[
https://issues.apache.org/jira/browse/TIKA-1614?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14510454#comment-14510454
]
Anya Yun Li commented on TIKA-1614:
-----------------------------------
Hi Nick,
I understand your concern. This is a content-based geoparser method, that is,
we identified location names from text, but in order to get the geographical
information(longitude, latitude), we need some kind of database for looking up.
Here I use Lucene to build index on GeoName's dataset, and this dataset
provides such information.
Those binary patch are Lucene index on GeoName's dataset.
If the above explanation does not answer your question, feel free to contact me.
Best,
Yun
> Geo Topic Parser
> ----------------
>
> Key: TIKA-1614
> URL: https://issues.apache.org/jira/browse/TIKA-1614
> Project: Tika
> Issue Type: New Feature
> Components: parser
> Reporter: Anya Yun Li
> Labels: memex
>
> ##Description
> This program aims to provide the support to identify geonames for any
> unstructured text data in the project NSF polar research.
> https://github.com/NSF-Polar-Cyberinfrastructure/datavis-hackathon/issues/1
> This project is a content-based geotagging solution, made of a variaty of NLP
> tools and could be used for any geotagging purposes.
> ##Workingflow
> 1. Plain text input is passed to geoparser
> 2. Location names are extracted from the text using OpenNLP NER
> 3. Provide two roles:
> * The most frequent location name choosed as the best match for the
> input text
> * Other extracted locations are treated as alternatives (equal)
> 4. location extracted above, search the best GeoName object and return the
> resloved objects with fields (name in gazetteer, longitude, latitude)
> ##How to Use
> *Cautions*: This program requires at least 1.2 GB disk space for building
> Lucene Index
> ```Java
> function A(stream){
> Metadata metadata = new Metadata();
> ParseContext context=new ParseContext();
> GeoParserConfig config= new GeoParserConfig();
> config.setGazetterPath(gazetteerPath);
> config.setNERModelPath(nerPath);
> context.set(GeoParserConfig.class, config);
>
> geoparser.parse(
> stream,
> new BodyContentHandler(),
> metadata,
> context);
>
> for(String name: metadata.names()){
> String value=metadata.get(name);
> System.out.println(name +" " + value);
> }
> }
> ```
> This parser generates useful geographical information to Tika's Metadata
> Object.
> Fields for best matched location:
> ```
> Geographic_NAME
> Geographic_LONGTITUDE
> Geographic_LATITUDE
> ```
> Fields for alternatives:
> ```
> Geographic_NAME1
> Geographic_LONGTITUDE1
> Geographic_LATITUDE1
> Geographic_NAME2
> Geographic_LONGTITUDE2
> Geographic_LATITUDE2
> ...
> ```
> If you have any questions, contact me: [email protected]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
