works-algorithm-notes.txt

Algorithm to Generate BibFrame *Works* from MaRC

reversed engineered from LoC xquery implementation of Marc -> BibFrame
  https://github.com/lcnetdev/marc2bibframe

  bf:Works  are created for bib records by:
    modules/module.MBIB-2-BIBFRAME-Shared.xqy


Each Bib Record creates a Work

PLUS

a Work is created when the Bib record has:

600t  # subject
533   # related reproduction
555u  # finding aid

# related works
730|740|780|785
787t
700t|710t|711t
490a
400t|410t|411t|430t|440t|534t|774t|775t|800t|810t|811t|830t
760|762|765|767|770|772|773|774

505 ind2==0, has |t  # contents notes

130|240  # uniform title
880 for 130|240  vernacular uniform title

================================================================================
Naomi's notes below this line


Looking for code that outputs bf:Work

# Each Bib Record
function mbshared:generate-work
  2607:    let $mainType := "Work"
  2780:        element {fn:concat("bf:" , $mainType)} {


# 600t
function mbshared:get-subject
  # It takes a specific 6xx as input.
  # It generates a bf:subject as output, which can have a bf:subject/bf:Work if 600t
  3018:    let $subjectType:= if ($d[@tag="600"][marcxml:subfield[@code="t"]]) then "Work" else $subjectType
  3148:          element {fn:concat("bf:",$subjectType)} {

  Called by:
    2744:             return mbshared:get-subject($d)
    from  function mbshared:generate-work


# 533 related repro
function mbshared:generate-related-reproduction
  # 533 to reproduction

  2205:     element bf:Work{

  Called by:
    2504:               return mbshared:generate-related-reproduction($d,$type)
    from  function mbshared:related-works

    Called by:
      # for instances - doesn't produce works
      745:   let $instance-relateds := mbshared:related-works($d/ancestor::marcxml:record,$workID,"instance")

      2751:   let $work-relateds := mbshared:related-works($marcxml,$workID,"work")
      from  function mbshared:generate-work


# 555u  finding aid
function mbshared:generate-finding-aid-work
  # 555 finding aids note may be related work link or a simple property

  2294:        element bf:Work{

  Called by:
    2748:                                 mbshared:generate-finding-aid-work($d)
    from  function mbshared:generate-work


# related work
# 730|740|780|785
# 787t
# 700t|710t|711t
# 490a
# 400t|410t|411t|430t|440t|534t|774t|775t|800t|810t|811t|830t
# 760|762|765|767|770|772|773|774
function mbshared:generate-related-work
  #  For RDA:   040$e = rda
  #  For AACR2:  Leader/18 = a
  #  Under AACR2, when two works were published together the first work in the compilation was given the 1XX/240, and the second work was given a 700 analytic (name/title).  This essentially resulted in identifying the aggregate work by only the first work in the compilation.
  #  Under RDA, we identify the aggregate work in the 240 (not just one of the works), and provide analytical added entries (name/title) for the works in the compilation.
  #  (245 would be the instance title, 240 the UT)

  2388:   element bf:Work {

  Called by:
    2500:  return mbshared:generate-related-work($d,$type, $workID)   ) # 730|740|780|785
    2508:  return      mbshared:generate-related-work($d,$type, $workID)  # 787t
    2512:  return      mbshared:generate-related-work($d,$type, $workID)  # 700t|710t|711t
    2516:  return mbshared:generate-related-work($d,$type, $workID)             # 490a
    2520:  return mbshared:generate-related-work($d,$type, $workID) # 400t|410t|411t|430t|440t|534t|774t|775t|800t|810t|811t|830t
    2526:  mbshared:generate-related-work($d,$type, $workID)  # 760|762|765|767|770|772|773|774

    all from  mbshared:related-works

    Called by
      # for instances - doesn't produce works
      745:   let $instance-relateds := mbshared:related-works($d/ancestor::marcxml:record,$workID,"instance")

      2751:   let $work-relateds := mbshared:related-works($marcxml,$workID,"work")
      from  function mbshared:generate-work


# 505 ind2==0, has |t  contents notes
function mbshared:generate-complex-notes
  # This function generates  constituent (hasPart) works from 505
  #   LC often uses only $a, but stanford has  $g$t, and dave reser says he's seen [$g]$t[$r][$u] and $u is always last.
  # It generates a bf:hasPart/bf:Work

  2901:                         element bf:Work {

  Called by
    2755:          return (mbshared:generate-complex-notes($d),
    from  function mbshared:generate-work


# 880 for 130|240  vern uniform title
function mbshared:generate-translationOf
  # This function generates a related work (rda expression?), as translation of from the 100, 240.
  # It takes a 130 or 240 element.
  # It generates a bf:translationOf/bf:Work

  3680:            element bf:Work {

  Called by
    3871:         mbshared:generate-translationOf($d)
    from  function mbshared:get-uniformTitle

    Called by
      2611:            return mbshared:get-uniformTitle($d)
      from   function mbshared:generate-work



# 130|240  uniform title
function mbshared:get-uniformTitle
  # This function generates a Work based on the uniformTitle.
  # It takes a specific datafield (130 or 240) as input.
  # It generates a bf:Work as output.

