Atomic User Guide v0.1.7
Contents
1. Atomic User Guide v0 1 7 Stephan Druskat Friedrich Schiller University Jena Copyright 2014 Friedrich Schiller University Jena Table of Contents Getting Started ri ae e eet hae eek het e ee ae ess ca de En ine cote reS 1 RESOUDCES E sccces nee reece sees cessach se ssaagesadaaass pi bang sh Coscegs ebosa od sees ee usslaveeas Ess odaw adh seomag essen S 1 THE workspace esete woes cosh Se ksteie a E desta ves coubareeveinen ied ity R EEEE desta EEES 1 Creating a Project ornest an ieron o ean e E tess E sotng ass E edendanes EREET 2 Annotation with the Annotation Graph Editor ssessssseesseressreersrrrrrrresrerrerrerrerrerreresree 3 Referee n ra Sones SPs esses ote ads oda duasasete ase vag ates odh ES OT EEEE E EEES REEE PE aE EELEE 7 AtOMICAL oieee reeet esonero tu EEVEE cost Weedsas yeenossbuycnvsteeudeoss bu nyesta veuuessbuedhestyeoueenss TIES 7 Install new plugs si dsstessabates eapstae a a a a ses ba dade a Mamta bea tooled ss 8 Getting started This guide gives a workflow example for annotating a corpus with Atomic Resources If you want to know more about Atomic please refer to the Atomic website http linktype iaa uni jena de atomic If you want to receive notifications about upcoming changes new releases etc or get in touch with other Atomic users please subscribe to the Atomic Users mailing list atomic user listserv uni jena de https lserv uni jena de mailman listinfo atomic user If you2. create a new node Then either drag and drop the tool onto a position of your choice in the editing area or click on the editing area at the position of your choice Drag and drop will create the node but will not open the annotation field for direct annotation of the newly created node Click and point will create the node and open the annotation field for direct annotation of the node If you have selected any nodes in the editing area before using a node tool the new node will be created and edges will be created from it to all of the selected nodes When using the Structure Tool the type of the newly created edges will be dominance relation When using the Span Tool the type of the newly created edges will be spanning relation and edges will only be created to selected nodes of type token as spanning relations only accept tokens as target nodes Creating edges To create an edge click on the respective tool then click on the node that should be the source of the edge and subsequently on the node that should be the target of the edge Atomic User Guide v0 1 7 The edge will be created but the annotation field will not automatically open In order to annotate edges please see Annotation of elements Annotation of elements In the Annotation Graph Editor annotations have a key and a value and optionally can be assigned to a namespace Annotations are entered in the annotation field using either of the following patte
3. with a null value marks an existing annotation with key key for deletion key value namespace key value key Atomic User Guide v0 1 7 Example a ni n3 t2 t5 d3 o4 x2 yoy Zes Deleting elements and annotations For the deletion of annotations see the section called Annotation of elements To delete elements use the d command It takes as arguments e an element or a list of elements or a set of elements or a combination of the three Example dni n5 dl 03 Reference AtomicAL Note that some AtomicAL commands can be issued not only for single elements but also for sets of elements where all elements are of the same type and have subsequent IDs The following command for example groups all tokens from tZ to t5 i e t t2 t3 t4 and t5 under a new parent structure node which is annotated with cat NP Pp 1 5 Cati NP Also note that whenever a command takes key value or key commands these can be grouped I e to annotate element n7 with annotations x x y y and z z you can do so in one single command as shown below a nl x x y y Z z Table 1 AtomicAL commands Command Arguments Syntax example help Show help help n New structure node key value n pos np s New span node element element key value s t1 t2 t3 cat NP element element key val s t1 t10 cat NP ue e New edge type b source target key e d n1 t3 cat
4. Atomic s Install Wizard Available Software Check the items that you wish to install af Work with Software Site For additional Eclipse plugins http download eclipse org releases indi v Add Find more software by working with the Available Software Sites preferences Name Version Ser Ter gee COMPONE mS Tor COUpsTe DIT VEU TTI VTL Se Eclipse CVS Client 1 3 100 v201 eae ve Eclipse EGit 1 3 0 201202151440 r amp Eclipse EGit Mylyn GitHub Feature 1 3 0 201202151440 r 4 Eclipse JGit 1 3 0 201202151440 r Select All Deselect All 1 item selected Details An Eclipse Git Team provider in pure Java Show only the latest versions of available software __ Hide items that are already installed Group items by category What is already installed C Show only software applicable to target environment Contact all update sites during install to find required software Example Install Git support in Atomic In order to be able to version your corpora and annotations and enable them for collaborative work you may want to use a versioning system such as git Subversion or similar And in order to be able to use these CVS from within Atomic you may want to have CVS integration in the application itself instead of having to use an external command line There are at least one or more plugins for this purpose available from the Eclipse p2 repository for all major version control systems In the case of git the plu
