Manual CisWeb application A gwt-based web service that provides
Contents
1. CisWeb very easily You only have to add a GUI at the client side and implement the RPC And itis a good chance to present CIS to the rest of the world because everybody who has a browser and internet can access the web service Even from Singapore it worked very well There are many possibilities how to upgrade this application But that s dreams of the future2. ir ee i Create new file View file Rename Copy Cut Download Delete eee eee eee eee ee eee Compress Uncompress file eee eee ee eee ee Picture 4 CisWeb FileMenu When you click with your right mouse button on a fileicon within the Filebrowser a small panel will appear that allows you to execute some operations on the file Double Click on a File will view the content of the file You will only see one part of the file because the server divides it into many parts and sends the parts to the client dynamically to improve the responsiveness Download file is only a download link to the file Delete file deletes the file on the server Compress Uncompress file Compress or uncompress the file to zip Copy Cut copy pastes the selected File into a Clipboard and this File can be coped moved to a new Directory by clicking the right mouse button situated over a directoryicon Create new file Create subfolder Paste Compress BEUuee Upload File Refresh IEFATAME TIT TT Picture 5 Paste a file into a new Directory The user can also upload new files to the server This can be done very easily by using the Upload File Option LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 7 VON 14 5 Computerlinguistic tools 5 1 Open files for the use within Tools Furthermore the user can execute tools which work on his personal files stored in his home directory Every tool works with Input Fi
3. LUDWIG MAXIMILIANS CENTRUM UNIVERSIT T F R M NCHEN INFORMATIONS UND SPRACHVERARBEITUNG Manual CisWeb application A gwt based web service that provides access to language processing tools of CIS Authors Version 1 0 Luitpold Gollas Version 2 0 Sebastian Bildner Version 3 0 Patrick Seebauer APPs from Authors Eduardo Torres Jorg Schuster Sebastian Nagel Max Hadersbeck Thorsten Voblt Florian Fink Project Manager Dr Max Hadersbeck 1 Introduction CisWeb is a web service that provides access to several language processing tools developed at the Centrum f r Informations und Sprachverarbeitung Munchen CIS The development was started during an internship of Luitpold Gollas at the professorship for Programming and Software techniques PST at the LMU from March 07 to August 07 The corner stone of the web service was laid by an application developed at the CIS called CisTex a standalone JAVA application which has to be installed from the user on his client and runs locally That means that every user had to download the whole suite before he could start working with it and every executable had to be compiled for his operating system The next step in developing a workbench for computer linguists was to offer a web service that holds the linguist tools on a server and the user communicates with the tools via an internet browser Everybody should have access to the tools without downloading the whole suite see P
4. a Plugin for your browser If you are experiencing problems please go to this site for testing your setup and for more information To access CisWeb you need a username and a password To get your authentication data ask the ClS system administrator for a new account Click the Log In button to be forwarded to the CisWeb application Authentication is required to establish a working directory on the CIS server for every user in order to save the private files you were working on that you can access them later again LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 4 VON 14 3 Main Screen Until now the main screen dispays the user interface of CisWeb that you will see all the time All other features are shown in popup windows or dialog boxes On the right side you will see the content of your working directory With the help of a Filebrowser you can access the files and navigate through your directories Together with your working directory you have access to a public directory were additional public files are stored see Picture 3 Information extraction EOS Statistics Automaton Convert Preferences Username max Current Language German Sign Out About mittel maxi Home 4 languages J English l J French A German J Russian jw lectures logs A sz_61560A_neu_snt console log derby log fileParams cfg sz_61560A_neu sz_61560A_neu snt Picture 3 Main menu of Ci
