to "author-date", citations with the same author and date will be disambiguated from each other by appending a letter to the date, in the order in which they occur in
bibliography
. (That means that any fixed sorting of references needs to be applied before the bibliography is presented to the method.) -
If it is set to nil (as is the default), no such disambiguation occurs.
The hash maps from the anchor of each bibitem (bibitem/@id
), to an object containing the following fields:
formattedref
-
the formatted reference, as output by
render(input)
, but with the date of the citation disambiguated if called withtype: "author-date"
. citation
-
the citation for the reference (Author-Date if called with
type: "author-date"
;docidentifier
if called withtype: nil
). author
-
the author designation for the reference (if called with
type: "author-date"
) date
-
the date designation for the reference, with disambiguating letter if necessary (if called with
type: "author-date"
). The author and date are differentiated so that consumers can separate them in citations — e.g. "Jones (1996) claims", "(Jones 1996, 1997)".
There is one template provided for each of the bibliographic types recognised by Relaton (/bibitem/@type
), and a default template:
Currently supported:
-
article
-
book
-
booklet
-
manual
-
proceedings
-
presentation
-
thesis
-
techreport
-
standard
-
unpublished
-
electronic resource
-
inbook
-
incollection
-
inproceedings
-
journal
-
website
-
webresource
-
dataset
Not currently supported:
-
map
-
audiovisual
-
film
-
video
-
broadcast
-
software
-
graphic_work
-
music
-
performance
-
patent
-
archival
-
social_media
-
alert
-
message
-
conversation
-
misc (default)
In Metanorma, not all types are used, but there are exemplars for all of these given on this site, following the human-readable style used in ISO 690. These can be overridden by supplying corresponding paramerers in the call to initialise Iso690Render.
Each template
is a string marked up with Liquid Markup, with the following fields
drawn from the bibliographic item:
Field | Relaton XPath | Multiple | Can come from host | Note |
---|---|---|---|---|
title |
./title |
If multiples, prioritise language match |
||
edition |
./edition |
Y |
If numeric value, is given internationalised rendering of "nth edition", as set in edition_numbering. Otherwise, the textual content of the tag is given. |
|
edition_raw |
./edition |
Y |
The strict textual content of the tag is given. |
|
edition_num |
./edition[@number] |
Y |
||
medium |
./medium |
Y |
||
place |
./place |
Y |
||
publisher |
./contributor[role/@type = 'publisher']/organization/name |
Y |
||
distributor |
./contributor[role/@type = 'distributor']/organization/name |
Y |
||
authorizer |
./contributor[role/@type = 'authorizer']/organization/name | ./contributor[role/@type = 'publisher']/organization/name |
Y |
||
authoritative_identifier |
./docidentifier[not(@type = 'metanorma' or @type = 'ordinal' or @type = 'ISBN' or @type = 'ISSN' or @type = 'DOI')] |
Y |
||
other_identifier |
./docidentifier[@type = 'ISBN' or @type = 'ISSN' or @type = 'DOI'] |
Y |
By default, each such identifier is prefixed with its type and colon |
|
doi |
./docidentifier[@type = 'DOI'] |
Y |
No prefix supplied |
|
status |
./status |
Rendering varies by flavour |
||
uri |
./uri[@type = 'citation' or @type = 'uri' or @type = 'src' or true] |
If multiples, prioritise language match. Always exclude DOI: that is not where the resource is available from |
||
access_location |
./accessLocation |
Y |
||
extent |
./extent |
Y |
Render with standard abbreviations for pp, vols, with n-dash, with delimiting of multiple locations |
|
creatornames |
./contributor[role/@type = 'author'] | ./contributor[role/@type = 'performer'] | ./contributor[role/@type = 'adapter'] | ./contributor[role/@type = 'translator'] | ./contributor[role/@type = 'editor'] | ./contributor[role/@type = 'distributor'] | ./contributor[role/@type = 'authorizer'] | ./contributor |
Y |
|
|
authorcite |
./contributor[role/@type = 'author'] | ./contributor[role/@type = 'performer'] | ./contributor[role/@type = 'adapter'] | ./contributor[role/@type = 'translator'] | ./contributor[role/@type = 'editor'] | ./contributor[role/@type = 'distributor'] | ./contributor[role/@type = 'authorizer'] | ./contributor |
Y |
|
|
role |
./contributor[role/description] | ./contributor[role/@type] |
|||
date |
./date[@type = 'issued'] | ./date[@type = 'circulated'] | ./date |
Y |
Always truncated to just year |
|
date_updated |
./date[@type = 'updated'] |
Y |
||
date_accessed |
./date[@type = 'accessed'] |
Y |
||
series |
./series[@type = 'main' or not(@type) or true] |
Y |
|
|
host_creatornames |
./relation[@type = 'includedIn']/ bibitem/contributor[role/@type = 'author'] |
Y |
Follows options for |
|
host_title |
./relation[@type = 'includedIn']/ bibitem/title |
Y |
Y |
Follows options for |
host_role |
./relation[@type = 'includedIn']/ bibitem/contributor[role/description] | ./relation[@type = 'includedIn']/ bibitem/contributor[role/@type] |
Y |
||
type |
./@type |
labels |
Missing dates and places of publication are rendered as "n.d." and "n.p." or the equivalent internationalisation (no_date
, no_place
in the internationalisation YAML files.) However, missing dates are left as nil in standards, as undated standards indicate that the citation applies to the latest version of the standard, and not that the date is unknown. Missing dates are also left as nil in webresources and websites, since they are continuously updated.
