This page is translated from installer_txm_dans_une_salle_de_cours_windows. It describes in detail TXM installation operations on Windows for easy batch installation in a computer classroom. The faq also contains other useful information in case of difficulties

Let $PROGRAMFILES be the installation path TXM:

• XP: C:\Program Files
• Vista and Seven: C:\Program

Let $HOMEDIR be the path of the user directory:

• XP: C:\Documents and Settings\UserID
• Vista and Seven: C:\Users\UserID

Installing TXM requires two main steps:

1. run the setup TXM (created with NSIS)
2. run TXM software once from a user account

The first step consists in placing all the necessary files into $PROGRAMFILES and the second copies some of the files into $HOMEDIR and performs the basic settings of TXM.

http://textometrie.ens-lyon.fr/files/software/TXM/0.7.9 (check the prerequisites on the page).

Run the Setup in a command line:

• with argument /S
• Warning: if TXM has already been installed, a popup is displayed to confirm the uninstallation. For automation, you need to:
  • remove the registry key: Software\Microsoft\Windows\CurrentVersion\Uninstall\TXM UninstallString
  • delete the folder $PROGRAMFILES/TXM

• A) If you wish to work with roaming accounts directories, you may tick the box “Do not use local user directory” This option may be useful in case the local user directories are systematically re-initialized and the students need to keep their corpora over several classes. This option requires that roaming accounts connection profiles have sufficient sufficient storage capacity (at least 100 Mb for demonstration corpora, additional 400 Mb for BROWN corpus).
• B) If you wish to work with directory located on a removable drive (e.g. USB drive), you may tick the box “Do not use user directory as TXM working directory”. This configuration may be useful in case local user directories are systematically re-initialized and the students need to keep their corpora over several classes but the storage capacity of roaming accounts is insufficient. In this case the path to the working directory will be asked every time the user starts TXM.
• Automation:
  • place an empty file names “ROAMINGINSTALL” in $PROGRAMFILES/TXM to activate the “Do not use local user directory” option
  • place an empty file names “ASKUSERDIRECTORY” in $PROGRAMFILES/TXM to activate the “Do not use user directory as working directory” option
• C) Additional configuration: After installation, if the options described above are not sufficient, it is possible to edit the script that launches TXM ($PROGRAMFILES\TXM\TXM.bat) in order to specify the path to the working directory.

If the firewall is not configured before installation, permissions are required interactively by the firewall during the first run of TXM.

• $PROGRAMFILES/R/bin/r.exe
• $PROGRAMFILES/TXM/cwb/bin/cqpserver.exe (since TXM 0.7, network access for this module is no longer required)

On some PCs, the application of anti-virus on network communications related to ports used by TXM can impact the TXM computing performance dramatically. (This is less the case since TXM 0.7 where the search engine does not use network by default)

See page http://textometrie.ens-lyon.fr/html/doc/treetagger/en/install.html.

• Copy additional binary corpus files to $PROGRAMFILES/samples so that they are loaded at TXM first session.

• Manually in a TXM menu: “File > Change language” or “Fichier > Changer la langue.” Attention in TXM 0.7 for 64bit machines, you must quit and restart TXM for the language change to take effect.
• Automatically in the file $PROGRAMFILES/TXM/TXM/TXM.ini
  • replace “-nl\nen” with “-nl\nfr” for French
  • replace “-nl\nen” with “-nl\nru” for Russian

Copy the file $HOMEDIR/Start Menu/Programs/TXM in $ALLUSERHOMEDIR/Start Menu/Programs/TXM where $ALLUSERHOMEDIR is the user directory for “All Users”.

• Assume $HOMEDIR/TXM does not exist
• Run TXM from the Start menu or $PROGRAMFILES/TXM/TXM.bat

At the first run of TXM, it is useful to make a first elementary calculation on a built-in corpus to check the general functioning of TXM.

• Manually: Start calculating the lexicon of the VOEUX corpus (context menu by right clicking on the icon of VOEUX corpus, entry 'Lexicon') if the calculation is successful, the installation is complete (the platform and access to the engine research work. To verify access to the R engine, you can run the command 'Information' on a partition that will display a histogram calculated R)
• Automatically: it is possible to run this calculation through a Groovy script: see manual TXM (0.7 in TXM it is not yet possible to pass the script name as a parameter of a TXM call)

1. Copy the following text (which you may copy from the TXM console):

 Start-up...
 Search Engine launched in memory mode.
 lancé.connecté statistical engine.
 Loading subcorpus and partitions ... Done.
 Ready.

2. Use TXM menu File> Import> Clipboard 3. Check in the last visible console are:

 pAttrs: [id, lbid, enpos, enlemma]
 sAttrs: [text: + name + id + type + path + base + project, s: + n, p: + id, txmcorpus: + lang]
 - EDITION - Building edition
 .
 Import Complete: 3 seconds (3966 ms)
 Search Engine launched in memory mode.
 lancé.connecté statistical engine.
 Loading subcorpus and partitions ... Done.
 TXM is ready.

(Note that the first line above must contain enpos and enlemma. But the indication of time after “Import Completed” can of course be different.)

Solution 1:

The above procedure can be transposed: TreeTagger needs a language model for the language of the environment, the copied text must be in that language, and the import console after the “fr” of frlemma and frpos are replaced by the two-letter language code, for example an English environment: <pre> pAttrs: [id, Ibid, frpos, frlemma] </pre> becomes

 pAttrs: [id, Ibid, enpos, enlemma]

Solution 2:

You choose and set the language you are going to test (called the 'XX'), unifying the following three settings:

1. Language model for TreeTagger: download and install a TreeTagger language model for the language you want to take the test, if not already done; the file name is the language code on two characters, as explained in the instructions for installing TreeTagger (eg 'en.par', 'fr.par.' - in our case 'XX.par').
2. Language selected for import clipboard: enter the code in two characters, eg. 'En', 'en', 'it' … - 'XX' in our case) in the following preference field: Tools> Parameters> TXM> User> Import> Clipboard, Default language.
3. Test corpus: choose a (piece of) the text in the same language (i.e. XX), you copy and then import via File> Import> Clipboard.

You should then find the following line in the console after import:

 pAttrs: [id, lbid, XXpos, XXlemma]

where XX is the two-character code of the selected language.

• Close TXM
• Run TXM for the second time
• A window informing on available TXM update may appear at the bottom right. In this case click in the window to start TXM update to the latest version
• If no update window appears, the update can be started manually by using the menu File> Check for updates
• To update to a BETA version, change Tools > Paremeters > TXM > Advanced > Update level to “BETA” and check for updates again

All remarks to improve this page are welcome.