Acrolinx IQ™

Acrolinx IQ Multilingual Term
Extraction
User Guide
Acrolinx IQ™
2
Contents
Installation - Major Steps
3
Installing Cygwin..................................................................................................................................4
Acquiring Third-party Tools....................................................................................................................5
Preparing Additional Tokenization Rules..................................................................................................6
Confirming and Editing Term Harvesting Settings....................................................................................7
Starting the Multilingual Term Extraction Tool for the First Time...............................................................8
Using the MTE Tool
11
Starting the Multilingual Term Extraction Tool.......................................................................................11
Performance Optimization........................................................................................................11
Context-specific Help..........................................................................................................................12
Running Multilingual Term Extraction....................................................................................................12
Opening Existing Results.....................................................................................................................12
Preparing the Results for Validation.....................................................................................................13
Validating and Importing the Results
15
Validating the Results.........................................................................................................................15
About the Term Extraction Results File......................................................................................16
Importing Validated Files.....................................................................................................................17
FAQ..................................................................................................................................................19
Troubleshooting
21
3
Chapter 1
Installation - Major Steps
You install the MTE (Multilingual Term Extraction) tool by extracting the
contents of the installation package to your computer. To complete the MTE
tool installation, you must be connected to the internet. Before you can use
the MTE tool, you must install several required components that automatically
download software files from the Internet during installation.
â
To install the MTE tool, follow these steps:
1 Unzip the MTE tool into an otherwise empty local folder.
Example: c:\MTE_tool
This guide refers to the installation directory as <INSTALL_DIR>.
2 Install Java, if it is not already installed, or upgrade Java if a version prior
to 1.6 is installed. You can determine what version of Java is installed by
entering java -version at a command prompt.
The Java installer can be downloaded from
http://www.java.com/en/download/manual.jsp
3 Install Cygwin, including certain necessary packages. (see "Installing
Cygwin" on page 4)
If you already have Cygwin installed, you can follow the Cygwin installation
procedure to add any necessary packages are not included in your current
installation.
4 Acquire necessary third-party tools by running an update script in your
installation directory (see "Acquiring Third-party Tools" on page 5).
The update script downloads components such as GIZA++, and certain
parts of Moses.
5 (Optional) Prepare any additional tokenization files (see "Preparing
Additional Tokenization Rules" on page 6).
If you do not plan to extract Japanese or Chinese terms and want to get
started quickly, you can skip this step. You can use the Moses tokenizer
which does not require any preparation.
6 (Optional) Confirm your Term Harvesting settings (see "Confirming and
Editing Term Harvesting Settings" on page 7).
The MTE tool uses the standard Acrolinx IQ Term Harvesting functionality
to extract monolingual terms.
•
•
If you plan to use Term Harvesting, you must have one or more rule
sets with Term Harvesting configured before you start multilingual
term extraction.
If you have already have list of source terms that you plan to use for
extracting translations, you can skip this step.
7 Start the MTE tool and configure the advanced settings (see "Starting the
Multilingual Term Extraction Tool for the First Time" on page 8).
Before you can start multilingual term extraction, you must configure
some compulsory advanced settings.
4
Installation - Major Steps
Installing Cygwin
Cygwin is a Unix-like environment and command-line interface for Microsoft
Windows. The MTE (Multilingual Term Extraction) tool uses some components
that normally only run on a Unix-based operating system. You install Cygwin
so that the MTE tool can run on Microsoft Windows.
â
To install Cygwin including the packages required by the MTE tool, follow
these steps:
1 Download the Cygwin setup tool from the Cygwin website.
2 Run the setup tool, and make a note of the directory where you install
Cygwin.
You will need to provide the Cygwin installation path in another step.
3 In the Select Packages screen, make sure the following packages are included:
•
Base:
•
•
gzip
Devel
•
•
•
gcc-g++
make
subversion
The subversion package not necessary if you already have a
DOS-based Subversion client installed.
•
•
Perl
•
•
zlib-devel
perl
Web
•
wget
Click the word Skip to mark a package for inclusion.
