Author: nick
Date: Mon May 8 18:13:15 2017
New Revision: 1794435
URL: http://svn.apache.org/viewvc?rev=1794435&view=rev
Log:
EMF, and mention the existance of the NER, NLP and OR parsers
Modified:
tika/site/src/site/apt/1.15/formats.apt
Modified: tika/site/src/site/apt/1.15/formats.apt
URL:
http://svn.apache.org/viewvc/tika/site/src/site/apt/1.15/formats.apt?rev=1794435&r1=1794434&r2=1794435&view=diff
==============================================================================
--- tika/site/src/site/apt/1.15/formats.apt (original)
+++ tika/site/src/site/apt/1.15/formats.apt Mon May 8 18:13:15 2017
@@ -209,6 +209,9 @@ Supported Document Formats
The {{{./api/org/apache/tika/parser/microsoft/WMFParser.html}WMFParser}}
class extracts simple text from Microsoft WMF drawings.
+ The {{{./api/org/apache/tika/parser/microsoft/EMFParser.html}EMFParser}}
+ class extracts simple text from Microsoft EMF drawings, along with
+ exposing any embedded other resources / files.
* {Video formats}
@@ -264,7 +267,7 @@ Supported Document Formats
extract email messages from the Microsoft Outlook PST email format.
The
{{{./api/org/apache/tika/parser/microsoft/OutlookExtractor.html}OutlookExtractor}}
(part of
- {{{./api/org/apache/tika/parser/microsoft/OfficeParser}OfficeParser}})
+ {{{./api/org/apache/tika/parser/microsoft/OfficeParser.html}OfficeParser}})
is able to extract email messages from the Microsoft Outlook MSG email
format.
@@ -352,6 +355,36 @@ Supported Document Formats
{{{http://www.digitalpreservation.gov/formats/fdd/fdd000325.shtml}
digitalpreservation.gov}}
for background on this format.
+* {Natural Language Processing}
+
+ Tika supports calling out to a number of Natural Language Processing and
+ Named Entity Recognition frameworks, tools and libraries.
+
+ These can be used to support additional formats, or to gain extra
information on
+ existing formats. In many cases, additional tools or REST services or
training
+ datasets are required to enable or power this support.
+
+ Details on the requirements and setup steps are generally given either in
+ the parser's javadocs, or on the {{{https://wiki.apache.org/tika/}Tika
wiki}}.
+
+ The
{{{./api/org/apache/tika/parser/sentiment/analysis/SentimentParser.html}SentimentParser}}
+ class classifies documents based on the sentiment of document, powered by
Apache
+ OpenNLP's Maximum Entropy Classifier.
+
+ {{{./api/org/apache/tika/parser/journal/JournalParser.html}JournalParser}}
uses
+ Grobid (via RESTful server) to extract additional metadata from the text of
+ journal publications. A number of other NLP and NER parsers are available
in the
+ {{{./api/org/apache/tika/parser/ner/}ner package}}
+
+* {Image and Video object recognition}
+
+ Tika supports calling out to a number of Object Recognition frameworks to
+ analyse the contents of images and videos. Large training datasets and or
+ frameworks are generally required, often accessed via REST services. The
+ {{{./api/org/apache/tika/parser/recognition/}recognition package}} contains
+ most of these. Details on the requirements and setup steps are generally
given
+ on the {{{https://wiki.apache.org/tika/}Tika wiki}}.
+
Full list of Supported Formats