5. application is provided Unitex was mainly Test your Java within the Browser developed by S bastien Paumier at the University of Paris Est Marne la Vall e France f It can be accessed via Unitex Open Applet Partners Paris Est Marne la Vallee Software Unitex In order to get this applet running you need to have installed the latest Sun now Oracle Java Plugin for your browser If you are experiencing problems please test your Java setup or go to the wiki for more information wittgensteinsource org Wittgenstein Typescript 213 mit freundlicher Genehmigung von Wittgenstein Source Bergen Text Edition BTE wittgensteinsource org Edited by Alois Pichler in collaboration with H W Kr ger D C P Smith T M Bruvik V Olstad Wittgenstein Archives at the University of Bergen Uni Digital Research Cooperation Alois Pichler and Max Hadersbeck Autors of CisWeb Luitpold Gollas Eduardo Torres Schuhmann Sebastian Bildner Current maintainer Patrick Seebauer Projectleader Dr Max Hadersbeck Autors of the computational tools are serveral students and staff of the CIS Username Password Login Brummen ooun MM Picture 2 Login screen of CisWeb On this page you can login to CisWeb and can test if your Browser has installed JAVA Plugins properly because some tools use them for using complex graphical Applications see In order to get this applet running you need to have installed the latest Jav
6. enstein LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN 0 40 148 zp gt lt s b Ts 213 wr n 16 gt Wie h ngen unsre Gedanken mit den Gegenst nden zusammen ber die wir Zeiger auffassen will aufa t s gt lt p gt lt p gt lt s b Ts 213 9r n 7 gt Wir k nnten uns den Marsbewohner sen wir dann diese Schrift als eine Sprache zur Beschreibung des Lautbildes auf so k nnte man sich rochen lt s gt lt p gt lt s b Ts 213 40r n 13 gt Ein Philosoph k nnte glauben einen Satz Ich k nnte mir da dies in einem Sinn nicht m glich ist lt s gt lt s b Ts 213 50r n 1 gt Ich kann mir eine Abmachung hm blau so hat er gelb zu malen u s w immer die komplement re Farbe und daher kann ch mir auch auch ohne eine vorhergehende Abmachung so deutet lt s gt lt s b Ts 213 50r n 2 gt Ich kann mir ferner mer eine gelblichere als ch Dir zeige und wieder kann ich mir die Deutung auch ohne m Vorgang des Kopierens Abbildens enthalten also aus ihm eindeutig abzulesen lt s gt lt p gt denken denken denken denken denken denken denken denken m SEITE 10 VON 14 Username max Cu lt s gt lt p gt s b Ts 213 wer n 17 gt Wie treten diese Gegenst nde in unsre Gedanken ein lt s gt lt s b Ts _ der auf der Erde erst nach und nach den sesichtsausdruck der Menschen als solchen 11 verstehen lernte da Einer d
7. icture 1 Advantages of this approach are that the tools have to be updated only on the server and every user has access to the latest versions all the time Furthermore for the CIS it is sufficient to compile only a Linux version of the executables The users who work with the web service are only provided with a graphical interface that will be shown in their browser window With this graphical interface the calls to the server side tools are performed and the server s results are presented as graphical output within the user s browsers or stored as files in a personal home directory on the server To maintain a home directory for every user the users must register them self if they start the web service at the server LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 2 VON 14 Client Server Architecture Picture 1 Client Server Architecture of CisWeb As CisWeb is a Client Server application it has to satisfy more demands than a local application like CisTex or Unitex CisWeb offers user management for different roles like administrators tutors or students Every registered user is associated to his own folders where he can upload download compress uncompress view or delete files Beside the private files the web service offers a public area where files can be exchanged between the administrator and the users With the help of the CisWebinterface the users can start end of sentence recognition concordance calls and language recogn
8. iese Schrift so auffal te als entspr che einfach jedem Buchstaben ein Laut und als h tt da ein Philosoph glaubte einen Satz mit in roter Farbe drucken lassen zu m ssen da er erst 50 wonach Einer dem ich eine gr ne Tafel zeige und sage male mir diese Farbe mir ein Rot malt wen da Einer meinen Befehl auch ohne eine vorhergehende Abmachung so deutet lt s gt lt s b Ts 213 50r n _ dal die Abmachung gelautet h tte auf den Befehl male mir diese Farbe male immer eine gelbliche lt is gt s b Ts 213 50r n s gt Aber kann man sagen dal einer ein rotes T felchen genau 2 lt s b T3 213 iv r n 9 gt Verk rpert der Vorgang des Abbildens sozusagen diese Regel lt s gt ee lt s b Ts 213 wo n 10 gt 5 274 lt s gt eae lt p gt lt s b Ts 213 wer n 11 gt Wie rechtfertigt man das Resultat der Abbildung mit der allgemeinen cin gar Regel der Abbildung lt s gt lt p gt Ve ee Es b 1s 213 10 7 n 12 gt 5 279 lt s gt Error EENE lt p gt lt s b Ts 213 iv r n 13 gt Der Vorgang der absichtlichen Abbildung der Abbildung mit der a SS ENE Intention abzubilden ist nicht wesentlich ein psychischer innerer lt s gt lt p gt espe are m m zs b Ts 213 iv r n 14 gt Ein Vorgang der Manipulation mit Zeichen auf dem Papier kann dasselbe 000200 leisten lt s gt ne renee lt s b Ts 213 1v r n 15 gt 5 203 lt s gt EEE Een lt p gt lt s b Ts 213 w