Figure 1: The
"gcc-g++" Package
When Selected for
Inclusion
4 Allow the setup tool to install any additional packages.
5
Acquiring Third-party Tools
The MTE (Multilingual Term Extraction) tool requires certain components such
as the "Moses" open-source statistical machine translation system. These
components can only be obtained by connecting to an online update site after
you install the main software files. You obtain these components by running
an update script in Cygwin.
â
To acquire the required third-party tools, follow these steps:
1 Create a directory on your computer to store the third-party tools.
The path name must not include any whitespaces.
Example: c:\MTE_third-party_tools
2 Open the Cygwin command-line window and change to the directory that
you created in the previous step by using the following syntax:
cd
/cygdrive/<DRIVE_LETTER>/<DIRECTORY_NAME>
Example command:
cd /cygdrive/c/MTE_third-party_tools
3 Run the update script by using the following syntax:
/cygdrive/<DRIVE_LETTER>/<INSTALL_DIR>/scripts/install-moses4-multilingualTE.sh
Example command:
/cygdrive/c/MTE_tool/scripts/install-moses-4-multilingualTE.
sh
You can also redirect the output of the script to a log file so that the update
events can be inspected for possible errors. To output the update events
to a log file instead of the command-line window, run the update script
using the following syntax:
<PATH_TO_UPDATESCRIPT> >& <PATH_TO_LOG_FILE>
Example command:
/cygdrive/c/MTE_tool/scripts/install-moses-4-multilingualTE.
sh >& /cygdrive/c/MTE_tool/install-moses.log
If you output to a log file, the Cygwin command-line window might appear
unresponsive until the script has completed. The script usually takes a
few minutes to complete.
Once the update script has completed, the following paths should exist
relative to the directory that you created in the first step:
moses4multilingualTE\bin\
moses4multilingualTE\moses-scripts\
Later, you will need to provide the path to these directories in the MTE
tool (Advanced Settings ➤ Translation extraction ➤ Moses directory ).
4 If you plan to do multilingual term extraction from documents that contain
Japanese or Chinese text, ensure Windows installation has support for
East Asian languages.
•
In Windows XP, ensure that the following option is selected:
6
Installation - Major Steps
Control Panel ➤ Regional and Language Options ➤ Languages ➤ Install Files for East
Asian languages .
If you find that the option is deselected, you must select it, which will
result in you being prompted to insert a Windows installation CD.
If you do not have the exact CD for which you are prompted, but a
similar one (e.g. you have SP2 instead of SP3), you might still be able
to use it. Click on the Browse button to indicate that you wish to use
the relevant files on the disk you have.
•
If you use Windows 7 or Vista, East Asian language support is installed
as a standard feature.
Preparing Additional Tokenization Rules
Tokenization is the process of breaking a stream of text up into words,
phrases, symbols, or other meaningful elements called tokens. The MTE
(Multilingual Term Extraction) tool can tokenize Roman-script languages by
using a specialist Acrolinx tokenizer or the generic Moses tokenizer. The MTE
tool can also process Japanese and Chinese text, but only with an Acrolinx
tokenizer. Depending on the tokenization methods you decide to use, you
might obtain extra tokenization rule files before you start using the MTE tool.
•
•
You might use the Moses tokenizer if you want to get started quickly and
do not require term extraction for Japanese or Chinese text. If you use
Moses tokenization for all languages, no extra files are required and you
can start the MTE tool.
You might use the Acrolinx tokenizer if you want more accurate results
for a specific language or if you require term extraction for Japanese or
Chinese text.
•
If you want to use the Acrolinx tokenizer for Japanese or Chinese text
only, you do not require extra files because tokenization rules for
Japanese and Chinese are included with the MTE tool package at the
following locations:
Japanese: <INSTALL_DIR>\tok\ja\tokenRules.tok
Chinese: <INSTALL_DIR>\tok\zh\tokenRules.tok
Note these paths and continue the installation.
