Table Of Contents
Importing Employee and Department Data
Importing Data from Multiple Sources
Data Import Methods
Data Import Process Flow
Importing Spoken Names
Importing Active Directory Employee Data
Active Directory Authentication Requirements
Mapping Speech Connect Database Fields to Active Directory and Cisco Unity Fields
Employee Data File Requirements
Description
File Format
Example
Importing an Employee File
Performing a Bulk Upload of Spoken Name Files
Department Data File Requirements
Description
File Format
Examples
Importing a Department File
Rebuilding the Grammar
Importing Employee and Department Data
This chapter describes how to format employee and department records, and explains how to import the data into the Speech Connect database. It also explains how to import Active Directory employee data from a Cisco Unity system into the Speech Connect database. Employee and department data files can be imported as often as needed to keep the database updated.
See the following sections:
•
Importing Data from Multiple Sources
•Importing Active Directory Employee Data
•Employee Data File Requirements
•Importing an Employee File
•Performing a Bulk Upload of Spoken Name Files
•Department Data File Requirements
•Importing a Department File
•Rebuilding the Grammar
Importing Data from Multiple Sources
This section describes the methods of importing data from multiple sources into the Speech Connect database, the flow of data as it is imported and processed, and the ways to import spoken name recordings into Speech Connect.
Data Import Methods
Speech Connect provides three ways to import data into its database:
•The Active Directory Import tool in the Speech Connect interface obtains employee data and spoken names from a Cisco Unity system by accessing Active Directory and importing the data into the Speech Connect database (see the "Importing Active Directory Employee Data" section).
•The Import Employees page in the Speech Connect interface allows you to upload an employee data file (which should be a CSV data file) from an HR database, Excel spreadsheet, or other data source, which then imports the data into the Speech Connect database (see the "Importing an Employee File" section).
•An automated data feed can be created by using sFTP to deposit an employee or department data file into the Speech Connect scfeeds directory on a periodic basis. Your sFTP login is scfeeds with the password that you set for scfeeds during installation. You must transfer the data file into the scfeeds directory.
Data Import Process Flow
Figure 4-1 shows the multiple source data import process flow.
Figure 4-1 Data Import Process Flow
With all three data import methods (Active Directory Import tool, Speech Connect interface import, and scfeeds import), the data file is placed in the Feeds/tenant_id directory and then processed in the same way for loading into the Speech Connect database. Data files that are sFTPed to the scfeeds directory are moved to the Feeds/tenant_id directory for processing when Scheduled Tasks are run.
The first processing step determines whether to load the file as an initial load (all new records) or as an update (mix of records for add, delete and update). If the process is an initial load, the data file is validated for the correct number of fields, correct order of fields, and compliance with field size constraints. Invalid records are rejected, and valid records are marked for adding to the database and placed in a transaction file for processing by the loader.
If the process is an update, the data file is validated for the correct number of fields, the correct order of fields, and compliance with field size constraints. Invalid records are rejected, and the remaining records are compared against the history file to determine which operation needs to be performed (add, delete, or update) and placed in a transaction file for processing by the loader.
When the transaction file has been processed successfully, the data file is saved as the current history file (for example, employee.csv.hst), which is used the next time an update is run for this same data feed. If the load process fails, the history file remains unchanged.
Note You can configure the Active Directory Import tool to query Active Directory and update the Speech Connect database when Scheduled Tasks are run.
Importing Spoken Names
Spoken names are played for disambiguation and transfer during the call session. During the rebuild process, Speech Connect creates a Text to Speech (TTS) recording for every employee who does not have a spoken name recording. These TTS recordings can be replaced later with spoken name recordings.
Recorded spoken names can be imported into Speech Connect in three ways:
•The Active Directory Import tool builds a new WAV file for every employee for which it finds a recorded spoken name in Active Directory, and imports the files into the spoken_name directory (see the "Importing Active Directory Employee Data" section).
•On the Modify Employee or Modify Department page, you can upload a recorded sound file (8k PCM Mu-law or WAV) to the spoken_name directory for an employee (see the "Employee Administration" section on page 7-1), or a department (see the "Department Administration" section on page 7-5).