9. ile 5 7 Automaton TGC Gramatics Tgc Transducer Grammar Compiler compiles Transducer Grammars to openfst Transducers implemented from J rg Schuster 5 8 Convert Recode recodes a File from a selectable Encoding to another Encoding Strip HTML Stripps all HTML Tags within an html formatted File and produces a Textfile out of it LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 14 VON 14 6 Logout Info s and Wiki s of CisWeb Current Language German Sign Qut About Picture 15 CisWeb logout About In the upper right corner of CisWeb you will find the logout Option of CisWeb In this corner you also find an Option About were you can find an Link to the CisWeb wiki where you get additional Documentation about CisWeb can deliver wishes for new features which should be implemented in CisWeb In the Wiki you also find a WiKi Page where you can report error s which happen in CisWeb 7 Prospect The current version is only a picture of the state of the art a lot of useful functionalities are not yet implemented However the range of use of this application is big One could think of using it as a lecture attendant tool where the professor tutor can make his text files available in the public folder The students could work with the tools provided by this web service without having to use a shell to execute the tools but a graphical user interface Furthermore every command line tool can be integrated into
10. ition on these files and the results can be seen in the browser window As a scenario one could imagine that a tutor provides some text files in the public folder of the web service so that every student in class has access to these files He can work on the files and upload the results again Due to the role management the tutor can collect the results from the student s folders A short description of how one can use the CisWeb will be provided in the following sections The workbench is based on the public Domain Tools gwt Google Web Toolkit http code google com intl de DE webtookkit Apache Derby Database http db apache org derby and Apache Tomcat http tomcat apache org LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 3 VON 14 2 Login Screen The entry point address is http cisweb cis uni muenchen de This lead so a forwarding to the real address of CisWeb which may be changing sometimes The first screen you will see will be a login screen Picture 2 Cisweb Computerlinguist s workbench Mozilla Firefox Datei Bearbeiten Ansicht Chronik Lesezeichen Extras Hilfe a Cisweb Computerlinguist s workben AKTUELLES CisWeb Computational linguist s workbench Links a mee This service provides access to software in the field of computational linguistics The Cis Web Manual Keine Neuigkeiten tools were programmed during software seminars at the CIS perros Among others an applet based on the Unitex
11. itstehende Lastwagen hebt Der Sherry Clan ist berall Auch sonst sind die ber hmtesten S hne meist Adoptivkinder der Region im Stadtbild von Jerez allcecenwartic Nach Alvaro einem weiteren Snro Picture 7 CisWeb File Viewer LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 8 VON 14 5 3 Information extraction Information extraction EO Open Applet Run script Graph Editor Concordance Suche F Wittgenstein Menu allows you to start 4 different tools a Open Applet here you can call the Computerlingustic Tool UNITEX which is developed as a public domain tool at Universite Paris Est Marne la Vall e see http www igm univ miv fr unitex The tool is opened in a Separate TAB CisWeb unitex of your Browser Cisweb unitex Mozilla Firefox L y Datei Bearbeiten Ansicht Chronik Lesezeichen Extras Hilfe Cisweb unitex x amp Cisweb Workbench x gt a http cisweb64 cis uni muenchen de 8090 CisWeb CisWeb html iL VETE 4 Text DELA FSGraph Lexicon Grammar Edit File Edition Windows Info Picture 8 UNITEX as JAVA Applet b Run Script here you can run UNITEX with a PERL Script via the command line LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 9 VON 14 c Concordance Here a new Concordance Tool is under Development with allows to locate pattern in a text file which can be specified as local grammar The tools accepts a Local Grammar as Gra
