Skip to content

Commit

Permalink
codemeta.json: updated according to (proposed) CLARIAH requirements (C…
Browse files Browse the repository at this point in the history 
…LARIAH/clariah-plus#38)
  • Loading branch information
@proycon
proycon committed Jul 22, 2022
1 parent 9c5cb51 commit 0cdf68a
Showing 1 changed file with 171 additions and 45 deletions.
216 changes: 171 additions & 45 deletions codemeta.json
View file
Original file line number Diff line number Diff line change
@@ -1,14 +1,15 @@
{
"@context": [
"https://doi.org/10.5063/schema/codemeta-2.0",
"http://schema.org"
"http://schema.org",
"https://w3id.org/software-types"
],
"@type": "SoftwareSourceCode",
"identifier": "foliautils",
"name": "foliautils",
"version": "0.18",
"description": "Command-line utilities for working with the Format for Linguistic Annotation (FoLiA).",
"license": "https://spdx.org/licenses/GPL-3.0",
"license": "https://spdx.org/licenses/GPL-3.0-only",
"url": "https://github.com/LanguageMachines/foliautils",
"author": [
{
Expand Down Expand Up @@ -83,7 +84,7 @@
"issueTracker": "https://github.com/LanguageMachines/foliautils/issues",
"contIntegration": "https://travis-ci.org/LanguageMachines/foliautils",
"releaseNotes": "https://github.com/LanguageMachines/foliautils/releases",
"developmentStatus": "active",
"developmentStatus": "https://www.repostatus.org/#active",
"keywords": [ "nlp", "natural language processing", "folia", "xml", "linguistic annotation" ],
"referencePublication": [
{
Expand Down Expand Up @@ -114,97 +115,222 @@
"url": "https://github.com/proycon/folia/raw/master/docs/folia.pdf"
}
],
"entryPoints": [
"targetProduct": [
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-2text",
"urlTemplate": "file:///FoLiA-2text",
"executableName": "FoLiA-2text",
"description": "Convert FoLiA documents into plain text",
"interfaceType": "CLI"
"consumesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
],
"producesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "text/plain"
}
]
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-txt",
"urlTemplate": "file:///FoLiA-txt",
"description": "Convert plain textinto FoLiA, the output will contain only <p> and <str> nodes. See ucto or rst2folia (FoLiA-tools) for alternatives.",
"interfaceType": "CLI"
"executableName": "FoLiA-txt",
"description": "Convert plain text to FoLiA, the output will contain only <p> and <str> nodes. See ucto or rst2folia (FoLiA-tools) for alternatives.",
"consumesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "text/plain"
}
],
"producesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
]
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-page",
"urlTemplate": "file:///FoLiA-page",
"executableName": "FoLiA-page",
"description": "Convert PAGE XML to FoLiA",
"interfaceType": "CLI"
"consumesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/page+xml"
}
],
"producesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
]
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-hocr",
"urlTemplate": "file:///FoLiA-hocr",
"executableName": "FoLiA-hocr",
"description": "Convert hOCR (as outputted by Tesseract) to FoLiA",
"interfaceType": "CLI"
"consumesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "text/html"
}
],
"producesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
]
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-alto",
"urlTemplate": "file:///FoLiA-alto",
"description": "Convert ALTO DIDL files into a series of FoLiA documents",
"interfaceType": "CLI"
"executableName": "FoLiA-alto",
"description": "Convert ALTO DIDL files into a series of FoLiA documents"
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-langcat",
"urlTemplate": "file:///FoLiA-langcat",
"executableName": "FoLiA-langcat",
"description": "Language Identification using textcat.",
"interfaceType": "CLI"
"consumesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
],
"producesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
]
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-idf",
"urlTemplate": "file:///FoLiA-idf",
"executableName": "FoLiA-idf",
"description": "Count words in a series of FoLiA documents and compute IDF statistics, which are outputted to a tsv file",
"interfaceType": "CLI"
"consumesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
],
"producesData": [
{
"@type": "Dataset",
"encodingFormat": "text/tab-separated-values"
}
]
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-stats",
"urlTemplate": "file:///FoLiA-stats",
"executableName": "FoLiA-stats",
"description": "Gather n-gram statistics over a series of FoLiA documents",
"interfaceType": "CLI"
"consumesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
],
"producesData": [
{
"@type": "Dataset",
"encodingFormat": "text/tab-separated-values"
}
]
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-collect",
"urlTemplate": "file:///FoLiA-collect",
"executableName": "FoLiA-collect",
"description": "Collect n-gram statistics from tsv files produced by FoLiA-stats, aggregating results.",
"interfaceType": "CLI"
"consumesData": [
{
"@type": "Dataset",
"encodingFormat": "text/tab-seperated-values"
}
],
"producesData": [
{
"@type": "Dataset",
"encodingFormat": "text/tab-separated-values"
}
]
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-correct",
"urlTemplate": "file:///FoLiA-correct",
"executableName": "FoLiA-correct",
"description": "Correct FoLiA documents using correction candidates generated by TICCL-rank (from ticcltools)",
"interfaceType": "CLI"
"consumesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
],
"producesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
]
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-wordtranslate",
"urlTemplate": "file:///FoLiA-wordtranslate",
"executableName": "FoLiA-wordtranslate",
"description": "Simple word-by-word translator on the basis of a dictonary and/or rewrite rules",
"interfaceType": "CLI"
"consumesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
],
"producesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
]
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-clean",
"urlTemplate": "file:///FoLiA-clean",
"executableName": "FoLiA-clean",
"description": "FoLiA-clean will produce a cleaned up version of a FoLiA file, or a whole directory of FoLiA files, removing specified annotation types and specified text classes",
"interfaceType": "CLI"
"consumesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
],
"producesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
]
},
{
"@type": "EntryPoint",
"@type": "CommandLineApplication",
"name": "FoLiA-pm",
"urlTemplate": "file:///FoLiA-pm",
"executableName": "file:///FoLiA-pm",
"description": "Convert Political Maskup XML to FoLiA",
"interfaceType": "CLI"
"producesData": [
{
"@type": "TextDigitalDocument",
"encodingFormat": "application/folia+xml"
}
]
}
]
}

0 comments on commit 0cdf68a

Please sign in to comment.