•
If you want to use Acrolinx tokenization for a language other than
Japanese or Chinese you must use additional tokenization rules from
your Acrolinx IQ Server installation. The Acrolinx linguistic resources
include tokenization rules for each language that you have purchased.
â
To prepare additional tokenization rules, follow this step:
•
Ensure that the MTE tool can access the Acrolinx tokenization rule files:
•
If the Acrolinx IQ Server is installed in the same computer as the MTE
tool, you use the tokenization rule file in your Acrolinx IQ Server
installation directory:
<ACROLINXIQ_INSTALL_DIR>\data\<LANG-ID>\resources\tokenRules.
tok
7
Note this path and continue the installation.
•
If the Acrolinx IQ Server is installed on a different computer to the MTE
tool, copy the entire resources directory to the computer where the
MTE tool is installed:
You copy the entire resources directory because the tokenization rules
reference other files in the same directory.
Confirming and Editing Term Harvesting Settings
When you run multilingual term extraction, the MTE (Multilingual Term
Extraction) tool connects to your Acrolinx IQ Server and runs a check that
includes monolingual term harvesting in at least one language. Before you
use the MTE tool, you must ensure that you have at least one rule set that
is configured for Term Harvesting in the source language. You can also
configure monolingual Term Harvesting for any additional languages that
your linguistic resources support.
The MTE tool runs Term Harvesting to produce the list of source terms that
is subsequently used to extract terms in the target languages. You might
also configure Term Harvesting for a target language to improve the quality
of the translations that are selected as primary candidates for each source
term.
For the source language and optionally for any other target languages, decide
which rule set you want to use for Term Harvesting.
Take into consideration that the Acrolinx IQ Server does not harvest terms
that already exist in your terminology database and are loaded for checking.
The rule set that you select will determine which terms are loaded for
checking. Consequently, the multilingual term extraction results can vary
depending on the rule sets that you select.
If you want the MTE tool to extract all possible terms regardless of the terms
in your terminology database, use a rule set that is not associated with any
term sets
â
To confirm and edit the term harvesting settings for your rule sets, follow
these steps:
1 Open the language configuration for each relevant language.
<ACROLINXIQ_INSTALL_DIR>\data\<LANG_ID>\configuration.
properties
2 Ensure that the following setting is present, adding or editing it if
necessary:
<RULE-SET>.termharvesting.onlyServerSide=true
where <RULE-SET> is the name of the rule set that you plan to use for
this language.
3 Ensure that the rule set has term harvesting rules configured by locating
the following setting:
<RULE-SET>.harvestingRules=<HARVESTING-RULES-FILE>
8
Installation - Major Steps
Typically, the value for <HARVESTING-RULES-FILE> is rules/<LANG_ID>
-harvesting.thrul .
Also ensure that the *.thrul file exists at the defined location which is
relative to the directory <ACROLINXIQ_INSTALL_DIR>\data\<LANG_ID>.
4 Confirm the term sets that your rule set is configured to load by searching
for the setting:
<RULE_SET>.terminology.sources
If this property does not exist, the rule set loads the term sets that are
configured in the fallback property
terminology.sources
5 If you made any changes to your language configuration, restart the
language servers.
You might need to repeat these steps if you uninstall and reinstall your
Acrolinx IQ server and/or resources.
Starting the Multilingual Term Extraction Tool for the
First Time
When you start the MTE (Multilingual Term Extraction) tool for the first time,
you must configure some advanced settings before you can run multilingual
term extraction.
â
To start the MTE tool and configure the advanced settings for the first
time, follow these steps:
1 Start the MTE tool by running the batch script at the following location:
<INSTALL_DIR>\run-mte-tool.bat
2 In the window that appears, click Extract Terms From Parallel Corpora.
3 In the Settings window, click Advanced Settings.
9
4 Configure the monolingual term extraction (Term Harvesting) settings:
a On the Monolingual TE tab, enter the address of the Acrolinx IQ Server
and a user name.