12. les which can be selected within Tool Dialog Box You have to select a file in the fileview menu and click on Choose file in the Tool Dialog Box Home klein tet Klein bet German Language Picture 6 Choose Files for tools 5 2 File Viewer The file viewer Picture 7 CisWeb File Viewer can handle files of different encodings The server automatically detects the encoding of the file and sends it with UTF 8 encoding to the client but the real encoding will be shown to the user So no matter in which encoding the file is stored at the server the client always sees the right umlauts As text files can be large and the downloading and displaying of the whole file would take a lot of time the files are loaded dynamically That means that only one part is downloaded to the client when he wants to view a file The user navigates through the file by clicking on the next or previous link Maxi 4 0 Home Save Encoding UTF 8 lt previous 1 of 33 next gt In den Bodegas von Jerez de la Frontera den wahren Heiligt mern der Stadt lernt der Besucher die jahrtausendealte Weinkultur dieser Region kennen Doch im Gegensatz zu den Hunderttausenden von gew hnlichen Besuchern die jedes Jahr in die k hlen schattigen Hallen mit ihren dicken Mauern ihren hohen Decken Pfeilern Kapitellen Rundb gen und Kuppeln pilgern haben die prominentesten G ste mit Spezialkreide ihren Namenszug auf dem dunklen H
13. olz einiger hier lagernder F sser hinterlassen Ein Speziallack sorgt daf r da die Unterschriften weder abbr ckeln noch verblassen Die Normalsterblichen hingegen sollen m glichst keine Spuren hinterlassen den Schlaf des kostbaren Inhalts der riesigen Tonnen nicht st ren m Die Bodegas sind die wahren Heiligt mer von Jerez de la Frontera und mit der entsprechenden Andacht vollzieht sich ihr Besuch In und um die Stadt im u ersten S dwesten Andalusien stehen mehrere Dutzend Sherry Tempel benannt nach dem lateinischen Wort apotheca was Waren oder Weinlager zu ebener Erde bedeutet Dort und eben nicht in muffigen Kellergew lben stapeln sich etagenweise mehrere hunderttausend F sser von denen jedes normalerweise f nfhundert Liter Rebensaft enth lt Die ersten Bodegas errichtete 1730 Pedro Domecq Sohn einer nach S dspanien ausgewanderten franz sischen Adelsfamilie Ihrem Beispiel folgten Weinbauern und h ndler aus England Schottland Irland und Holland so da heute die weltweit wichtigsten Sherry und Brandy Produzenten fest in Jerez etabliert sind Sandeman Williams amp Humbert Wisdom amp Warter Burdon Gordon Harvey Croft Ihre Pr senz kann der Besucher f rmlich erschnuppern wenn er durch die Stra en und Gassen der Stadt flaniert hier hinter offenen Eisentoren einen Blick auf riesige Aluminiumtanks erhascht dort einem Gabelstapler begegnet der versandfertige Kartons mit den bekannten Logos auf bere
14. ph File and a Electronic Dictionary in DELA Format With the help of the Dictionary the tool can work lemmatized on the text The tool uses an implementation of automaton developed by Florian Fink CIS 2011 The Concordance is programmed from Patrick Seebauer CIS 2010 Choose file Graph File Choose fille Dictionary File Choose file W Hide Advanced Options Left Stop at last SF Right Stop at last Sy 100 Max Occurences Lenght Picture 9 Concordance Search and Concordance Viewer d Suche gt Wittgenstein This tool is connected to the Big Typescript of Ludwig Wittgenstein http www wittgensteinsource org Wittgenstein Typescript 213 mit freundlicher Genehmigung von Wittgenstein Source Bergen Text Edition BTE wittgensteinsource org Edited by Alois Pichler in collaboration with H W Kr ger D C P Smith T M Bruvik V Olstad Wittgenstein Archives at the University of Bergen Uni Digital Research Cooperation Alois Pichler and Max Hadersbeck lf your CisWeb Account has the Right to read the TS 213 Big Typescript the tool locates the Pattern which is supplied in the Menu The tool uses an implementation of automaton developed by Florian Fink CIS 2011 The Concordance is programmed from Patrick Seebauer CIS 2010 Information extraction FOS Statistics Automaton Conv denken Jptions Picture 9 Concordance Search and Concordance Viewer connected to the Big Typescript of Ludwig Wittg