  3888:        element bf:Work {

  Called by
    2611:            return mbshared:get-uniformTitle($d)
    from   function mbshared:generate-work








NOTE:  switches:
-----------------

 to generate instances from secondary ISBNs


declare variable $mbshared:transform-rules :=(
<rules>
  <rule status="off" id="1" label="isbn" category="instance-splitting">New instances on secondary unique ISBNs</rule>
  <rule status="on" id="2" label="issn" category="instance-splitting">New instances on secondary unique ISSNs</rule>
  <rule status="on" id="3" label="260" category="instance-splitting">New instances on multiple 260s (not serials)</rule>
  <rule status="on" id="4" label="250" category="instance-splitting">New instances on multiple 250s</rule>
  <rule status="on" id="4" label="300" category="instance-splitting">New instances on multiple 300s</rule>
  <rule status="on" id="5" label="246" category="instance-node">246 domain is instance</rule>
  <rule status="on" id="6" label="247" category="instance-node">247 domain is instance</rule>
  <!--<rule status="on" id="7" label="856" category="instance-splitting">New instances on secondary 856s that are resources (ind2=0)</rule>??? -->
</rules>
);




Works:
-----------------

mbshared:generate-work
from https://github.com/lcnetdev/marc2bibframe/blob/master/modules/module.MBIB-2-BIBFRAME-Shared.xqy#L2559-l2829


Every bib record generates a work.

If 555u, generate a finding aid work by calling  mbshared:generate-finding-aid-work

mbshared:generate-finding-aid-work




mbshared:generate-related-work







mbshared:get-resourceTypes
https://github.com/lcnetdev/marc2bibframe/blob/master/modules/module.MBIB-2-BIBFRAME-Shared.xqy#L3493-L3538

leader|06
  <type leader6="a">Text</type>
  <type leader6="c">NotatedMusic</type>
  <type leader6="d">NotatedMusic</type>
  <type leader6="e">Cartography</type>
  <type leader6="f">Cartography</type>
  <type leader6="g">MovingImage</type>
  <type leader6="i">Audio</type>
  <type leader6="j">Audio</type>
  <type leader6="k">StillImage</type>
  <type leader6="m">Multimedia</type>
  <type leader6="m">Dataset</type>
  <type leader6="o">MixedMaterial</type>
  <type leader6="p">MixedMaterial</type>
  <type leader6="r">ThreeDimensionalObject</type>
  <type leader6="t">Text</type>

007|00
  t: Text
  q: NotatedMusic
  adr: return Cartography
  m: return MovingImage
  v: return MovingImage
  s: Audio
  o: MixedMaterial
  f: Tactile

336a based on specific strings matched
  <type sf336a="(text|tactile text)">Text</type>
  <type sf336a="(notated music|tactile notated music)">NotatedMusic</type>
  <type sf336a="(notated movement|tactile notated movement)">NotatedMovement</type>
  <type sf336a="(cartographic dataset|cartographic image|cartographic moving image|cartographic tactile image|cartographic tactile three-dimensional form|cartographic three-dimensional form)">Cartography</type>
  <type sf336a="(three-dimensional moving image|two-dimensional moving image|cartographic moving image)">MovingImage</type>
  <type sf336a="(performed music|sounds|spoken word)">Audio</type>
  <type sf336a="(still image|tactile image|cartographic image)">StillImage</type>
  <type sf336a="computer program">Multimedia</type>
  <type sf336a="(cartographic dataset|computer dataset)">Dataset</type>
  <type sf336a="(three-dimensional form|tactile three-dimensional form|three-dimensional moving image| cartographic three dimensional form|cartographic tactile three dimensional form)">ThreeDimensionalObject</type>
  <type sf336a="(cartographic tactile image|cartographic tactile three-dimensional form|tactile image|tactile notated music|tactile notated movement|tactile text|tactile three-dimensional form)">Tactile</type>

336b based on specific strings matched
  <type sf336b="(txt|tct)">Text</type>
  <type sf336b="(ntm|ccm)">NotatedMusic</type>`
  <type sf336b="(ntv|tcn)">NotatedMovement</type>
  <type sf336b="(tcrd|cri|crm|crt|crn|crf)">Cartography</type>
  <type sf336b="(tdm|tdi)">MovingImage</type>
  <type sf336b="(prm|snd|spw)">Audio</type>
  <type sf336b="(sti|tci|cri)">StillImage</type>
  <type sf336b="cop">Multimedia</type>
  <type sf336b="(crd|cod)">Ddataset</type>
  <type sf336b="(tdf|tcf|tcm|crf|crn )">ThreeDimensionalObject</type>
  <type sf336b="(crt|crn|tci|tcm|tcn|tct|tcf)">Tactile</type>