•The ssp/prompts/seaa/1/spoken_name directory is not directly accessible via sFTP. However, you can bulk upload recorded names that have been obtained from a source other than Active Directory (see the "Performing a Bulk Upload of Spoken Name Files" section).
•Name Collect can be used to request that employees record their names. Enable Name Collect on the Application Setup page (see the "Name Collect Settings" section on page 6-10), and when employees with TTS spoken names next call the system, they will be asked to record a name. If employees want to later re-record the spoken names, you can set the Record Name option in the employee records to "Yes" and they will be prompted to record a name when they next call the system. Make sure that Name Collect is enabled on the Application Setup page in order for it to operate, even when re-recording employee names.
Importing Active Directory Employee Data
Revised May 2009
Caution When Active Directory data is imported into Speech Connect, any manual changes that you later make on the Modify Employee page will be preserved when the Active Directory Import tool is run again, unless the record was changed in Active Directory. Changes made in Active Directory take precedence over changes made in the Speech Connect interface.
The Active Directory Import tool collects Cisco Unity employee records from an existing Active Directory with a Cisco Unity-extended schema, and imports them into a Speech Connect database, as follows:
•Connects to the server where Active Directory is installed
•Binds to a user account in Active Directory
•Queries for all user information
•Creates a CSV file that contains the user data
•Builds a new WAV file for every employee for which it finds a recorded spoken name, which will be played as is for disambiguation and transfer. (The WAV file is named employee_id.wav, where employee_id is the actual ID stored in Active Directory. For example, 90564CB5-85D2-421A-833C-388E6CB6B33E.wav.)
•Imports the file into the local Speech Connect database
•Installs the spoken name files on the local Speech Connect system
Note The user login to Active Directory must have permissions to search and retrieve data. The server must be a Global Catalog (GC) server. Data import from a Domain Controller (DC) server will not work.
To Import Active Directory Data
Step 1 In the Speech Connect interface, click Employees and then click Active Directory Import. The Active Directory Import Employees page appears.
Step 2 Enter data in all fields.
The Active Directory Import fields are described in Table 4-1.
Step 3 To verify that the Active Directory profile settings are valid, click Validate.
Step 4 Click Save to save the configuration settings of this profile (only validated profiles are allowed to be saved).
Step 5 To run the import process, click Execute. Active Directory Import performs the following processes:
•Queries Active Directory for employee data.
•Creates a employee data file.
•Loads the data file onto the Speech Connect system.
•Validates the data.
•Imports the data into the database.
Note Any new or updated employee names or tiebreakers will take effect in speech processing after the next rebuild. After Scheduled Tasks are run, the grammar is automatically rebuilt. For instructions on manually rebuilding the grammar, see the "Rebuilding the Grammar" section.
Caution Running the Active Directory Import process may affect CPU and network performance. We recommend that you run this process during off-peak hours.
Table 4-1 Active Directory Import Fields
Field
|
Description
|
Values
|
Profile
|
Name that you gave this import when you clicked New or New Copy and saved this setup.
|
Alphanumeric characters
|
User Name
|
Enter the Active Directory User ID.
|
User name, for example, administrator
|
Password
|
Enter the Active Directory Password.
|
Password, for example, administrator
|
Server DNS Name
|
Enter the host name of the Active Directory system where the Active Directory is installed. The server must be a Global Catalog (GC) server. Data imported from a Domain Controller (DC) server will not work.
|
Host name, for example, "unity-test"
|
Transfer Attribute
|
Selects the attribute that is used to define the extension that callers will be transferred to.
|
•ciscoEcsbuTransferId
•ciscoEcsbuDtmfId
|
Tiebreaker 1
|
Selects the tiebreaker 1 attribute used for disambiguation during call recognition.
|
•Physical Delivery Office Name
•City or Locality
•State
•Country/Region
•Job title
•Department
•Company
|
Tiebreaker 2
|
Selects the tiebreaker 2 attribute used for disambiguation when tiebreaker 1 is insufficient to resolve a match during call recognition.
|
•Physical Delivery Office Name
•City or Locality
•State
•Country/Region
•Job title
•Department
•Company
|
Security
|
Selects the security type for the connection to Active Directory. For more information, see the "Active Directory Authentication Requirements" section.
|
•SSL—Secure Socket Layer
•Digest—Encryption handshake protocol
•Clear Text—Password and login are unprotected (not recommended; use only if necessary)
|
Basedn
|
Revised July 2008
The Base Distinguished Name defines where to start the search, including the domain controller name of the system and any organizational units (OU) or containers (CN) to search.