You must enter a user name for a user who does not require a password
to connect to the Acrolinx IQ Server.
b Click New to add a language and specify which rule set to use for Term
Harvesting (see "Confirming and Editing Term Harvesting Settings" on
page 7).
At a minimum, you must add the source language.
5 On the Tokenization tab select the tokenization behavior (see "Preparing
Additional Tokenization Rules" on page 6).
10
Installation - Major Steps
6 On the Translation extraction tab, specify the directories where you installed
Cygwin (see "Installing Cygwin" on page 4) and Moses (see "Acquiring
Third-party Tools" on page 5).
7 Click OK to save your settings.
The MTE tool is now ready to use.
11
Chapter 2
Using the MTE Tool
Starting the Multilingual Term Extraction Tool
Start the MTE (Multilingual Term Extraction) tool to open existing extraction
results or to run multilingual term extraction.
â
To start the MTE tool, follow these steps
1 Run the MTE tool batch script at the following location:
<INSTALL_DIR>\run-mte-tool.bat
2 In the window that appears, click one of the buttons to indicate how you
want to use the MTE tool.
•
•
Click Extract Terms From Parallel Corpora, if you are opening the MTE tool for
the first time, or if you want to run term extraction.
Click Open Existing Results, if you want to review the results from a previous
term extraction.
3 If you are starting the MTE tool for the first time, configure the advanced
settings before you run term extraction (see "Starting the Multilingual
Term Extraction Tool for the First Time" on page 8).
Performance Optimization
If you want to extract terms from a relatively large file (perhaps 250,000+
segments ) , or to open the results of such an extraction, then you may need
to increase java's heap size.
The run-mte-tool.bat script allows for this via optional arguments.
For example, if you invoke the script in the following way from a command
prompt, then the java virtual machine's initial heap size will be 512 megabytes
and its maximum heap size will be 1 gigabyte:
run-mte-tool.bat java 512m 1g
(The first argument -- 'java' -- is the java command to use. Modify this
argument only if you have multiple versions of java installed and want to
specify a non-default one to use.)
If you try to extract terms from a very large file (perhaps 900,000+
segments), you may find that the process crashes with a segmentation fault
or a 'Hangup' during a call to the program snt2cooc.out. The log contains
something like this:
line 910000 line 911000 line 912000 sh: line 1: 172 Hangup ...
In that case, increase Cygwin's maximum memory per the instructions at
the following URL, then restart the extraction:
http://www.cygwin.com/cygwin-ug-net/setup-maxmem.html
12
Using the MTE Tool
Context-specific Help
If you are unsure about the purpose of a specific setting or the type of input
required, you can use the context-specific help to get more help for that
setting.
Most settings in the MTE tool have context-specific help that displays when
you move your mouse over the label for the setting.
Running Multilingual Term Extraction
You run multilingual term extraction to generate a list of translation candidates
for each source term in all of the specified target languages.
1 Start the MTE tool and click Extract Terms From Parallel Corpora in the window
that appears.
2 Enter a source language and one or more target languages.
3 Decide how the source terms will be provided.
•
•
If you want to use Acrolinx IQ Term Harvesting to extract terms and
generate a list of term candidates in the source language, select Extract
candidates from source side of parallel corpora.
If you already have a list of source terms that you want to find
translations for, select Use terms listed in file and browse the file that
contains your source terms.
4 Provide one or more input files.
•
•
Click Add TMX to add a TMX that contains segments in the source and
target languages.
Click Add paired texts... to add text files that contain identical lists of
sentences in the source language and one of the target languages.
5 Enter an output directory and click Start Term Extraction.
Opening Existing Results
You might open existing results to prepare the results for validation or to
quickly review the quality of the term extraction for a specific document.
1 Start the MTE tool and click Open Existing Results in the window that appears.
2 In the file browser that appears, browse to the results file.
The results of a given multilingual term extraction run can be found at:
13
<OUTPUT_DIR>\*.term-translations
TIP: You can also open multiple results files simultaneously by holding
down the CTRL key and selecting the files.