5. NP value a Annotate element key value anl pos np element key Deletes the annotation with key key d Delete element d element element dnl n2n3 d element element dnl nlO Atomic User Guide v0 1 7 Command Arguments Syntax example p Group under new parent p element element key val p nl n2 cat NP p element element key val p n1 n8 cat NP For reasons of readability namespace annotations have been left out completely Please note that whenever a command can take a key value it can also take a namespace key value argument Cf Table 2 AtomicAL edge type flags Also note that if a flag has been set in an e command before this flag can be re used in subsequent uses of the command i e the edge type flag remains active until a new flag is explicitly set In order to re use the active flag don t specify a flag with the command e g e n1 n2 Table 2 AtomicAL edge type flags Flag Edge type d Dominance relation s Spanning relation Pointing relation Order relation Install new plugins In Atomic you can very easily add new functionality to the application yourself at runtime This is done by installing additional plugins in Atomic There are generally two way to achieve this depending on how the plugin is being made available e Install new plugins via the Eclipse Marketplace e Install new plugi
6. clipse is a pdf viewer plugin for Eclipse One of its notable features is a complete integration of SyncTeX which allows forward and inverse search in pdf by B von Loesch EPL Install pdf viewer latex PDF SyncTeX B o lt Back Next gt Cancel Finish In order to install a new plugin go to Options Install new plugins from Eclipse Marketplace This will open the Eclipse Marketplace client Its use is fairly self explanatory You can browse different categories or search for plugins Once you have found a plugin you want to install simply click the provided install button next to the search result and follow the instructions Atomic User Guide v0 1 7 Example Install XML editing support in Atomic If you are working on corpus annotations available in an XML dialect such as SaltXML which is used in Atomic you may want to be able to edit the source files comfortably with a dedicated XML editor There is a variety of solutions available from the Eclipse Marketplace including XMLEspresso http marketplace eclipse org content xmlespresso xml editor eclipse plugin which can automatically gener ate grammars for unknown XML schemas and DTDs To install this plugin do the following e Open the Eclipse Marketplace Client see above e Search for xmlespresso The first result should be the XMLEspresso plugin e Click the install button e The EM client will calculate requirements and dependencies Once its fin
7. displays the name of the document it is currently working on as well as the first few characters of the corpus text For a complete list of available commands see AtomicAL reference Creating nodes To create a new structure node use the n command It takes as arguments e one or more annotations in one of the following forms key value namespace key value Atomic User Guide v0 1 7 Example n syntax cat NP lemma Atomic To create a new span node use the s command It takes as arguments e one or more annotations in one of the following forms key value namespace key value Example s coref reftype group text The board Creating edges To create a new edge use the e command It takes as arguments e an edge type flag which defines the type of the edge to be created For a list of valid flags see Table 2 AtomicAL edge type flags Also see there for an explanation of how edge type flags are re used e exactly one source node referenced via its ID e exactly one target node referenced via its ID e one or more annotations in one of the following forms key value namespace key value Example e p n2 n1 coref coreftype anaph Annotation of elements To annotate elements use the a command It takes as arguments e an element or a list of elements or a set of elements or a combination of the three e one or more annotations in one of the following forms where the last one an annotation
8. ect Restore from Local History Properties Alt Enter amp pcc2 12 Atomic User Guide v0 1 7 Figure 6 Configure a git repository in Atomic Configure Git Repository Configure Git Repository Please select an existing repository or create a new one D Use or create repository in parent folder of project Repository Working directory No repository selected Path within repository Project Name Current Location Target Location pcc2 home stephan atomic workspace Cancel 13