OU and CN values can be used to restrict searches, but are not required.
|
Revised July 2008
Organizational units and domain controller name. For example:
OU=Voicemail2-2,OU=Voicemail2-1,
OU=Voicemail2,OU=Voicemail,
DC=unity-test,DC=com
or
CN=Users,DC=unity-test,DC=com
or
OU=Boston,
CN=Users,
DC=unity-test,DC=com
|
Scope
|
Selects the depth of the Active Directory search.
|
•Base—Search only the Basedn
•One—Search one level below the Basedn
•Sub—Search all of the branches (containers) below the Basedn
|
Search Filter
|
Selects the attributes used to filter the Active Directory search:
(Users Only)—Includes only Cisco Unity users.
(Users and Contacts)—Includes Cisco Unity users plus external users. Note that external users will cause calls to be transferred to the operator because external users do not have extensions defined in their employee record.
|
•Users Only—Cisco Unity users only
•Users and Contacts—Cisco Unity users plus external users
|
Automated Processing
|
Sets the Active Directory Import tool to query Active Directory and update the Speech Connect database when Scheduled Tasks are run. If no updates in Active Directory are available, the import process does not run. (For more information about Scheduled Tasks, see the "Scheduling Names Tuning Tasks" section on page 8-2.)
|
•Yes—Runs Active Directory Import tool when Scheduled Tasks is run
•No—No automated processing
|
Active Directory Authentication Requirements
Table 4-2 describes Active Directory authentication requirements.
Table 4-2 Active Directory Authentication Requirements
Security Mechanism
|
Username
|
DNS Name
|
Notes
|
Clear Text
|
UPN (User Principal Name). For example, username@domain.com.
|
FQDN (Fully Qualified Domain Name). For example, server.domain.com.
|
A DNS name cannot be a numerical IP address.
|
Domain qualified user name. For example, domain\username.
|
samAccountName. For example, username.
|
SSL
|
UPN (User Principal Name). For example, username@domain.com.
|
FQDN (Fully Qualified Domain Name). For example, server.domain.com.
|
•A certificate must be installed on the Active Directory domain controller in order for SSL to work. See the Microsoft Knowledge Base article #321051 on the Microsoft support website.
•A DNS name cannot be a numerical IP address.
|
Domain qualified user name. For example, domain\username.
|
samAccountName. For example, username.
|
Digest
|
UPN (User Principal Name). For example, username@domain.com.
|
FQDN (Fully Qualified Domain Name). For example, server.domain.com.
|
•The FQDN must be a registered SPN (Server Principal Name) on the Active Directory domain controller. If the FQDN needs to be added to the Active Directory server as a registered SPN, the user must reset his or her password to ensure that the credentials are up to date, or authentication will fail.
•A DNS name cannot be a numerical IP address.
•The domain and forest levels can both remain at Windows 2000; there is no need to change these to Windows 2003. The important factor is that there are no longer Windows 2000 domain controllers running.
•If you do not have any Windows 2000 domain controllers in your environment now, but did in the past, and if the user account was created while the domain controllers were Windows 2000, the password should work. However, if the password was also created then and has not been changed, resetting the password of the existing account (even resetting it to the same password) will allow the account to work with Digest.
|
Domain qualified user name. For example, domain\username.
|
samAccountName. For example, username.
|
Mapping Speech Connect Database Fields to Active Directory and Cisco Unity Fields
Table 4-3 shows a mapping of Speech Connect database fields to Active Directory and Cisco Unity fields.
