# dbs/lucegene/bindxml.properties
# d.gilbert, fall 2004


LIB_NAME = bindxml
title = BIND Protein Interactions

# can use absolute paths here, or assume ARGOS_SERVICE_ROOT is set

DATA_ROOT=web/data/extdb/bind/
INDEX_PATH=indices/lucene/bindxml/

MIME_TYPE=text/xml


## --------- search/report keys -------------------

# docid is now interaction id; gi: for Geninfo-id
docurl=http://bind.ca/Action?idsearch=bindid:

batch.forward=lookup.jsp

## parser chopped at '-' ; ok now?
linkto= Object-id_str>fbgn-docid
linkto.fbgn=Object-id_str>docid

searchfield=all

# Fields to display when xml chosen on search result page.
outfields=docid,Molecule_a,MolID_a,Molecule_b,MolID_b,BIND-descr_simple-descr,url
storefields=docid,BIND-descr_simple-descr,Molecule_a,Molecule_b,MolID_a,MolID_b,url
## title == BIND-descr_simple-descr but not yet stored

## for xslt to produce the result page tables
resultxsl = conf/bindxml_result.xsl
resultspage = resultxsl.jsp

header.native=<?xml version="1.0"?>\n\
  <?xml-stylesheet href="conf/bindxml_report.xsl" type="text/xsl"?>\n\
  <BIND-Submit>\n

footer.native=\n</BIND-Submit>

# <!DOCTYPE BIND-Submit PUBLIC "-//NCBI//BIND/EN" "BIND.dtd">
# SAXException Relative URI "BIND.dtd"; can not be resolved without a base URI.
# >> dang these parsers, how do you turn off this w/o editing data?
# perl -pi -e's/^<\!(DOCTYPE.*)>/<\!-- $1 -->/'

header.xml=<?xml version="1.0"?>\n<LuceGene>
footer.xml=\n</LuceGene>

# ? this returns flybase FBrf reports now ; change
batchformat = text/xml
batchformats = text/xml, text/plain, text/csv,  text/tsv
title.text/xml=BIND XML

nativeformat = text/xml

# batchheadlines=1
# batch.outfields= docid, docclass, ArticleTitle, PMID, LastName, Year, all


fieldlabel.docid=ID
fieldlabel.docclass=Class

## for search only?
searchskip.docclass=0
searchskip.title=0
searchskip.docid=1
searchskip.summary=1

# ---- indexing values; NOTE: need these for proper (cased) searches

# locate data with regex file, folder patterns
regex_folder=
regex_file=^bind.*\.xml$
regex_skipfile=
regex_skipfolder=.*

INDEX_CLASS=org.eugenes.index.LuceneXmlIndexer

## append existing index or create new
INDEX_APPEND=false

## index <tag> names as values (as well as field names)?
INDEX_TAGS=false

## index <tag attr="val"> values ?
## none of medline attributes look interesting
INDEX_ATTRIBUTES=true

## use fieldname xpath; full top.middle.last field name?
#? INDEX_XPATH=true -- Medline.xml leaf names are unique enough we can drop the Xpath mess
INDEX_XPATH=false

## INDEX_LEVEL=0 means index main records one level below xml <top> tag 
INDEX_LEVEL=3
INDEX_BLANKS=false

# <BIND-Submit>      level=0
#   <BIND-Submit_date>
#     <Date>
#     </Date>
#   </BIND-Submit_date>
#   <BIND-Submit_sub>
#     <BIND-Submitter>
#     </BIND-Submitter>
#   </BIND-Submit_sub>
#   <BIND-Submit_interactions>   level=1
#     <BIND-Interaction-set>     level=2
#     </BIND-Interaction-set>
#   </BIND-Submit_interactions>
#   <BIND-Submit_complexes>
#     <BIND-Complex-set>     level=2
#     </BIND-Complex-set>
#   </BIND-Submit_complexes>
#   <BIND-Submit_pathways>
#     <BIND-Pathway-set>     level=2
#     </BIND-Pathway-set>
#   </BIND-Submit_pathways>
# </BIND-Submit>

# zcat fbmedline*.gz |\
# perl -pe'if(m,ID:(FBrf\d+),){$id=$1} s,<MedlineID>,<FlyBaseID>$id</FlyBaseID>\n<MedlineID>,;'\
# > medline.xml
# grep -c FlyBaseID med*xml = 32640


