Permalink
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
113 lines (105 sloc) 2.81 KB
akka.loglevel = DEBUG
IkeToolWebapp = {
port = 8080
indices = [
{
name = WebCorpus1
location = datastore
description = "A collection of ~1.5 million sentences from web documents skewed towards elementary science topics"
item = {
type = directory
datastore = public
group = org.allenai.dictionary.indexes
name = WaterlooFiltered
version = 1
}
},
{
name = WebCorpus2
location = datastore
description = "A collection of ~1.1 million sentences from web documents skewed towards elementary science topics"
item = {
type = directory
datastore = public
group = org.allenai.dictionary.indexes
name = WaterlooFilteredV2Shard5
version = 1
}
}
// NOTE: You can also specify a path in the file system to point to an index by using `location = file`, for e.g.:
//{
// name = MyCorpus
// location = file
// description = "My Blacklab-indexed corpus"
// path = /path/to/index-directory
//}
]
}
Tablestore = {
db = {
// NOTE: Before running locally, uncomment the below lines and specify values for keys `url` to point to your PostgresSql JDBC
// link, `user` and `password`.
// AI2 Internal users: use the database in the `test` deploy environment and the same username. You can get the password
// from this file in the ops-keystore in S3.
url = "jdbc:postgresql://"${POSTGRES_DB}":5432/okcorpus"
user = "okcorpus"
password = ${POSTGRES_PASSWORD}
}
}
QuerySuggester = {
narrow = {
prefixSize = 3
suffixSize = 3
minSimilarityDifference = 5
suggestWord = true
suggestPos = true
suggestSetRepeatedOp = true
}
broaden = {
suggestWord = true
suggestPos = true
minSimilarityDifference = 5
wordPOSSampleSize = 200
}
maxUnlabelledBiasCorrection = 50
pruneOperatorsIfMoreMatchesThan = 1500
pruneOperatorsIfLessThan = 2
minMaxOpReuse = 3
maxOpReuseReturn = 4
maxOpReusePercentOfBeamSize = 0.35
numToSuggest = 11
percentUnlabelled = 0.4
timeoutInSeconds = 55 s
}
word2vecPhrasesSearcher = {
vectors = {
// word2vec embedding (300 dim) bin file
type = file
datastore = public
group = org.allenai.okcorpus
name = citeseer-ten-percent.3.mc100.bin
version = 1
}
format = "binary"
embeddingSize = 300
}
pmiPhrasesSearcher = {
vectors = {
// function PPMI SVD embedding (300 dim) text file
type = file
group = org.allenai.okcorpus
datastore = public
name = functionPPMI.300.vec.txt
version = 1
}
format= "text"
embeddingSize = 300
}
combinationPhraseSearcher = {
// combinationStrategy can take values from [sum, average, min, max]
// default is "average"
combinationStrategy = "average"
}
spray.can.client {
request-timeout = 30 s
}