9. ent salt Die Jugendlichen in Zossen wollen ein Musikcaf Atomic User Guide v0 1 7 Tool based annotation The Annotation Graph Editor includes a tools palette which can be used for tool based annotation The following paragraphs describe how to use the tools in the palette to create an annotation graph for the corpus document Keyboard navigation The Annotation Graph Editor can be navigated via the keyboard using the following keys Arrows Left Right Up Down Change the currently selected element Shift Arrows Extend the currently selected element to include the elements to the left the right above or below Ctrl Arrows Move focus to a different element without changing the selection Ctrl Shift Arrows Scroll the GEF viewer left right up or down Alt Down Arrow Change the currently selected element to be the first nested element within the element that currently has focus Alt Up Arrow Change the currently selected element to be the parent element containing the element that currently has focus Change the selection to be a connection associated with the element with focus Repeatedly pressing either of these keys cycles the current selection forward or backward through the connections associated with the element with focus e Extend the selection to include connections associated with the element with focus Creating nodes Click on the Structure Tool or the Span Tool in order to
10. gin you would want to install is Eclipse EGit To install this plugin do the following e Open the Install Wizard see above e Select the p2 repository to work with in this case Software Site for additional Eclipse plugins at http download eclipse org releases indigo http download eclipse org releases indigo This is the default site and is already defined in Atomic e Search for egit or browse the category Collaboration until you find the entry Eclipse EGit e Check the checkbox next to the plugin name e Click Next 11 Atomic User Guide v0 1 7 e The wizard will calculate requirements and dependencies Once its finished review the items that you want to install and click Next e Read the license and accept it e Atomic will now install EGit Once its finished you will be prompted to restart Atomic Click Restart Now e Once Atomic is restarted EGit is installed and accessible via right click on a project in the Navigation View cf Figure 5 Sharing an Atomic project to a git repository You can now use EGit to version your corpus documents or any other files you have in your workspace For more information on how to use EGit please refer to the EGit User Guide http wiki eclipse org EGit User_Guide Figure 5 Sharing an Atomic project to a git repository ri ow ka Ctrl C Delete Rename Import Export Refresh F5 Close Project Apply Patch Compare With Share Proj
11. have suggestions feature requests problems or want to file a bug report please create a new issue at Atomic s issue tracker at github com https github com infraling atomic issues When reporting bugs please don t forget to mention the version of Atomic you are using and your operating system and architecture e g Windows 7 64 bit Atomic is open source under the Apache License Version 2 0 http www apache org licens es LICENSE 2 0 html and you can find the source code at github com infraling atomic https github com infraling atomic If you want to learn about Salt Atomic s data model or Pepper the converter framework for linguis tic formats included in Atomic please go to the SaltNPepper homepage https korpling german hu berlin de saltnpepper The workspace In Atomic workspaces can be defined A workspace is a physical location 1 e path on the file system where Atomic stores projects folders and files You are asked to pick a workspace location when running Atomic User Guide v0 1 7 Atomic for the first time and each subsequent time you run the application if you haven t activated the option to remember the workspace Creating a project There are two ways to create a project in Atomic e Create a new project from scratch providing the corpus text e Import an existing corpus from one of the formats which can be read by Pepper Create a new project To create a new Atomic project ope
12. ished review the items that you want to install and click Next e Read the license and accept it e Atomic will now install XMLEspresso Once its finished you wil be prompted to restart Atomic Click Restart Now e Once Atomic is restarted XMLEspresso is installed and accessible via right click on an XML or salt file in the Navigation View and the Open With Other menu item Install new plugins from a p2 repository Atomic can install plugins from p2 repositories e g when they are not available via the Eclipse Market place All of the plugins provided by Eclipse are available from the official Eclipse Software Site Third party plugins may have their own update sites usually specified on the plugins homepage Atomic in cludes an Install Wizard to search the available plugins and install them as needed You can access it via Options Install new plugins from p2 repository In the Wizard you have to provide the URL of a p2 repository Atomic already includes the reference to the official Eclipse repository Select the Software Site for additional Eclipse plugins from the Work with field and wait until the contents are fully loaded The plugins are shown in categories but you can switch categorization off in the self explanatory Install Wizard as well For further information on how to install plugins from a p2 repository cf Example Install Git support in Atomic 10 Atomic User Guide v0 1 7 Figure 4