337a based on specific strings matched
  <type sf337a="audio">Audio</type>

337b based on specific strings matched
  <type sf337b="s">Audio</type>




For Marc bib record:

if leader06 is
  'a'
    && leader07 is 'a', 'c', 'd' or 'm':   008type = BK
    && leader7 is 'b', 'i' or 's':   008type = SE

  't':  008type = BK
  'p':  008type = MM
  'm':  008type = CF
  'e', 'f', 's':  008type = MP
  'g', 'k', 'o', 'r':  008type = VM
  'c', 'd', 'i', 'j':  008type = MU


generate instances based on 008type
generate instances from 856|859 if ind2 is '0' and sub 3 doesn't contain 'contributor'

types = mbshared:get-resourceTypes
uniformTitle if 130 or 240
names if 100, 110 or 111 OR ( 700,710,711,720 without |t)
titles = 243, 245, 247

label:  first of:
  uniformTitle[madsrdf:authoritativeLabel]
  uniformTitle[bf:workTitle]
  regular titles[bf:workTitle]

  remove trailing period

bf:authorizedAccessPoint from label

events from 033
dissertations from 502
audience from 008|23 and 008type (BK|CF|MU|V)
audience from 521
genre from 008|24
genre from




Instances
-----------------


mbshared:generate-instances
from https://github.com/lcnetdev/marc2bibframe/blob/master/modules/module.MBIB-2-BIBFRAME-Shared.xqy#L2039-L2082


If we have an ISBN from 020a, then


isbn-sets =  if 020a exists, we get sets of 020a (if it matches one of our ISBNs) and bf:isbn 10 and 13 chars

If we have an 020a with an ISBN
  then look in  260, 261, 262, 264 or 300 to  generate-instance-from260

  else if we asked to generate instances from secondary unique ISBNS, call  generate-instance-fromISBN







mbshared:generate-instance-from260
https://github.com/lcnetdev/marc2bibframe/blob/master/modules/module.MBIB-2-BIBFRAME-Shared.xqy#L523-L784

  title from (245|246|247|222|242|210)
  resp_statement from  245c
  edition from 250





mbshared:generate-instance-fromISBN
L1480-L1654

for each ISBN pair, make an instance:
  instanceOf work
  title from 245, and info after 020a ISBN  (volume or physical form or raw)




ISBN:
-----------------

mbshared:process-isbns
https://github.com/lcnetdev/marc2bibframe/blob/master/modules/module.MBIB-2-BIBFRAME-Shared.xqy#L2422-L2461

#   This is the function that finds and dedups isbns, delivering a complete set for generate-instance from isbn
#  If the 020$a has a 10 or 13, they are matched in a set, if the opposite of a pair doesn't exist, it is calculated

  for 020a,
    get string after parens (?), 'normalize' into pairs of equiv 10 and 13 digit ISBNs  via call to  mbshared:get-isbn
    get unique 13 digit isbns

    return bf:set* as element() containing both marcxml:subfield [code=a] (if it matches one of our ISBNs) and bf:isbn calculated nodes


mbshared:get-isbn
https://github.com/lcnetdev/marc2bibframe/blob/master/modules/module.MBIB-2-BIBFRAME-Shared.xqy#L3901-L3972

note:  refers to   http://www.isbn.org/converterpub.asp  as comment

  remove '-' and ' '
  remove everything that isn't a digit at beginning and end
  if length of string is 9, add X at end

  if isbn is 10 chars, create the 13 char version  (prepend 978, replace last digit with checksum)
  if isbn is 13 chars, create the 10 char version  (lop first 3 chars, replace last digit with checksum)


  then If the 020$a has a 10 or 13, they are matched in a set, if the opposite of a pair doesn't exist, it is calculated


isbn-s

:  If the 020$a has a 10 or 13, they are matched in a set, if the opposite of a pair doesn't exist, it is calculated

:   This is the function that finds and dedups isbns, delivering a complete set for generate-instancefrom isbn
:  If the 020$a has a 10 or 13, they are matched in a set, if the opposite of a pair doesn't exist, it is calculated
:   @param  $marcxml        element is the MARCXML record
:   @return wrap as as wrapper for bf:set* as element() containing both marcxml:subfield [code=a] and bf:isbn calculated nodes



https://github.com/lcnetdev/marc2bibframe/blob/master/modules/module.MBIB-2-BIBFRAME-Shared.xqy#L2422-L2461