## IndexWriter opts
## merge=10 is default; 4 == less mem usage ; 2 minimum  
merge_factor=4
## max_field_length is max # terms/field
max_field_length=1000000
MAX_FIELDS=10000


## field indexing parameters
## sumfields list needs to match field.xxx common summary fields
sumfields=docid,docclass,title

# special summary fields -- replace w/ fieldalias.TAG=newtag
# these must always have fieldtype Text or UnIndexed to be useful

fieldtype=UnStored

fieldtype.docid=Text
fieldtype.Geninfo-id=Text
fieldtype.Interaction-id=Text
field.docid=Interaction-id
# this is BIND GI id
# Interaction-id is interaction id; bindid:

fieldtype.BIND-descr_simple-descr=Text
field.title=BIND-descr_simple-descr
fieldtype.title=Text

fieldtype.Pub_muid=Text
fieldtype.PubMedId=Text
fieldtype.BIND-object_short-label=Text
fieldtype.BIND-object_other-names_E=Text
fieldtype.BIND-object_descr=Text

fieldtype.Object-id_id=Text
fieldtype.Object-id_str=Text
# ^^ FBgn ids are here
#field.docid=Object-id_str
#fieldalias.FlyBaseID=docid



analyzer=org.eugenes.index.BiodataAnalyzer2

# all field defaults
tokenizer=fbacode$LowerWordTokenizer
tokenfilter=fbacode$DebugFilter

## make DBX FB id field instead of relying on Object-id_str ?

## want to separately index BIND-Interaction_a and _b fields ??
## esp want FBid and BIND-object_short-label from both for results views

## this wont work with  INDEX_XPATH=false

#fieldalias.BIND-Interaction_a.BIND-object.BIND-object_short-label=Molecule_a
#fieldalias.BIND-Interaction_b.BIND-object.BIND-object_short-label=Molecule_b

fieldrecoder.BIND-object_short-label=fbacode$BINDInter_Recoder
fieldrecoder.Object-id_str=fbacode$BINDInter_Recoder,fbacode$FBID_Recoder

fieldalias.BIND-object_short-label_a=Molecule_a
fieldalias.BIND-object_short-label_b=Molecule_b
fieldalias.Object-id_str_a=MolID_a
fieldalias.Object-id_str_b=MolID_b

fieldtype.Molecule_a=Text
fieldtype.Molecule_b=Text
fieldtype.MolID_a=Text
fieldtype.MolID_b=Text

# fieldrecoder.Object-id_str=fbacode$FBID_Recoder
fieldtype.DBX=Text

## this full xpath is hopeless..
# fieldalias.BIND-Interaction_b.BIND-object.BIND-object_id\
#  .BIND-object-type-id.BIND-object-type-id_protein.BIND-id.BIND-id_other\
#  .Seq-id.Seq-id_general.Dbtag.Dbtag_tag.Object-id.Object-id_str
#   =FBID_B
# fieldtype.FBID_B=Text

tokenfilter.docid=fbacode$DebugEndOfRecordFilter
#tokenfilter.EOR=fbacode$DebugEndOfRecordFilter
 
# to create "contents" field of all text
indexall=false
 
# # AddCommonField_FieldRecoder
# # generate base fieldname fields when INDEX_XPATH=true
# # this doesnt prevent default indexing of full xpath
# # using property 'fieldalias.att_timestamp=timestamp' will collapse all
# #   for given field.
# 
# #fieldrecoder.name=LucegeneIndexers$AddCommonField_FieldRecoder
# #fieldrecoder.att_id=LucegeneIndexers$AddCommonField_FieldRecoder
# #fieldrecoder.synonym=LucegeneIndexers$AddCommonField_FieldRecoder
# #fieldrecoder.description=LucegeneIndexers$AddCommonField_FieldRecoder
# 
# # tokenfilter.att_length=org.eugenes.index.BiodataAnalyzer$NumberFilter
# 
# ## these look useless:
# fieldtype.QualifierName.att_MajorTopicYN=skip
# fieldtype.DescriptorName.att_MajorTopicYN=skip
# fieldtype.RegistryNumber=skip
# fieldtype.GrantList.att_CompleteYN=skip

# #----------------- sample debug output

# bin/lucegene-search.sh -verbose=1 -lib medline