13. n the New Project Wizard either by going to File gt New Project or by pressing CTRL N In the wizard choose New Atomic Project Figure 1 The New Project Wizard at Step 2 New Atomic project Create a new Atomic project This wizard helps you create an Atomic project from a raw corpus text Enter the name For the new project and choose the text file containing the raw corpus text Alternatively you can enter or paste the corpus text into the text field NOTE If you want to import a pre formatted existing corpus into Atomic you can do so by starting the Corpus Import Wizard Right click into the navigation window and choose Import gt Corpus Import Choose corpus text file Browse Enter corpus text Raw corpus text Click Browse to pick text file containing the corpus text or to enter corpus text Choose tokenizer TreeTagger Tokenizer lt Back Next gt Cancel In the next step cf Figure 1 The New Project Wizard at Step 2 fill in the project name and e choose a txt file containing the corpus text and select the file s encoding or Atomic User Guide v0 1 7 e enter the corpus text via the text field provided after clicking the button Finally choose the tokenizer that should be used to tokenize the corpus text currently only the TreeTagger Tokenizer is available and press Finish The corpus text will now be tokenized and the project created Impor
14. ns from a p2 repository Install new plugins via the Eclipse Marketplace The Eclipse Marketplace http marketplace eclipse org is a platform for third party Eclipse based so lutions i e plugins or complete applications Plugins available from the marketplace can be consumed by any Eclipse based application such as Atomic Searching the marketplace and installing software from it is achieved via the Eclipse Marketplace client included in Atomic Atomic User Guide v0 1 7 Figure 3 The Eclipse Marketplace client Eclipse Marketplace Eclipse Marketplace Select solutions to install Press Finish to proceed with installation Press the information button to see a detailed overview and a link to more information Search Recent Popular installed Find Q latex All Markets AllCategories gt Go COOL by Budapest University of Technology and Economics EPL Install Ecore html latex oe StatET forR Share Eclipse based IDE for the statistical programing language R It offers a set of mature tools For R coding and package building This includes a Fully integrated R by WalWare de EPL Learn more editor console debugger r latex e TeXlipse Share The TeXlipse plugin provides both LaTeX and BibTeX editors a project creation wizard and a complete user manual of the editor functions Additional features by Texlipse sf net community EPL Install tex latex Pdf4Eclipse Share Pdf4E
15. rns namespace key value key value Key and value are shown directly on the element the namespace is shown in a pop up when hovering over an annotation Elements i e nodes as well as edges can be annotated by selecting the element via the mouse or the keyboard s arrow buttons and then either clicking on it once or pressing Return This action will open the annotation field for direct editing You can enter one or more annotations with one annotation per line To add a line break press Return To commit the annotations press CTRL Return To abort editing the annotations press ESC To edit a single existing annotation on an element select the annotation by mouse or keyboard and either click on it or press Return This will open the single annotation field Change the annotation and press Return to commit it Press ESC to abort the editing process Deleting elements and annotations In order to delete elements and annotations select the respective element s or annotation s per mouse or keyboard and either press DEL or click the delete symbol in the menu bar Command line based annotation with AtomicAL Atomic provides the native annotation language Atomic Annotation Language AtomicAL for rapid com mand line annotation The AtomicAL console is available when an editor with AtomicAL support is open The AtomicAL con sole can be opened via the menu View gt AtomicAL Console The AtomicAL console
16. t an existing corpus In order to import an existing corpus right click into the Navigation window and choose Import from the pop up menu In the Import Wizard choose Atomic Corpus Import Follow the instructions provided by the Import Wizard to import a corpus from one of the supported formats These are currently CoNLL EXMARaLDA Elan MMAX2 PAULA RelANNIS SaltXML Treetagger UAM The corpus will be imported and wrapped in a new project Annotation with the Annotation Graph Editor To open acorpus document of your choice in the Annotation Graph Editor double click the file containing the respective corpus document in the Navigation view Should you have opened this file with another editor before Atomic will have remembered the editor the file was opened with last In this case right click onto the file and choose Open With Annotation Graph Editor from the pop up menu When opening a newly created document the editing area will show the tokenized corpus text Figure 2 Atomic application window with Annotation Graph Editor Atomic Palette b Y amp jugendliche Is Y corpus amp Nodes B corpus_document s CI Structure B saltProject salt Clspan Edges Dominance relation Spanning relation Pointing relation Order relation T1 T2 T3 T4 T5 T Die Jugendlichen in Zossen wollen ein console Ex BE rt tiv O AtomicAL Console To display a list of available commands type help Working on corpus docum
Download Pdf Manuals
Related Search