15. r n 16 Wie h ngen unsre Gedanken mit den Gegenst nden zusammen er ber die wir denken lt s gt lt p gt es b Ts 213 10 7 n 1 gt Wie treten diese Gegenst nde in unsre Gedanken ein lt s gt s b Ts 213 0 r n 18 gt Sind sie in ihnen durch etwas Andres etwa hnliches vertreten lt s gt lt s b Ts 213 iver n 19 gt Wesen des Portr ts die Intention lt s gt Picture 10 Output of the Concordance and Search Tool 5 4 EOS End Of Sentence recognizer LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 11 VON 14 EOS Statistics Wastl s EOS EOS gt EOS 2 gt Picture 11 End of Sentence Menu Behind the EOS Menu 3 different End Of Sentence Tools are hidden A textfile Choose File is parsed and at the end of a sentence an EOS mark is added The Result is stored in an output file Output filename in your working directory If you don t specify an output Filename a resultfile with the name Input filename eos will be added to your working directory a Wastl s EOS This tool is developed from Sebastian Nagel a former researcher at the CIS See http www cis uni muenchen de wastl Options Enable EOS detection S Each snt in separate line P Text instead of file de win cp1252 de us utf 5 de english iso 888 en win cp1252 en u ut 5 ru russian iso 6659 5 ru win cp1251 ru us utf amp Execute Ad
16. sWeb with Filebrowser and Menu LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 5 VON 14 Furthermore you can upload files from your local client to your working directory on the server with the Upload file box If you log on the web service for the first time your working directory may be empty because you have not uploaded files yet and there are no files in the public folder You must click with the right button of your mouse which has to be positioned over a Directory Create new file ww g Create subfolder gl i Paste 7 ee Compress Upload File ee Refresh Picture 4 Upload File and Manage Directories Left to the CisWeb Filebrowser you see the CisWeb Menu with all of the CIS Tools that can be accessed over the web service will go more into detail on these tools in Features Information extraction This tool offers various tools to extract Phrases from a text file with the help of Local Grammars a Open Applet Here you start the UNITEX Programingsystem developed at University of Paris Est Marne la Vall e France see htip www igm univ mlv fr unitex Run Script Here you can start the UNITEX Programingsystem from a PERL Script Graph Editor in work This will offer soon a Graphical Editor for Local Grammars Concordance This implements a search flotex and build concordance tool suche This search tool is connected to the Big Typescript 213 of Ludwig Wittgenstein http www wi
17. ttgensteinsource org Wittgenstein Typescript 213 mit freundlicher Genehmigung von Wittgenstein Source Bergen Text Edition BTE wittgensteinsource org edited by Alois Pichler in collaboration with H W Kr ger D C P Smith T M Bruvik V Olstad Wittgenstein Archives at the University of Bergen Uni Digital Research Cooperation Alois Pichler and Max Hadersbeck f EOS 2295 This tool offers various end of sentence recognizer Statistics Here you will find a tool that automatically detects the language of a text file and another tool which can produce Frequency Lists Automaton A tool that implements a Tgc Transducer Grammar Compiler It compiles Transducer Grammatics to openfst Transducers The tools use the Opensource Transducerimplemenation from AT amp T Convert Here you can Recode the Coding of a textfile and Strip HTML Code from an html File Preferences Here you can change your Password and select your Language which is used from Unitex LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 6 VON 14 4 Features of CisWeb In the last section have already introduced all of the tools that are or will be integrated into CisWeb When you read this manual there might be other tools available because the application is still under development Additionally to these tools other features are implemented as well File operations
18. vanced options Picture 12 End of Sentence detection from Sebastian Nagel This tool allows to Segment German Englisch and even Russian Textfiles stored in different encodings b EOS This is the first version of an speed optimized C Version EndOfSentence See Demopage http maxdemo cis uni muenchen de home demos eosv1 index html LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 12 VON 14 Home klein tt sheet klein tet German Language Picture 13 End of Sentence detection First C Version c EOS This is the actual C Version of EndOfSentence Detection This version is permanently enhanced and new languages are added French Italian Croatian Russian See Demopage http maxdemo cis uni muenchen de home demos eos index html See Readme File http maxdemo cis uni muenchen de home demos eos readme htm Input Homej sz bt eos Choose file File Output File W Hide Advanced Options Regex Internal Choose file Agent EOS Mark ens IPrimus agent YlAbbreviation agent Ylcross agent Muttilin e Paragraph Detection nded Tagging Print All Tags lone Eos per line YINormalize Picture 14 End of Sentence detection actual C Version LUDWIG MAXIMILIANS UNIVERSIT T M NCHEN SEITE 13 VON 14 5 6 Statistics Language Detection of a language of a textfile Caution Files with UTF16 Encoding are not supported Frequency List Produces a Frequency List of an utf8 textf
Download Pdf Manuals
Related Search