3 If you configured Term Harvesting (monolingual term extraction) for one
or more target languages as well as the source language, add the Term
Harvesting candidates to the results.
When you add the Term Harvesting results for a target language, you
improve the quality of the translations that are selected as primary
candidates for each source term.
a) In the "Extracted Multilingual Terms" window, click Filtering Settings ....
b) In the dialog box that appears, select the table row for the relevant
language and then click Add.
c) In the file browser that appears, browse to the file that contains the
term candidates for the relevant language.
If Term Harvesting was run successfully for the relevant language, you
can find the file containing the term candidates at the following location:
<OUTPUT_DIR>\monolingualTE_<LANG_ID>\terms
After you have specified the file, the file path should appear in the
table.
d) After you have selected a terms file for each relevant language, click
OK.
4 If you are satisfied with the preliminary results, you can prepare the
results for validation.
Preparing the Results for Validation
If you are satisfied with the preview of the results that is displayed in the
"Extracted Multilingual Terms" window, you can prepare the results for
validation. When you prepare the results for validation, you generate a file
that users can view and edit in Microsoft Excel and eventually import into
the Acrolinx IQ Terminology Manager.
â
To prepare the results for validation, follow these steps
1 Open some existing results and select File ➤ Prepare results for validation...:
2 In the dialog box that appears, configure the settings for the output file,
including the file format and output directory.
You can use the context-sensitive help (see "Context-specific Help" on
page 12) to get more information on individual settings.
3 Click OK to generate the validation files and distribute for validation (see
"Validating the Results" on page 15).
14
Using the MTE Tool
TIP: Validation files are also generated automatically after multilingual
term extraction has completed. The files are stored in a subdirectory with
the name validation_<FORMAT>. For example, if you selected the
preferred format "Excel 97-2003 Workbook (*.xls)", a directory with the
name validation_xls is created. You can configure how these files are
generated on the Validation tab of the advanced settings. Note that formats
other than Excel 97-2003 Workbook (*.xls) are not recommended.
15
Chapter 3
Validating and Importing the Results
Validating the Results
The extracted term candidates must be manually validated before you can
import them into the terminology database.
â
To validate the results of the multilingual term extraction, follow these
steps:
1 Open the XLS file with the results in Microsoft Excel.
2 Validate the results by editing the content of the individual rows:
a) (Optional) Check the Cluster ID and edit the Subcluster ID to group
term candidates and their proposed translations together and to identify
term variants.
Example: The term candidates back and back panel were identified as
possible synonyms and assigned to the Cluster ID 1.
Additionally both candidates were numbered consecutively with the
Subcluster ID 1 and 2.
If you want to confirm that both term candidates are in fact synonyms
you assign them an identical Subcluster ID (for example 1).
b) Assign the desired status to the term candidate and the translations.
c) Make sure that the connections between the term candidate and the
translation are correct by checking the Contexts row.
d) (Optional) Add any missing term variants and translations in a new
line.
IMPORTANT: Add additional term variants in the source language in
a line above the automatically extracted term candidate.
16
Validating and Importing the Results
Additional translations must not be added to the same line as the term
candidates.
Figure 2: Adding
translations
Figure 3: Adding
variants
3 Save the file.
NOTE: Save the file as .xls if you received an XML file.
About the Term Extraction Results File
The term extraction results are saved in one or more Excel files. By default,
these files contain the following rows:
Row
Description
Editable
Cluster ID
Possible synonyms are assigned an No
identical Cluster ID.
Subcluster ID
You can use the Subcluster ID to
Yes
select if possible synonyms (with an
identical Cluster ID) are in fact
synonyms.
You mark synonyms by assigning
an identical Subcluster ID.
Status
The status of the term candidate in Yes
the source language.
(mandatory)
The default status values in the
dropdown menu are preferred,
admitted, proposed, deprecated und
rejected (you can also type only the
beginning of the status name, for
example r, re, or rej).