Many fields are populated either by the description of the bibliographic item itself, or by the description of the item containing it (the host item: ./relation[@type = 'includedIn']/bibitem
). For example, in a paper included in an edited volume, the edition will typically be given for the editor volume, rather than for the paper. Those fields are indicated by "Can come from host" in the table.
The Liquid templates use the filters defined in Liquid, such as upcase
. We have defined some custom filters:
-
capitalize_first
capitalises only the first word in a string, and does not lowercase other words in the string. So "third edition" becomes "Third edition", but "3. Aufl." does not become "3. aufl."
The Liquid template surrounds each field by preceding and following punctuation.
-
Fields are space-delimited. So
<em>{{ title }}</em> [{{medium}}]
are two separate fields. -
If fields are not space-delimited, this is indicated by inserting
|
. So{{ title }}|{{ medium}}
is two fields, rendered with no space separation. -
If the field is empty, its surrounding markup is also removed. So if there is no medium, then
[{{medium}}]
is not rendered, and the brackets will be stripped. -
Underscore is treated as space, attaching to the preceding or following field. So
,{{ edition }}{{ labels['edition'] }}
is treated as the one field. -
Underscore is escaped by \. So
<span_class="std\_note">
maps to<span class="std_note">
. -
If punctuation is space delimited, it is inserted regardless of preceding content. So
{{ creatornames }} ({{date}}) .
will insert the full stop whether or not the date is present. -
Space between punctuation and before punctuation is automatically removed.
-
Spaces within fields are globally converted to underscores. For that reason, any filter operations in Liquid need to refer to underscore instead of space.
-
There are primary and secondary quotation marks defined as labels, and subject to internationalisation:
{{ labels['qq-open'] }}
,{{ labels['qq-close'] }}
,{{ labels['q-open'] }}
,{{ labels['q-close'] }}
. By default in Latn and Cyrl, these are<em>
/</em>
and empty, respectively; they are 《…》 , 〈…〉 in Hans, and wavy underline, empty in Hant. If these are used, they need not to be space-delimited from what they quote; e.g.{{ labels['qq-open'] }}{{ title }}{{ labels['qq-close'] }}
.
For example:
"{{ creatornames }} ({{date}}) . <em>{{ title }}</em> [{{medium}}] ,_{{ edition }}_{{ labels['edition'] }} ."
If a type uses another type’s template, the type is mapped to the other type’s name; e.g.
template: book: ... booklet: book
The nametemplate
is a hash of Liquid templates for the formatting of contributor names in particular positions. It
draws on the following fields drawn from the bibliographic item:
Field | Relaton XPath | Multiple | Note |
---|---|---|---|
surname[0] |
./contributor[1]/person/name/surname | ./contributor[1]/person/name/completename |
i.e. surname is the name default |
|
surname[1] |
./contributor[2]/name/surname |
||
surname[2] |
./contributor[3]/name/surname |
||
initials[0] |
./contributor[1]/name/formatted-initials | ./contributor[1]/name/forename/@initial |
If not supplied, the first letter of each given name is used instead |
|
initials[1] |
./contributor[2]/name/initial |
||
given[0] |
./contributor[1]/name/forename[1] |
If not supplied, initials are used instead |
|
given[1] |
./contributor[2]/name/forename[1] |
||
middle[0] |
./contributor[1]/name/forename[not(first())] |
Y |
|
middle[1] |
./contributor[2]/name/forename[not(first())] |
Y |
|
nonpersonal[0] |
./contributor[1]/organization/name |
Y |
|
nonpersonal[1] |
./contributor[2]/organization/name |
Y |
The formatted-initials
field is presumed to contain full stops, and so do the surrogates of that
field done by using individual forenames' initial
attributes, or the forename initials.