Table 4-3 Mapping of Speech Connect Database Fields to Active Directory and Cisco Unity Fields
Speech Connect Field Name
|
Active Directory Attribute
|
Description
|
Active Directory Display Name
|
Cisco Unity Display Name
|
subscriberid
|
objectGUID
|
Unique identifier
|
First name
|
1
|
firstname
|
givenName
|
Subscriber first name
|
Last name
|
|
lastname
|
sn
|
Subscriber last name
|
|
|
middlename
|
|
Subscriber middle name
|
|
|
nickname
|
|
Subscriber nickname
|
|
|
spokenname
|
msExchRecordedName
|
Name recorded in Active Directory
|
|
|
language
|
|
Preferred language of subscriber
Default: English
|
|
|
email_addr
|
mail
|
Subscriber e-mail address
|
Email Addresses
|
|
list_in_ext_dir
|
ciscoEcbsuListInUM
Directory
|
List in External Directory
|
|
List in external directory
|
tiebreaker1
|
PhysicalDelivery
OfficeName
l
st
co
title
department
company
|
Physical delivery
Office name
City or locality
State
Country/region
Job title
Department
Company
|
Office
City
State/Province
Country/Region
Title
Department
Company
|
|
tiebreaker2
|
PhysicalDelivery
OfficeName
l
st
co
title
department
company
|
Physical delivery
Office name
City or locality
State
Country/region
Job title
Department
Company
|
Office
City
State/Province
Country/Region
Title
Department
Company
|
|
office_extension
|
ciscoEcsbuTransferID
ciscoEcsbuDtmfID
|
Subscriber phone extension for transfer
|
|
Transfer ID
DTMF Access ID
|
office_phone
|
|
Subscriber office phone number
|
|
|
mobile_phone
|
|
Subscriber mobile phone number
|
|
|
home_phone
|
|
Subscriber home phone number
|
|
|
other_phone
|
|
Subscriber other phone number
|
|
|
other_office
|
|
Subscriber other office phone number
|
|
|
Employee Data File Requirements
This section describes how to format an employee data file, which can be created from an HR database, Active Directory, an Excel spreadsheet, or another data source, and which is used to import employee records into the Speech Connect database.
Description
The employee data file contains a record for each person to be included in the Speech Connect database. Record data includes an employee ID, full name, nickname (if any), e-mail address, office extension, office phone, cell phone, home phone number, other phone, other office phone, and tiebreaker values such as physical delivery office name, city or locality, state, country/region, job title, department, or company.
File Format
The employee data file is a comma delimited ASCII file. Each line contains a record for one person, and must adhere to the file format rules described in this section. Any record that is missing required data or that contains invalid data is logged as containing an error and is not processed. Empty data fields must be denoted by consecutive commas. The record must end with a new line.
General employee data file format rules are:
•One header record is followed by one or more data records.
•All records are delimited by a new line.
•Record fields are delimited by a comma; field values are enclosed in double quotes ("value").
•All fields must be present; empty values are represented by a pair of double quotes ("").
•Blank fields are allowed except when indicated as required.
Caution If your CSV file is handled in Microsoft Excel, be sure to format the cells as text. Otherwise, numbers entered into a cell may be converted to scientific notations and then loaded as invalid data in the database.
The header record must be formatted as follows:
subscriberid,firstname,lastname,middlename,nickname,spokenname,language,email_addr,
list_in_ext_dir,tiebreaker1,tiebreaker2,office_extension,office_phone,mobile_phone,home_phone,
other_phone,other_office
The employee data fields, which are defined in Table 4-4, must be in the same order as the header record.
Table 4-4 Employee Data Fields
Field
|
Type/Length
|
Required
|
Comment
|
subscriberid
|
40 characters (max)
|
Yes
|
Must be a unique identifier for each subscriber.
This identifier will be associated with a subscriber and should not change.
No two records can have the same value.
|
firstname
|
2 characters (min) to
32 characters (max)
|
Yes
|
Should be a proper first name consisting of letters, -, and _. The system provides common synonyms such as Steve for Stephen.
To use initials such as DJ for a firstname, you must put an underscore between the characters. For example d_j.
|
lastname
|
2 character (min)
32 characters (max)
|
Yes
|
Should be a full last name consisting of letters, -, and _. The system splits hyphenated last names into all varieties. For example, smith-jones results in smith, jones, and smith jones as possible last names. You can also use an apostrophe ('). For example, D'Auria.
|
middlename
|
32 characters (max)
|
No
|
Middle name or initial of the employee.
|
nickname
|
2 character (min)
32 characters (max)
|
No
|
Can be another known name consisting of letters, -, and _.
