gmane.comp.apache.uima.general

Re: Lucene cas consumer

Dan McCreary — 2008-12-04T19:32:43

Hello, I am somewhat new to UIMA so I apologize if I misunderstand some things. But this is a very interesting question for me. I see Lucene as a very wildly adopted but *Java-only framework* of tools for building and maintaining keyword *indexes *on many types of documents. Lucene also has great support for HADOOP and MapForce-type saleability. But Lucene is also designed to work with many front end tools like POI libraries to extract text from Microsoft Word, Excel, PowerPoint etc. I see Apache UIMA as a general purpose *analytic pipeline architecture *with the strengths of a very advanced common in-memory processing model. I thin there is a huge win-win for both projects if we can make UIMA enrich text documents with entities before they are indexed by Lucene and also make these tools much easier to install and work together. You should not have to be a Java developer just to install these tools and have them index and search our file systems. I have spent many hours trying to get UIMA to work with

Re: Lucene cas consumer

Niels Ott — 2008-12-04T18:36:59

Hi all, I'm using both Lucene and UIMA in one project. Lucene is primarily an information retrieval API. It provides a framework and default implementations for analyzing several languages. Analyzing means tokenization, stop words, etc. Furthermore, it brings the key functionality to build an inverted index and to search it. Lucene can be extended easily. E.g. one can implement an analyzer that does lemmatization or that looks up synonyms in Wordnet and adds them to the index. What Lucene cannot do - or at least not without a lot of hacking - is aggregating analyses as UIMA can using the CAS. Usually your knowledge grows during an UIMA-based NLP-pipeline: you add the a token annotation, a lemma annotation, a POS-annotation and so on... In Lucene, you have the classical pipeline: the output replaces the input. (Yes, by subclassing Lucene's "Token" class, one can fiddle around the issue, but it is not elegant at all.) What makes Lucene + UIMA interesting for me is a simple fact: I can do all the NLP I w

Re: Lucene cas consumer

Greg Holmberg — 2008-12-04T18:12:18

Roberto-- It does seem like there should be a close relationship between the two groups. I don't know much about Lucene--can you educate me? For example, have you given any thought to what to do with UIMA annotations? From what little I've read about Lucene, they seem to have a thing called a document analyzer, but they don't mean the same thing we mean by analysis in the NLP community. They appear to mean something more like "tokenizer". So I haven't yet found a place to put UIMA annotations, say for example, named entities or parts of speech. I'm wondering if Lucene needs a major feature enhancement before its truly useful with UIMA? What are your thoughts on how the integrate the two? What functionality is possible? Greg Holmberg -------------- Original message ---------------------- From: "Roberto Franchini" public.gmane.org>

Printing UIMA PDF documentation

Thilo Goetz — 2008-12-04T09:07:43

There's a problem with printing some of the current UIMA PDFs. Some pages will not print. Marshall has fixed this and regenerated the docs on our website. So if you're having problems with printing the docs, get the PDFs from the website. This will be fixed in the next release. --Thilo

Lucene cas consumer

Roberto Franchini — 2008-12-03T15:06:23

Hi, I'm going to write a Lucene CAS consumer. The porpouse is to create a Lucene document, or more than one, for each CAS. Last year (2007) the JENA university lab (JULIE lab? is it right?) delivered such a component, named LUCAS. Then it disappeared. LUCAS seems a good piece of software. The Technische Universität Darmstadt developed one too: http://www.ukp.tu-darmstadt.de/projects/dkpro/. (I will write to them). There's anybody interested to share knowledge and/or code to do that component? I think that Lucene and UIMA can be very good friends :) Roberto PS: I apologize for my bad English.

Re: Setting Aggregate AE Parameters

Niels Ott — 2008-11-26T18:51:14

Hi Christoph, yes, the specifier actually should have been named description. I followed your instructions and they worked for me. Thank you very much! I'm using a Properties object and of course I can only obtain String objects from it. settings.setParameterValue(name, value) requires the type of value to match the type specified in the descriptor. Here's a little trick that does it: private Object autoConvertParameter(Object originalParameter, String value) { if ( originalParameter instanceof Boolean) { return Boolean.parseBoolean(value); } if ( originalParameter instanceof Integer) { return Integer.parseInt(value); } if ( originalParameter instanceof Float) { return Float.parseFloat(value); } // fallback: return as string return new String(value); } Object genericTypeValue = autoConvertParameter(settings.getParameterValue(key), value); settings.setParameterValue(key, genericTypeValue); If there is a more elegant solution, let me know. The hack above will of course fail in certa

