CERN Accelerating science

BibConvert Admin Guide

Contents

A. Overview
B. XML-oriented mode
      1 Configuration File Examples
      2 Running BibConvert
C. Plain text-oriented mode
      1 Configuration File Examples
      2 Running BibConvert
      3 BibConvert Configuration Guide
          Conventions
          3.1 Step 1 Definition of Source Record
          3.2 Step 2 Definition of Source Fields
          3.3 Step 3 Definition of Target Record
          3.4 Formatting in BibConvert
             3.4.1 Definition of Formatting Functions
             3.4.2 Generated Values

A. Overview

BibConvert utility enables you to convert metadata records from various metadata formats into another metadata format supported by the Invenio local database. It is designed to process XML harvested metadata records, converting them into MARC21 before they are uploaded into the database. However, BibConvert is flexible enough to deal also with other structured metadata according to your needs, and offers a way to actually insert what you want into the database.

BibConvert is suitable for tasks such as conversion of records received from multiple data sources, or conversion of records from another system that may support a different metadata format.

In order to cover a wider range of possible conversions, BibConvert has 2 different modes, each dealing with different types of data, and each using different configuration files.

Plain text-oriented mode:
Deals with source data being typically structured with line breaks, and character-based separators. You can use this mode when you need to process line-based data, such as comma/tab separated values. Still, this mode is powerful enough to convert complex structures, at the cost of a more complex configuration.
XML-oriented mode
Convert source data being encoded in XML. Provided you have installed a supported XSLT processor on your machine, BibConvert can make use of standard XSLT to interpret your XML data.

In addition to XSLT, we provide a home-made solution for converting XML source data. It uses our own BFX language as transformation language, extended with XPath for node selections.
You should consider using this solution only in the case where you have not installed (or do not want to install) an XSLT processor on your machine.

B XML-Oriented Mode

1 Configuration File Examples

Using XSLT

You can use standard XSL stylesheets to transform your source data. Have a look at the provided samples in etc/bibconvert/config directory of your Invenio installation.

Using BFX

BFX (BibFormat for XML) uses a syntax similar to XSLT. Roughly they only differ in the name of the tags.
More documentation about it is to be added soon (FIXME).
Have a look at the provided samples in etc/bibconvert/config directory of your Invenio installation to learn more about it.

2. Running BibConvert

BibConvert in XML-oriented mode has only 1 parameter: -c. It is used to specify which transformation stylesheet to apply to the piped XML.
$ bibconvert -coaidc2marcxml.xsl < sample.xml > /tmp/record.xml
If the stylesheet you want to use is installed in the etc/bibconvert/config directory of your Invenio installation, then you can just refer to it by its filename. Otherwise use the full path to the file.

C Plain Text-Oriented Mode

1 Configuration File Examples

OAI DublinCore into MARC21 and OAI MARC into MARC21 configurations will be provided as default configuration, ensuring the standard uploading sequence (incl. OAIHarvest and BibUpload utilities). Other configurations can be created according to your needs. The configuration file that has to be created for each data source is a text file with following structure:

  ### the configuration starts here
  ### Configuration of bibconvert templates
  ### source data : 
 
  === data extraction configuration template ===
  ### here comes the data extraction configuration template
  #   entry example:
 
  AU---%A---MAX---;---
 
  #   extracts maximum available data by field from metadata record
  #   the values are found between specified tags
  #   in this case between the '%A' tag and other tags defined
  #   repetitive values are recognized by a semicolon separator
  #   resp. by multiple presence of '%A' tag
 
  ===   data source configuration template   ===
  ### here comes the data source configuration template
  #   entry example:
 
  AU---<:FIRSTNAME:>-<:SURNAME:>
 
  #   describes the contents of extracted source data fields
  #   in this case, the field AU is described as having two distinct subfields
 
  ===   data target configuration template   ===
  ### here comes the data target configuration template
  #   entry example:
 