Initials are considered delimited by a full stop followed by space or an alphabetic character.
If the full stops are to be stripped, as is often required by bibliographic styles, that needs to occur within the
Liquid template. Bibliographic styles also govern whether initials are separated by space; this gem treats full stop,
not space, as the initials delimiter.
(So D. X. is two initials, as is D.X., but M.-J. is a single initial, and so is de S.)
There are at least three distinct nametemplate
instances that need to be provided, one for a single contributor (one:
), one for two contributors (two:
), one for three or more (more:
), and optionally one for "et al." (etal:
). The number of contributors for which "et al." starts being used is indicated by etal_count
.
For example:
{ one: "{% if nonpersonal[0] %}{{ nonpersonal[0] }}{% else %}{{ surname[0] }}, {{ given[0] }} {{ middle[0] | slice : 0 }}{% endif %}", two: "{% if nonpersonal[0] %}{{ nonpersonal[0] }}{% else %}{{ surname[0] }}, {{ given[0] }} {{ middle[0] | slice : 0 }}{% endif %} & {% if nonpersonal[1] %}{{ nonpersonal[1] }}{% else %}{{ given[1] }} {{ middle[1] | slice : 0 }} {{ surname[1] }}{% endif %}", more: "{% if nonpersonal[0] %}{{ nonpersonal[0] }}{% else %}{{ surname[0] }}, {{ given[0] }} {{ middle[0] | slice : 0 }}{% endif %}, {% if nonpersonal[1] %}{{ nonpersonal[1] }}{% else %}{{ given[1] }} {{ middle[1] | slice : 0 }} {{ surname[1] }}{% endif %} & {% if nonpersonal[2] %}{{ nonpersonal[2] }}{% else %}{{ given[2] }} {{ middle[2] | slice : 0 }} {{ surname[2] }}{% endif %}", etal: "{% if nonpersonal[0] %}{{ nonpersonal[0] }}{% else %}{{ surname[0] }}, {{ given[0] }} {{ middle[0] | slice : 0 }}{% endif %}, {% if nonpersonal[1] %}{{ nonpersonal[1] }}{% else %}{{ given[1] }} {{ middle[1] | slice : 0 }} {{ surname[1] }}{% endif %} <em>et al.</em>", etal_count: 6 }
In the case of more
, the (name)[1]
entries are repeated for all additional authors above 2 and before the final author.
The behaviour of et al. can be specified as follows:
-
etal_count
: the number of authors to trigger et al. in bibliographic rendering -
etal_display
: how many authors to show if using et al. in bibliography (by default, same asetal_count
)
So the Chicago Manual of Style behaviour:
For more than ten authors (not shown here), list the first seven in the reference list, followed by et al.
is realised with etal_count = 10, etal_display = 7
The authorcitetemplate
is a subclass of the name template, configured for rendering author names for author-date citations.
That means that it typically selects only surnames for rendering.
The behaviour of et al. in author-date citations can be specified as follows:
-
etal_count
: the number of authors to trigger et al. in bibliographic rendering -
etal_display
: how many authors to show if using et al. in bibliography (by default, same asetal_count
)
The seriestemplate
is a template for the rendering of series information. It draws on the following fields drawn from the bibliographic item:
Field | Relaton XPath | Multiple | Can come from host | Note |
---|---|---|---|---|
series_title |
./series[@type = 'main' or not(@type) or true]/name |
Y |
||
series_abbr |
./series[@type = 'main' or not(@type) or true]/abbreviation |
Y |
||
series_num |
./series[@type = 'main' or not(@type) or true]/number |
Y |
||
series_partnumber |
./series[@type = 'main' or not(@type) or true]/partnumber |
Y |
||
series_run |
./series[@type = 'main' or not(@type) or true]/run |
Y |
||
series_place |
./series[@type = 'main' or not(@type) or true]/place |
Y |
||
series_organization |
./series[@type = 'main' or not(@type) or true]/organization |
Y |
||
series_dates |
./series[@type = 'main' or not(@type) or true]/from, ./series[@type = 'main' or not(@type) or true]/to |
Y |
For example: {% if series_abbr %}{{series_abbr}}{% else %}{{series_title}}{% endif %} ,_({{series_run}}) {{series_num}}|({{series_partnumber}})
The journaltemplate
is a template for the rendering of series information, when they relate to articles in a journal. The template is distinct because of longstanding practice of rendering journal information differently from monograph series information. The template draws on the same fields as the seriestemplate
, but because the journal title is typically italicised and the numeration is not, any italicisation needs to occur within the template.