Re: Setting Aggregate AE Parameters

Christoph Büscher — 2008-11-26T17:31:10

Hi Niels, you might want to try to modify what you call the 'specifier' (I guess its a AnalysisEngineDescription) before you call UIMAFramework.produceAnalysisEngine(specifier) I used this in Unit tests during setup, something like this: -------- URL descResource = LexProcessingWriterTest.class.getClassLoader().getResource( "MyDescriptor.xml"); AnalysisEngineDescription) aeDescription = UIMAFramework.getXMLParser().parseAnalysisEngineDescription( new XMLInputSource(descResource)); ConfigurationParameterSettings settings = aeDescription.getAnalysisEngineMetaData().getConfigurationParameterSettings(); ------------ Now that you have the settings you can set/overwrite values like: settings.setParameterValue(name, value); At the end when you produce the analysis engine with the modified values there shouldn't be any problem: myAE = UIMAFramework.produceAnalysisEngine(aeDescription); Hope that helps, Christoph Niels Ott schrieb:

Setting Aggregate AE Parameters

Niels Ott — 2008-11-26T17:02:11

Hello everybody, I'm new to this mailing list, however I'm not so new to UIMA. Here's my issue: I am trying to set the parameters of an Aggregate AE from within my Java program. I tried the following the solution I found here: http://www.mail-archive.com/uima-user-d1GL8uUpDdXTxqt0kkDzDmD2FQJk+8+b< at >public.gmane.org/msg01241.html# which is basically to produce an annotator and then to configure it using annotator.setConfigParameterValue("ParamName","NewValue"); This does not work for me as annotator = UIMAFramework.produceAnalysisEngine(specifier); which I'm using before to produce the Aggregate initializes its AEs with the default parameters. These default parameters are of course not suitable for use in the given scenario, causing the initialization to fail before I correct the configuration. I tried to use produceAnalysisEngine(ResourceSpecifier aSpecifier, Map aAdditionalParams) instead but the additional parameters don't do the job. Perhaps I misunderstood the purpose of

Re: Additional Runtime Arguments

Thilo Goetz — 2008-11-26T13:27:23

UIMA doesn't really support this. The reason is that the UIMA philosophy is that all data must live in the CAS. Otherwise your analysis is not portable, remotable etc. Howsoever that may be, there are various ways you can get around this restriction, if you really need to. For example, if your documents have a unique ID, you can write that into the CAS somehow. Then you could have a static factory where your annotators can go and get the object for the document ID they're working on. Somehow, you also need to clear this cache after processing has finished. You should *not* subclass the CAS, for any number of reasons. HTH, Thilo

Re: Additional Runtime Arguments

Eddie Epstein — 2008-11-26T13:11:41

Not sure if this will help, but the CAS supports a concept of multiple "views", where each view is itself a CAS object. The process(cas) method can then be passed a specified view, so that the annotator only deals with a subset of the entire CAS. Some have used views to deal with complex documents, creating a view for each of several document parts and then passing specified view(s) to the appropriate AE. For more on CAS views, see http://incubator.apache.org/uima/downloads/releaseDocs/2.2.2-incubating/docs/html/tutorials_and_users_guides/tutorials_and_users_guides.html#ugr.tug.mvs Regards, Eddie On Tue, Nov 25, 2008 at 4:11 PM, Thomas Polzin public.gmane.org> wrote:

Additional Runtime Arguments

Thomas Polzin — 2008-11-25T21:11:51

I probably missed something in the documentation but here is my problem: I'm trying to process a pretty complex structured document and I do not want to translate the entire structure into annotations. Instead I would like to pass some object containing this additional structure to the respective analysis engines. Note that a static object will not work. Also, I do not want pass this object to the engines when they are constructed/initialized. The best solution I came up with so far is to subclass the CAS object and then to downcast to it in the process( CAS cas) function. Does this make sense? Is there a better solution? Thanks so much in advance Thomas

Re: Mailing list in a forum

Jukka Zitting — 2008-11-25T10:35:55

Hi, On Sun, Nov 23, 2008 at 7:19 PM, Eric Quinn public.gmane.org> wrote: You might also want to try Nabble that works pretty well for this sort of use. BR, Jukka Zitting

Re: Mailing list in a forum

Marshall Schor — 2008-11-25T04:26:20

GMane seems to be somewhat broken. The forum views for the uima-users list seems to be "missing" many posts, including the original one from Eric Q. However, the "RSS Feed" view has all the posts, I think. Not sure what the problem is. I've attempted to contact the site maintainers but have gotten no response. Here's the details: I can see some messages for the uima-user (apache) mailing list at http://dir.gmane.org/gmane.comp.apache.uima.general are appearing in the RSS feeds, but not on the web in either of the Frames/Threads or the Blog-like interfaces. Examples of messages on the RSS feed, but missing in the others are the ones dated: Sunday Nov 23 2008 1:19 pm (Mailing list in a forum) and Tuesday, Nov 18, 2008 11:21 AM (Donation of a widely used ....) Another thing: The graph on http://dir.gmane.org/gmane.comp.apache.uima.general seems to be for a date range of 6 june 07 thru 1 October 07. and the date range for http://dir.gmane.org/gmane.comp.apache.uima.devel is 1 April 07 to 1 Oct 07. Our ma