  AU::CONF(AU,,0)---<datafield id="700" ind1="" ind2=""><subfield code="a"><:AU*::SURNAME::CAP():>, <AU*::FIRSTNAME::ABR():></subfield></datafield>
 
  #   This section concerns rather the desired output, while previous two were focused on the data source structures.
  #   Each line equals to one output line, composed of given literals and values from extracted source data fields.
  #   In this example, the XML Marc21 output line is defined,
  #   containing re-formatted values of source fields SURNAME and FIRSTNAME
 
  ### the configuration ends here

Having prepared a configuration, the BibConvert will convert the source data file according to it in a batch mode. The BibConvert is fully compatible with the Uploader1.x configuration language. For more information, have a look at the BibConvert Configuration Guide section below.

2. Running BibConvert

For a fully functional demo, consider the following sample input data:

sample.dat -- sample bibliographic data to be converted and inputted into Invenio
sample.cfg -- sample configuration file, featuring knowledge base demo

To convert the above data into XML MARC, use the following command:

$ bibconvert -b'<collection>' -csample.cfg -e'</collection>' < sample.dat > /tmp/sample.xml
and see the XML MARC output file. You would then continue the upload procedure by calling BibUpload.

Other useful BibConvert configuration examples:

dcq.cfg -- Qualified Dublin Core in SGML to XML MARC example
dcq.dat -- corresponding data file, featuring collection identifiers demo
dcxml-to-marcxml.cfg -- OAI XML Dublin Core to XML MARC example
bibtex.cfg -- BibTeX to XML MARC example

3 BibConvert Configuration Guide

Conventions


- comment line starts with '#' sign in the first column
- each section is declared by a line starting with '===' (further characters on the line are ignored)
- values are separated by '---'

3.1 Step 1 Definition of Source record

- Create/edit "data extraction configuration template" section of the configuration file.
- Each line of this section stands for a definition of one source field:

name---keyword---terminating string---separator---

- Choose a (valid) name allowed by the system
- Enter keyword and terminating string, which are boundary tags for the wanted value extraction
- In case the field is repetitive, enter the value separator
- "---"is mandatory separator between all values, even zero-length
- MAX/MIN keywords can be used instead of terminating string
 

Example of a definition of author(repetitive) and title (non-repetitive) fields:

  === data extraction configuration template ===
  ### here comes the data extraction configuration template
 
  AU---AU_---MAX---;---
  TI---TI_---EOL------

3.2 Step 2 Definition of Source fields

Each field extracted from the source according to the definition done in the first step can have an internal structure, which is described in this section.

- Create/edit "data source configuration template" section of the configuration file.
- Each line of this section stands for a definition of one source field
- corresponds to the name defined in the step 1

name---{CONST<:SUBFIELD:>[CONST]}}

- Enter only constants that appear systematically.
- Between two discrete subfields has to be defined a constant of a non zero length
- "---"is a mandatory separator between the name and the source field definition

Example of a definition of author(repetitive) and title (non-repetitive) fields:

===   data source configuration template   ===
TI---<:TI:>
AU---<:FIRSTNAME:>-<:SURNAME:>

3.3 Step 3 Definition of target record

This definition describes the layout of the target record that is created by the conversion, together with the corresponcence to the source fields defined in step 2.

- Create/edit "data target configuration template" section of the configuration file.
- Each line of this section stands for an output line created by the conversion.
- <name> corresponds to the name defined in the steps 1 and 2

CODE---CONST<:name::SUBFIELD::FUNCT():>CONST<:GENERATED_VALUE:>

- CODE stands for a tag for readability (optional)
- "::"is a mandatory separator between the name and the subfield definition
- optionally, you can apply the appropriate formatting function(s) and generated values
- "::"is a mandatory separator between the subfield definition and the function(s)
- "---"is a mandatory separator between the tag and the output code definition
- mark repetitive source fields with an asterisk (*)