Do not include common nicknames such as Mike for Michael, because they are already listed in the master synonym list. This field can be empty.
To use initials such as DJ for a nickname, you must put an underscore between the characters. For example, d_j.
|
spokenname
|
120 characters (max)
|
No
|
Contains the spoken name filename (8k PCM Mu-law or WAV)
|
language
|
5 characters (max)
|
Yes
|
Five-character representation of the user's spoken language from the document "IETF (Internet Engineering Task Force). RFC 1766: Tags for the Identification of Languages, ed. H. Alvestrand. 1995."
The default is en-US (English United States).
|
email_addr
|
80 characters (max)
|
No
|
E-mail address of the subscriber.
|
list_in_ext_dir
|
integer
|
Yes
|
Determines if the employee is included in the grammar for an external directory. Used primarily when you are setting up internal and external directories. Grammars can be configured to include all employees, or to exclude employees that are set to 0 in this field.
•0—Do not list in directory. (Note that employee can still be listed in the internal directory.)
•1 (default)—List in directory.
|
tiebreaker1
|
60 characters (max)
|
No
|
Tiebreaker1 value to be announced as is when more than one match occurs. Examples include a physical delivery office name, city or locality, state, country/region, job title, department, or company.
Must consist of alphanumerical characters, spaces, and underscores.
|
tiebreaker2
|
60 characters (max)
|
No
|
Tiebreaker2 value to be announced as is when tiebreaker1 is insufficient to resolve a match.
Must consist of alphanumerical characters, spaces, and underscores.
|
office_extension
|
20 characters (max)
|
No
|
Employee office extension.
|
office_phone
|
20 characters (max)
|
No
|
Complete phone number of the employee, including country code (optional, if needed) and area code. This string is used by the system to play back the phone number if "Play Phone Number" is set to Yes in the application. (If this field is blank, then the extension is played back.)
|
mobile_phone
|
20 characters (max)
|
No
|
Complete phone number of the employee cell phone.
|
home_phone
|
20 characters (max)
|
No
|
Complete phone number of the employee home phone.
|
other_phone
|
20 characters (max)
|
No
|
Complete phone number of any other phone that is associated with the employee.
|
other_office
|
20 characters (max)
|
No
|
Other office extension used by the employee.
|
Caution Do not use extra spaces in fields. The system recognizes spaces as characters. Therefore, you must be careful to use spaces only when necessary for separation. Do not use leading or trailing spaces.
|
Example
The following is an example of an employee data file:
subscriberid,firstname,lastname,middlename,nickname,spokenname,language,email_addr,
list_in_ext_dir,tiebreaker1,tiebreaker2,office_extension,office_phone,mobile_phone,
home_phone,other_phone,other_office<CR>
"5555009","John","Smith","","Clyde","5555009.wav","en-US","jsmith@abcfinance.com","1",
"Engineering","Seattle","5009","9784423009","5085170696","7812249043","","5042"<CR>
Note The header and every record, including the last record, must end with a carriage return. Make sure that the last record in the file ends with a carriage return, or the data file will be rejected.
This example data file will import the following data:
Subscriberid: 5555009
FirstName: John
LastName: Smith
Middle Name:
NickName: Clyde
SpokenName: 5555009.wav
Language: en-US
Email_addr: jsmith@abcfinance.com
List_in_ext_dir: 1 (Yes)
Tiebreaker1: Engineering
Tiebreaker2: Seattle
Office_extension: 5009
Office_phone: 9784423009
Mobile_phone: 5085170696
Home_phone: 7812249043
Other_phone:
Other_office: 5042
Importing an Employee File
Revised February 2008
The Import Employees page allows you to import employee data from a CSV file into the Speech Connect database. This feature is used to either load all data from the employee file into the database, or to update the database with any data from an employee file not currently in the database.
To Import an Employee File Into the Database
Step 1 In the Speech Connect interface, click Employees and then click Import. The Import Employees page appears.
The Import Employees fields are described in Table 4-5.
Step 2 Click Browse to locate and select the file. Click Open. The employee data file appears in the Upload File field.
Note To avoid conflicts with CSV files that reside on a remote system drive, we recommend that your CSV file reside on and be loaded from your local PC system drive.