Assistant Scientist position available at the MBLWHOI library

Hamed, Ahmed A. — 2008-11-24T22:16:46

Apologies for cross-listing! Go to : https://mbl.simplehire.com/applicants/jsp/shared/frameset/Frameset.jsp Under Scientist link: *Assistant Scientist* View 01-10-2008 Scientist MBLWHOI library Posting Details Posting Number: 0000005 Reference Code: Department: MBLWHOI library Position Type: Scientist Position Title: Assistant Scientist Position Summary: The MBLWHOI Library at the Marine Biological Laboratory seeks highly motivated information scientists to join a growing team of library-based researchers. The successful candidate will join a community of librarians and researchers that are internationally recognized as defining the current trends and practices in marine information sciences and biological/biomedical informatics. Additional Information: In addition to providing leadership in advancing current research projects at the Library (such as the uBio and MBLWHOI Library Herbarium projects), the successful candidate will develop an extramurally funded research pro

Re: Mailing list in a forum

Marshall Schor — 2008-11-23T18:32:09

It works, but took maybe about 5 minutes... Let's see if this message gets "threaded" in the GMane view. -Marshall Eric Quinn wrote:

Mailing list in a forum

Eric Quinn — 2008-11-23T18:19:34

Hello, this is Eric Quinn. I am working with Marshall Schor to find a way to view things sent to the mailing list without being forced to subscribe to the mailing list. This is a test to see whether it is also possible to send the things to the mailing list using Gmane, the website we are testing with this message. Gmane has been mirroring the mailing list since approximately May 2007, and you can view it here: http://news.gmane.org/gmane.comp.apache.uima.general/ Thank you.

Re: UIMA CPE GUI issue

Matt McClain — 2008-11-18T19:32:29

Eddie, They were a couple of parameters in an analysis engine that I wrote. I haven't tried to replicate this on any built-in analysis engines. Thanks for the tip about the cpeGui, too. I've been using it as a quick way to create CPE descriptor files, but maybe I should stop that. Thanks again, Matt

Donation of a widely used type system for multi-modal text analysis

Burn Lewis — 2008-11-18T16:21:13

NOTE: This was foolishly posted to just uima-dev in September. The type system and some of the sample code are used in the recently mentioned UCC tool from CMU. We would like to add a UIMA type system and sample annotators to the Apache Incubator project as an example of a rich multimodal application. Our hope is that others will find the techniques and types useful, and will find it a good starting point for developing other multimodal applications. GTS is a type system designed for multi-modal applications that combine analytics from multiple sources and modalities, such as speech recognition, language translation, entity detection. etc. It is currently used by 10 cooperating groups participating in the Darpa GALE project ( http://www.darpa.mil/ipto/programs/gale/gale.asp) to transcribe, translate, and extract information from foreign language news broadcasts. This application requires that all the data is cross-referenced so that, for example, any English sentence can be traced back to the precise r

granting code to Apache UIMA project

Eric Riebling — 2008-11-18T15:48:25

Greetings, Carnegie Mellon would like to grant code to the Apache UIMA project, in the form of two support tools we have created: * UIMA Component Repository (UCR): A web portal for upload/download/review of UIMA components, type systems, etc. Based on Java and existing tools (e.g. Tomcat). * UIMA Component Container (UCC): A Java system that supports server-side execution of distributed UIMA services on data supplied from a client web browser. This message is intended to initiate discussion about whether such a grant makes sense, and if so, what the concrete next steps should be in preparing the software. (On the legal side, we have engaged CMU's tech transfer office regarding the grant letter from CMU, and we are working out the details with them now.) Thank you for your consideration, Eric Riebling, Senior Research Programmer Eric Nyberg, Professor School of Computer Science Carnegie Mellon University

Assistant Scientist position available at the MBL

Hamed, Ahmed A. — 2008-11-14T16:33:10

Apologies for cross-listing! To apply: You must fill out the online application : https://mbl.simplehire.com/applicants/Central?search=scientist Thanks, Ahmed

U-Compare first public release announcement

Yoshinobu Kano — 2008-11-14T16:05:29

Dear UIMA users, The U-Compare Initiative is pleased to announce the first public release of U-Compare system ( http://u-compare.org/ ); an easy-to-use system for the construction and comparison of NLP workflows. U-Compare is an integrated text mining/natural language processing system based on the UIMA Framework which provides access to the worlds largest repository of type system compatible UIMA components. The components come from a wide range of different research groups, and currently include corpus readers, syntactic parsers, semantic entity recognizers, as well as a variety of other tools. New components are being added all the time, and we expect to be adding a selection of more resources like relation extraction tools in the near future. From these, or any UIMA components, U-Compare allows users to build complex NLP workflows by simply dragging-and-dropping components into order. These workflows may then be run from within the U-Compare system which automatically makes available a range of data v