Example of a definition of author (repetitive) and title (non-repetitive) codes:


AU::CONF(AU,,0)---<datafield id="700" ind1="" ind2=""><subfield code="a"><:AU*::AU:></subfield></datafield>
TI::CONF(TI,,0)---<datafield id="245" ind1="" ind2=""><subfield code="a"><:TI::TI::SUP(SPACE, ):></subfield></datafield>

- preserve newlines in a source field for later use by formatting functions by marking them with "^"

Example of a definition of a book editors field in which the newlines are preserved so that they can be processed by the JOINMULTILINES formatting function:


AU---<datafield id="773" ind1=" " ind2=" "><:BOOKEDITOR^::BOOKEDITOR::JOINMULTILINES(<subfield code="a">,</subfield>):></datafield>

With a value such as:
Test Case, A
The results may be:

<datafield tag="773" ind1="" ind2=""><subfield code="a">Test</subfield><subfield code="a">Case, A</subfield></datafield>

3.4 Formatting in BibConvert

 3.4.1 Definition of formatting functions
Every field can be processed with a variety of functions that partially or entirely change the original value.
There are three types of functions available that take as element either single characters, words or the entire value of processed field.
 

Every function requires a certain number of parameters to be entered in brackets. If an  insufficient number of parameters is present, the function uses default values. Default values are constructed with attempt to keep the original value.

The configuration of templates is case sensitive.

The following functions are available:

ADD(prefix,suffix) - add prefix/suffix
KB(kb_file,[0-9]) -lookup in kb_file and replace value
ABR(x,suffix)/ABRW(x,suffix) - abbreviation with suffix addition
ABRX() - abbreviate exclusively words longer
CUT(prefix,postfix) - remove substring from side
REP(x,y) - replacement of characters
SUP(type) - suppression of characters of specified type
LIM(n,L/R)/LIMW(str,L/R) - restriction to n letters
WORDS(n,side) - restriction to n words from L/R
MINL(n)/MAXL(n) - replacement of words shorter/greater than n
MINLW(n) - replacement of short values
EXP(str,1|0)/EXPW(type) - replacement of words from value if containing spec. type/string
IF(value,valueT,valueF) - replace T/F value
UP/DOWN/CAP/SHAPE/NUM - lower case and upper case, shape
SPLIT(n,h,str,from)/SPLITW(sep,h,str,from) - split into more lines
CONF(field,value,1/0)/CONFL(value,1/0) - confirm validity of a field
RANGE(from,to) - confirm only entries in the specified range
DEFP() - default print
IFDEFP(field,value,1/0) - IF condition is met, default print
JOINMULTILINES(prefix,suffix) - Join a multiline string into a single line with each segment having prefix and suffix
 

ADD(prefix,postfix)

default: ADD(,)    no addition

Adds prefix/postfix to the value, we can use this function to add the proper field name as a prefix of the value itself:

ADD(WAU=,)    prefix for the first author (which may have been taken from the field AU2)
 

KB(kb_file)    -    kb_file search

default: KB(kb_file,1/0/R)

The input value is compared to a kb_file and may be replaced by another value. In the case that the input value is not recognized, it is by default kept without any modification. This default can be overridden by _DEFAULT_---default value entry in the kb_file

The file specified in the parameter is a text file representing a table of values that correspond to each other:

{input_value---output_value}

KB(file,1) searches the exact value passed.
KB(file,0) searches the KB code inside the value passed.
KB(file,2) as 0 but not case sensitive
KB(file,R) replacements are applied on substrings/characters only.

bibconvert look-up value in KB_file in one of following modes:
===========================================================
1 - case sensitive / match (default)
2 - not case sensitive / search
3 - case sensitive / search
4 - not case sensitive / match
5 - case sensitive / search (in KB)
6 - not case sensitive / search (in KB)
7 - case sensitive / search (reciprocal)
8 - not case sensitive / search (reciprocal)
9 - replace by _DEFAULT_ only
R - not case sensitive / search (reciprocal) replace

Edge spaces are not considered. Output value is not further formated.