17
Row
Description
Editable
You can also assign your own status
values that you have to map
correctly during import.
If no status value is selected the
default value is proposed.
Use the status rejected to mark
incorrect term candidates and
ignore them during import. The
corresponding translations are
automatically ignored and do not
need a status value.
Term Candidate
The term candidate that was
identified during term extraction.
No
Frequency
The frequency of the term candidate No
in the translation memory.
Status
The status of the proposed
translation.
Yes
(mandatory)
Use the status rejected to mark
incorrect translations that will be
ignored during import.
Language
The language of the proposed
No
translation. This row is displayed if
the term extraction was run on
more than two languages.
Proposed Translation The proposed translation for the
term candidate.
No
Frequency
The frequency of the proposed
No
translation in the translation
memory as translation for the term
candidate.
Contexts
The contexts of the possible
No
translations that can be used for
validation. The term candidates are
highlighted.
TIP: A maximum of five contexts is
displayed.
Importing Validated Files
After you validated the results of the multilingual term extraction and saved
the Excel files you can import the files with the term import wizard to use
the new terms and their translations in the terminology database. You must
convert the files to the compatible OLIF format before you can import them.
18
Validating and Importing the Results
The conversion can be done with the Java based tool ValidatedTermsTo
Olif.jar which is available from your Acrolinx project consultant.
â
To import the validated Excel files, follow these steps:
1 Convert the validated files into the OLIF format:
a) Open a command prompt and change to the directory where you saved
the file ValidatedTermsToOlif.jar.
Example:
cd c:\MTE_tool\ExcelOlifConverter\
b) Enter the following command:
java -jar ValidatedTermsToOlif.jar <INPUT_DIR> <OUTPUT_FIL
E>
IMPORTANT: The input directory must contain only the validated files.
2 Open the Dashboard and navigate to
Import.
Terminology > Import and Export >
TIP: To ensure that any validated terms and translations are imported
correctly, you should set the dulicate detection (Terminology > Customize
> Duplicate Detection) to UUID (Universally Unique Identifier) is the same
as an existing term (allows most duplicate terms). This way terms are also
imported if for example two source terms have an identical translation.
OLIF and select the file.
Field Mapping page and set the following options:
a) Map the element status to the field Status.
b) Do not map the element adminStatus.
3 In the Import Options set the File Format to
4 Click Next to open the
c) If you want to import individual contexts, map each of the numbered
context elements up to the maximum number of contexts that you
want to import per term to a corresponding field.
For example, if you want to import the first two contexts for each term,
map the elements context1 and context2.
19
If you are importing validated results for the first time, create the
required context fields by using the Add Field button.
Figure 4: Field Mapping
5 Click Next to open the List Item Mapping page and set the following options:
a) Click Map It for Me to automatically map any known status values.
b) Manually map any other status values.
6 Click Finish. In the dialog box Source Language click OK to start the import.
You can display the validated terms and their translations in the
Terminology Manager.
FAQ
This section contains the answers to several frequently asked questions.
•
How can I correct spelling errors?
The best way to correct spelling errors is to add the correct form in a new
line directly under the incorrect form.
•
Where can I add additional columns to the Excel file?
You can add additional columns to any part of the table. However, you
must not edit existing titles or add duplicate titles. Manually added columns
will be imported but must be mapped manually in the field mapping.
•
What do the different colors in the Excel file mean?
The colors have no special meaning. They are used to highlight editable
fields and to make it easier to distinguish the individual clusters.
•
On what basis are the results sorted?
The multilingual term candidates are grouped into clusters which contain
the possible synonyms in the source language. The clusters are then
sorted in descending order according to the frequency of the term
candidates in the source language. The results are split into several files.
Files with a lower number contain the most frequently used clusters.
20
Validating and Importing the Results
•
Can I delete terms instead of assigning them the status rejected?
Yes. You can also manually delete terms from the table.
•
The results are split between several files and the Cluster IDs are repeated
in each file. Are term candidates with these identical IDs linked during
import?
No. Term candidates from different files are never linked.