Step 3 Select Load or Update, depending on whether this is an initial load or an update.
Step 4 Click Validate. The employee data is validated, and a status message appears. If the data passes validation, a Load button appears on the page.
Step 5 Click Load. The employee data file is loaded on the system, and the data is processed as an initial load or as an update to the database, depending on the process you selected.
Step 6 If you need to upload recorded sound files (8k PCM Mu-law or WAV), you can upload the files individually within each employee record (see the "Employee Administration" section on page 7-1), or upload files in bulk directly to the spoken_name directory (see the "Performing a Bulk Upload of Spoken Name Files" section).
Note Any new or updated names or tiebreakers will take effect in speech processing after the next rebuild. The grammar is rebuilt automatically when Scheduled Tasks are run. To rebuild the grammar manually, see the "Rebuilding the Grammar" section.
Caution Importing employee data can affect CPU and network performance. We recommend that you import data during off-peak hours.
Table 4-5 Import Employees Fields
Field
|
Description
|
Values
|
Upload File
|
Allows you to search your local drive or network and select an employee data file.
|
Any valid file containing employee data. For example, employee.csv.
|
Load
|
Imports all data from the employee data file into the database.
|
Load—Loads all employee data into the database.
|
Update
|
Compares the employee data file with its history file, and imports only the data that is not currently in the database.
|
Update—Updates the database with any data from the employee file that is not currently in the database.
|
Performing a Bulk Upload of Spoken Name Files
Do the following "To Import Spoken Names into the spoken_name Directory" procedure to perform a bulk upload of spoken name recordings (8k PCM Mu-law or WAV) into the spoken_name directory on the Speech Connect server.
To Import Spoken Names into the spoken_name Directory
Step 1 Log in to the Speech Connect server as scadmin, and enter the following:
cd /tmp
mkdir name_files
cd name_files
Step 2 By using sFTP, log in to the server as scadmin, and transfer the spoken name files into /tmp/name_files.
Step 3 Log in to the server as root, and enter the following:
cd /tmp/name_files
chown seaa *
chgrp seaa *
cp /tmp/name_files/* /opt/cisco/speechconnect/ssp/prompts/seaa/1/spoken_name
Step 4 While logged in as root, clean up the temporary files and directory by entering the following:
cd /tmp/name_files
rm *
cd ..
rmdir name_files
Department Data File Requirements
This section describes how to format a department file, which may be created from an HR database, an Excel spreadsheet, or another data source, and which is used to import department records into the Speech Connect database.
Description
The department data file contains a record for each department included in the Speech Connect directory. This list may include departments such as Human Resources, Customer Service, and Payroll.
File Format
The department data file is a comma delimited ASCII file. Each line contains a record for one department and must adhere to the following rules. Any record that is missing required data or contains invalid data is logged as containing an error and is not processed. Empty data fields must be denoted by consecutive commas. The record must end with a new line.
General department file format rules are:
•One header record is followed by one or more detail records.
•All records are delimited by a new line.
•Record fields are delimited by a comma; field values are enclosed in double quotes ("value").
•All fields must be present except alternate_department_title; empty values are represented by a pair of double quotes ("").
Caution If your CSV file is handled in Microsoft Excel, be sure to format the cells as text. Otherwise, numbers entered into a cell may be converted to scientific notations, and then loaded as invalid data in the database.
A department-only load file can consist of a header record and one or more data records. The alternate_department_title field can be repeated as many times as needed. Blank fields are allowed except when indicated as required.
The header record with one alternate_department_title (optional) is as follows:
department_code,department_title,phone_transfer,phone_play,alternate_department_title
If one or more alternate_department_title is included, the header record must contain the words alternate_department_title for each additional field, separated by a comma.
The department data fields, described in Table 4-6, must be in the same order as the header record.