ABR(x,trm),ABRW(x,trm)  - abbreviate term to x places with(out) postfix

default: ABR(1,.)
default: ABRW(1,.)

The words in the input value are shortened according to the parameters specified. By default, only the initial character is kept and the output value is terminated by a dot.
ABRW takes entire value as one word.

 
example input output
ABR() firstname_surname f._s.
ABR(1,) firstname_surname f_s
ABR(10,COMMA) firstname_surname firstname,_surname,

ABRX() - abbreviate exclusively words longer than given limit

default: ABRX(1,.)

Exclusively words that reach the specified length limit in the input value are abbreviated. No suffix is appended to the words shorter than specified limit.

CUT(prefix,postfix) - remove substring from side

default: CUT(,)

Remove string from the value (reverse function to the "ADD")

REP(x,y)   - replace x with y

default: REP(,)    no replacement

The input value is searched for the string specified in the first parameter. All such strings are replaced with the string specified in the second parameter.

SUP(type,string)   - suppress chars of certain type

default: SUP(,)    type not recognized

All groups of characters belonging to the type specified in the first parameter are suppressed or replaced with a string specified in the second parameter.

Recognized types:

SPACE .. invisible chars incl. NEWLINE
ALPHA .. alphabetic
NALPHA .. not alphabetic
NUM .. numeric
NNUM    .. not numeric
ALNUM  .. alphanumeric
NALNUM  .. non alphanumeric
LOWER  .. lower case
UPPER  .. upper case
PUNCT  .. punctuation
NPUNCT  .. not punctuation
 

 
example input output
SUP(SPACE,-) sep_1999 sep-1999
SUP(NNUM) sep_1999 1999
SUP(NUM) sep_1999 sep_

LIM(n,side)/LIMW(str,side)   - limit to n letters while trimming L/R side

default: LIM(0,)        no change
default: LIMW(,R)        no change

Limits the value to the required number of characters by cutting excess characters either on the Left or Right.
LIMW locates the first occurrence of (str) string and cut either Left or Right side.

 
example input output
LIM(4,L) sep_1999 1999
LIM(4,R) sep_1999 sep_
LIMW(_,R) sep_1999 sep_

WORDS(n,side)  - limit to n words while trimming L/R side

default: WORDS(0,R)

Keeps the number of words specified in the first parameter and cuts the excessive characters either on Left or Right.

 
example input output
WORDS(1,R) Sep 1999 Sep
WORDS(1,L) Sep 1999 1999

MINL(n)   - exp. words shorter than n

default: MINL(1)

All words shorter than the limit specified in the parameter are replaced from the sentence.
The words with length exactly n are kept.

 
example input output
MINL(2) History of Physics History of Physics
MINL(3) History of Physics History Physics

MAXL(n)   - exp. words longer than n

default: MAXL(0)

All words greater in number of characters than the limit specified in the parameter are replaced. Words with length exactly n are kept.

 
example input output
MAXL(2) History of Physics of
MAXL(3) History of Physics of

MINLW(n) - replacement of short values

default: MINLW(1) (no change)

The entire value is deleted if shorter than the specified limit.
This is used for the validation of created records, where we have 20 characters in the header.
The default validation is MINLW(21), i.e. the record entry will not be consided as valid, unless it contains at least 21 characters including the header. This default setting can be overriden by the -l command line option.

In order to increase the necessary length of the output line in the configuration itself, apply the function on the total value:

AU::MINLW(25)---CER <:SYSNO:> AU    L <:SURNAME:>, <:NAME:>
 
 

EXP(str,1|0) - exp./aprove word containing specified string

default: EXP   (,0)     leave all value

The record is shortened by replacing words containing the specified string.
The second parameter states whether the string approves the word (0) or disables it (1).

for example, to get the email address from the value, use the following

 
example input output
EXP(@,0) mail to: libdesk@cern.ch libdesk@cern.ch
EXP(:,1) mail to: libdesk@cern.ch mail libdesk@cern.ch
EXP(@) mail to: libdesk@cern.ch libdesk@cern.ch

EXPW(type)   - exp. word from value if containing spec. type

default: EXPW        type not recognized
 

The sentence is shortened by replacing words containing specified type of character.