•
Is it necessary to convert every Excel file individually into the OLIF format?
No. During the conversion all files in the input directory are converted
into one single OLIF file.
21
Chapter 4
Troubleshooting
You might experience the following problems when using the MTE tool.
Problem description
Error messages like this one keep showing up in the log of your extraction
run:
1 [main] perl 2052 C:\cygwin\bin\perl.exe: *** fatal error unable to remap \\?\C:\cygwin\lib\perl5\5.10\i686-cygwin\auto\
Storable\Storable.dll to same address as parent: 0xB20000 != 0x
BD0000
Stack trace:
Frame
Function Args
0088B508 6102796B (0088B508, 00000000, 00000000, 00000000)
0088B7F8 6102796B (6117EC60, 00008000, 00000000, 61180977)
0088C828 61004F1B (611A7FAC, 612489D4, 00B20000, 00BD0000)
End of stack trace
0 [main] perl 6980 fork: child 2052 - died waiting for dll
loading, errno 11
Explanation
Cygwin's memory layout was somehow disrupted. This can happen from time
to time.
Solution
Follow these steps to correct Cygwin's memory layout:
1 Make sure the 'rebase' package is included in your Cygwin installation.
You can check whether it is included by typing 'which rebase' at a Cygwin
command prompt.
If it is included, you will see a message like '/usr/bin/rebase'.
If it in not included, you will see a message like 'which: no rebase in ...
'.
If it is not included, add it using the Cygwin setup.exe utility.
2 Close all Cygwin processes on your machine (e.g. Cygwin command
prompts).
3 From a DOS command prompt, enter the following sequence of commands:
cd \cygwin\bin
ash
/usr/bin/rebaseall
If the problem persists after you have carried out the above steps, solution
2 is to restart your computer.
Problem description
The monolingual term extraction step fails, no monolingual term candidates
are extracted.
22
Troubleshooting
Possible explanation #1
There is no Acrolinx IQ language server running for the relevant language
at the IP address you provided, or you are not licensed to run checks in the
given language.
Solution when explanation #1 applies
Make sure you have the appropriate license and that you have a language
server running for the relevant language.
If you do not have a license for the relevant language, you can still do
multilingual term extraction on this language, but only if you either make it
a target language (in which case monolingual term extraction is a helpful but
not required part of the multilingual term extraction process) rather than the
source language, or provide a list of the terms in this language for which you
want to extract translations.
Possible explanation #2
The language server configuration properties are configured in a way that is
preventing monolingual term extraction via the Multilingual Term Extraction
tool. Even if you have previously set the configuration properties correctly,
this problem can emerge as a result of the settings being overwritten, e.g.
due to resources being upgraded.
Solution when explanation #2 applies
See the topic Confirming and Editing Term Harvesting Settings on page 7.
Modify the configuration properties accordingly, then restart the affected
language server(s).
Possible explanation #3
The rule set that you specified for the relevant language in Advanced Settings ➤
Monolingual TE ➤ Rule set by language does not exist. If this is the case, your log
will contain something like
[ERROR] Error while trying to check "C:\my_output_dir\monolingual
TE_EN\input1.clc"
com.acrolinx.util.soap.AcrocheckException: Unsupported rule set
"NonRules"
Solution when explanation #3 applies
In Advanced Settings ➤ Monolingual TE ➤ Rule set by language change the setting for
the relevant language to one of the existing rule sets for that language (and
make sure that the configuration properties are correct with respect to that
rule set -- see explanation #2).
Index
23
Index
A
P
Acquiring Third-party Tools • 5, 6
Performance Optimization • 11
Preparing the Results for Validation • 13
C
Context-specific Help • 12
R
Running Multilingual Term Extraction • 12
I
Installation Major Steps • 3
Installing Cygwin • 4
O
Opening Existing Results • 12
S
Starting the Multilingual Term Extraction Tool • 11
Starting the Multilingual Term Extraction Tool for
the First Time • 8
V
Validating the Results • 15