Table 4-6 Department Data Fields
Field
|
Type/Length
|
Required
|
Description
|
department_code
|
16 Characters (max)
|
Yes
|
A unique identifier for each department. This identifier should not change. No two records can have the same value.
|
department_title
|
60 Characters (max)
|
Yes
|
The title of the department. This field is used to record the TTS prompt, which is played back during disambiguation. The entire title has to be spoken for recognition.
|
phone_transfer
|
20 Characters (max)
|
No
|
The phone extension of the department. This is used by the system as is to transfer the call from the local system.
|
phone_play
|
20 Characters (max)
|
No
|
The phone number or extension of the department, which is played as is during call transfer.
|
alternate_department_title
|
60 Characters (max)
|
No
|
An alternate name for the department. You can enter values for Name 1 through Name 20. If no alternate names are specified, then only the Title as is will be used for recognition.
|
Examples
The first example does not include an alternate_department_title:
department_code,department_title,phone_transfer,phone_play<CR>
"1022","Human Resources","5551212",""<CR>
The second example includes one alternate_department_title:
department_code,department_title,phone_transfer,phone_play,alternate_department_title<CR>
"1022","Human Resources", "5551212","","Employment Resources"<CR>
The third example includes three alternate_department_title fields:
department_code,department_title,phone_transfer,phone_play,alternate_department_title,
alternate_department_title,alternate_department_title<CR>
"1039","Accounting","","","Accounting","Finance","Payroll"<CR>
Importing a Department File
Revised February 2008
The Import Department page allows you to import department data from a CSV file into the Speech Connect database. This feature is used to either load all data from the department file into the database, or to update the database with any data from the department file that is not currently in the database.
Caution The first department file that is loaded on to the system is added to the database. The next file must have the same name as the first file in order to be used as an update to the database because the contents of the first file are compared to the current file to determine how transactions are processed.
To Import a Department File into the Directory
Step 1 In the Speech Connect interface, click Departments and then click Import. The Import Departments page appears.
The Import Departments fields are described in Table 4-7.
Step 2 Click Browse to locate and select the file. Click Open. The department data file appears in the Upload File field.
Note To avoid conflicts with CSV files that reside on a remote system drive, we recommend that your CSV file reside on and be loaded from your local PC system drive.
Step 3 Select Load or Update, depending on whether you are performing an initial load or update.
Step 4 Click Validate. The department data is validated, and a status message appears. If the data passes validation, a Load button appears on the page.
Step 5 Click Load. The department data file is loaded on the system, and the data is processed as an initial load or as an update to the database, depending on the process you selected.
Step 6 If you need to upload recorded sound files (8k PCM Mu-law or WAV), you can upload the files individually within each department record (see the "Department Administration" section on page 7-5), or upload files in bulk directly to the spoken_name directory (see the "Performing a Bulk Upload of Spoken Name Files" section).
Note Any new or updated department names will take effect in speech processing after the next grammar rebuild. The grammar is rebuilt automatically when Scheduled Tasks are run. To rebuild the grammar manually, see the "Rebuilding the Grammar" section.
Table 4-7 Import Departments Fields
Field
|
Description
|
Values
|
Upload File
|
Allows you to search your local drive or network and select a department data file.
|
Any valid file containing department data.
|
Load
|
Imports all data from the department data file into the database.
|
Load—Loads all department data into the database.
|
Update
|
Compares the department data file with its history file, and imports only the data that is not currently in the database.
|
Update—Updates the database with any data from the department file that is not currently in the database.
|
Rebuilding the Grammar
After importing employee and department records into the database, you must rebuild the grammar so that the names can be used for call recognition. Your system is shipped with the Linguistic Definitions Dictionary (LDD), which is a master phonetic dictionary used to update your database with phonetic pronunciations of your names. During the Rebuild Grammar process, Speech Connect compiles your records into a speech recognition grammar by building phonetic pronunciations of your names for use during call recognition.
As you add and update user records, the new records are queued for the grammar rebuild process. The Speech Connect interface allows you to run the Rebuild Grammar process immediately or to schedule it for periodic updates, as follows:
•To rebuild the grammar immediately, go to the Grammar Setup page by clicking Config > Grammar. This process builds the grammar immediately, but does not create Text to Speech (TTS) files for employees who do not have a recorded name file. For more information about setting up grammars for your applications, see the "Grammar Setup" section on page 6-11.
•To rebuild the grammar at scheduled intervals and to create TTS files for employees who do not have a recorded name file, go to the Scheduled Tasks page by clicking Config > Scheduled Tasks. For more information about scheduling grammar rebuilds, see the "Scheduling Names Tuning Tasks" section on page 8-2.