Types supported in EXPW function:

ALPHA .. alphabetic
NALPHA .. not alphabetic
NUM .. numeric
NNUM    .. not numeric
ALNUM  .. alphanumeric
NALNUM  .. non alphanumeric
LOWER  .. lower case
UPPER  .. upper case
PUNCT  .. punctuation
NPUNCT  .. non punctuation

Note: SPACE is not handled as a keyword, since all space characters are considered as word separators.

 
example input output
EXPW(NNUM) sep_1999 1999
EXPW(NUM) sep_1999 sep

IF(value,valueT,valueF) - replace T/F value

default: IF(,,)

Compares the value with the first parameter. In case the result is TRUE, the input value is replaced with the second parameter, otherwise the input value is replaced with the third parameter.
In case the input value has to be kept, whatever it is, the keyword ORIG can be used (usually in the place of the third parameter)

 
example input output
IF(sep_1999,sep) sep_1999 sep
IF(oct_1999,oct) sep_1999
IF(oct_1999,oct,ORIG) sep_1999 oct_1999

UP    - upper case

Convert all characters to upper case

DOWN   - lower case

Convert all characters to lower case

CAP    - make capitals

Convert the initial character of each word to upper case and the rest of characters to lower case

SHAPE    - format string

Supresses all invalid spaces

NUM    - number

If it contains at least one digit, convert it into a number by suppressing other characters. Leading zeroes are deleted.

SPLIT(n,h,str,from)

Splits the input value into more lines, where each line contains at most (n+h+length of str) characters, (n) being the number of characters following the number of characters in the header, specified in (h). The header repeats at the beginning of each line. An additional string can be inserted as a separator between the header and the following value. This string is specified by the third parameter (str). It is possible to restrict the application of (str) so it does not appear on the first line by entering "2" for (from)

SPLITW(sep,h,str,from)

Splits the input value into more lines by replacing the line separator stated in (sep) with CR/LFs. Also, as in the case of the SPLIT function, the first (h) characters are taken as a header and repeat at the beginning of each line.  An additional string can be inserted as a separator between the header and the following value. This string is specified by the third parameter (str). It is possible to restrict the application of (str) so it does not appear on the first line by entering "2" for (from)

CONF(field,value,1/0)  - confirm validity of a field

The input value is taken as it is, or refused depending on the value of some other field. In case the other (field) contains  the string specified in (value), then the input value is confirmed (1) or refused (0).

CONFL(str,1|0) - confirm validity of a field

The input value is confirmed if it contains (1)/misses(0) the specified string (str)

RANGE(from,to) - confirm only entries in the specified range

Left side function of target template configuration section to select the desired entries from the repetitive field.
The range can only be continuous.

The entry is confirmed in case its input falls into the range from-to specified in the parameter, border values included. As an upper limit it is possibe to use the keyword MAX.

This is useful in case of AU code, where the first entry has a different definition from other entries:

AU::RANGE(1,1)---CER <:SYSNO:> AU2    L <:AU::SURNAME:>, <:AU::NAME:>    ... takes the first name from the defined AU field
AU::RANGE(2,MAX)---CER <:SYSNO:> AU     L <:AU::SURNAME:> , <:AU::NAME:>    ... takes the the rest of namesfrom the AU field
 

DEFP() - default print

The value is printed by default even if it does not contain any variable input from the source file.