For example, the recommended practice in the current edition of ISO 690 is to give explicit volume labels:
<em>{% if series_abbr %}{{series_abbr}}{% else %}{{series_title}}{% endif %}</em> {{ labels['volume'] }}_{{series_num}} {{ labels['part'] }}_{{series_partnumber}}
A common template that drops those labels is:
<em>{% if series_abbr %}{{series_abbr}}{% else %}{{series_title}}{% endif %}</em> {{series_num}}|({{series_partnumber}})
The extent of a bibliographic item may be expressed differently depending on the type of bibliographic item. For example, the extent of a book chapter may be expressed as pp. 9–20, while the extent of an article may be expressed as just 9–20.
To capture this, a separate template is supplied under extenttemplate
for each bibliographic item type. For those types where none is supplied, the template given for misc
is used as the default.
The template draws on the defined types of locality of extents; the most common of these is volume
, issue
(within volume; "number" for journals), and page
. Locality types are the fields used in the Liquid templates; for example:
{ article: "{{ volume_raw }}|({{ issue_raw }}) : {{ page_raw }}" misc: "{{ volume }}, {{ page }}" }
The internationalisation files define a singular and a plural version of the locality types, under labels['extent']
.
-
The plural label is always used if the extent is a range (with a
<from>
and<to>
). -
The singular label is used if the extent is not a range (pp. 2–4 vs. p. 3).
-
The internationalisation files include a slot where the number or number range is inserted, indicated by
%
, since this varies by language. (For instance, English haspp. %
, whereas Chinese has第%页
.) -
The number of the volume, issue/number, or page, without accompanying labels, is given in
volume_raw
,issue_raw
, andpage_raw
.
The size of a bibliographic item is distinct from the extent: the size is how large the item is (e.g. how many pages are in the book), whereas the extent is how much of the host item the item covers (e.g. which pages of the book are in the current chapter.) They can be displayed quite differently from extent; for example, while extent pages is given in English as pp. 9–20 or p. 3, size pages is given as 3 pp..
To capture this, a separate template is supplied under sizetemplate
for each bibliographic item type. Again, for those types where none is supplied, the template given for misc
is used as the default.
The template draws on the defined types of locality of extents; the following are currently recognised:
Field | Relaton XPath | Note |
---|---|---|
volume |
./medium/size[@type = 'volume'] |
With internationalisation of label |
volume_raw |
./medium/size[@type = 'volume'] |
|
issue |
./medium/size[@type = 'issue'] |
With internationalisation of label |
issue_raw |
./medium/size[@type = 'issue'] |
|
page |
./medium/size[@type = 'page'] |
With internationalisation of label |
page_raw |
./medium/size[@type = 'page'] |
|
data |
./medium/size[@type = 'data'] |
Unit of size is included in value |
duration |
./medium/size[@type = 'time'] |
Expressed in ISO 8601 duration |
Locality types are the fields used in the Liquid templates; for example:
{ dataset: "{{ data }}" misc: "{{ volume }}, {{ page }}, {{ data }}, {{ duration }}" }
The internationalisation files define a singular and a plural version of the locality types, under labels['size']
.
-
The plural label is always used if the extent is a range (with a
<from>
and<to>
). -
The label is singular only if the value is
1
, else it is plural (1 p., 2 pp.) -
Again, the internationalisation files include a slot where the number or number range is inserted, since this varies by language.
-
The number of volumes or pages, without accompanying labels, is given in
volume_raw
andpage_raw
. -
Multiple spans of the same type are joined by
+
; e.g. xlii + 76 pp..
In addition, the configuration includes different configuration options for rendering:
The internationalisation file sets the following variables, which can be overridden in configuration parameters:
edition_number
:: has following values corresponding to the rule-based number rules defined in Twitter CLDR.
for a language. For example, English 4th is defined as ["OrdinalRules", "digits-ordinal"]
, because under twitter-cldr, 4th
is generated as 4.localize(:en).to_rbnf_s("OrdinalRules", "digits-ordinal")
. If missing, the raw number is given.
edition
:: is the localised expression for edition, with the edition number given as %. So _4th ed.` is generated with edition
as % ed.
.
date
:: date format default, taken from Twitter CLDR: to_full_s
, to_long_s
, to_medium_s
, to_short_s
, or one of the to_additional_s
formats. One value is given for each of "month_year", "day_month_year", and "date_time"; e.g. { month_year: to_long_s, day_month_year: to_long_s, date_time: to_long_s }
.