IFDEFP(field,value,1/0) - IF condition is met, default print

The line is printed by default (even if it does not contain any variable input from the source file) IF a condition is met that depends on the value of some other field. The condition is basically either that "field" contains "value" (in which case the 3rd parameter should be set to 1), or that "field" does NOT contain "value" (in which case the 3rd parameter should be set to 0).

For example, given the following line:

690C::REP(EOL,)::IFDEFP(comboYEL,BOOK,1)---<datafield tag="690" ind1="C" ind2=" "><subfield code="a">BOOK</subfield></datafield>

We want to print the line if the (field) "comboYEL" contains the (value) "BOOK", otherwise we don't want to print it. Therefore, the 3rd parameter is set to "1". However, in the following line:

690C::REP(EOL,)::IFDEFP(comboYEL,BOOK,0)---<datafield tag="690" ind1="C" ind2=" "><subfield code="a">OTHER</subfield></datafield>

We want to print the line if the (field) "comboYEL" does NOT contain the (value) "BOOK", otherwise we don't want to print it. Therefore, the 3rd parameter is set to "0".

This is achieved by using "IFDEFP". If the line had contained variables, the "CONF" function would have been used instead.

JOINMULTILINES(prefix,suffix) - Join a multiline string into a single line with each segment having prefix and suffix

Given a field-value with newlines in it, split the field on the new lines (\n), separating them with prefix, then suffix. E.g.:
For the field XX with the value:

  Test
  Case, A

And the function call:

  <:XX^::XX::JOINMULTILINES(<subfield code="a">,</subfield>):>

The results would be:

  <subfield code="a">Test</subfield><subfield code="a">Case, A</subfield>

One note on this: <:XX^::XX:
Without the ^ the newlines will be lost as bibconvert will remove them, so you'll never see an effect from this function.

3.4.2 Generated values

In the template configurations, values can be either taken from the source or generated in the process itself. This is mainly useful for evaluating constant values.

Currently, the following date values are generated:
 

DATE(format,n)

default: DATE(,10)

where n is the number of digits required.

Generates the current date in the form given as a parameter. The format has to be given according to the ANSI C notation, i.e. the string is composed out of following components:

  %a    abbreviated weekday name
  %A    full weekday name
  %b    abbreviated month name
  %B    full month name
  %c    date and time representation
  %d    decimal day of month number (01-31)
  %H    hour (00-23)(12 hour format)
  %I    hour (01-12)(12 hour format)
  %j    day of year(001-366)
  %m    month (01-12)
  %M    minute (00-59)
  %p    local equivalent of a.m. or p.m.
  %S    second (00-59)
  %U    week number in year (00-53)(starting with Sunday)
  %V    week number in year
  %w    weekday (0-6)(starting with Sunday)
  %W    week number in year (00-53)(starting with Monday)
  %x    local date representation
  %X    local time representation
  %y    year (no century prefix)
  %Y    year (with century prefix)
  %Z    time zone name
  %%    %
 

WEEK(diff)

Enters the two-digit number of the current week (%V) increased by specified difference.
If the resulting number is negative, the returned value is zero (00).
Values are kept up to 99, three digit values are shortened from the left.

WEEK(-4)    returns 48, if current week is 52
WEEK           current week
 

SYSNO

 
Works the same as DATE, however the format of the resulting value is fixed so it complies with the requirements of further record handling. The format is 'whhmmss', where:

w     current weekday
hh    current hour
mm    current minute
ss    current second

The system number, if generated like this, contains a variable value changing every second. For the system number is an identifier of the record, it is needed to ensure it will be unique for the entire record processed. Unlike the function DATE, which simply generates the value of format given, SYSNO keeps the value persistent throughout the entire record and excludes collision with other records that are generated in period of one week with one second granularity.

It is not possible to use the DATE function for generating a system number instead.

The system number is unique in range of one week only, according to the current definition.
 
 

OAI

Inserts OAI identifier incremented by one for earch record Starting value that is used in the first record in the batch job can be specified on the command line using the -o<starting_value> option.