Using Apple e.g. Search Engine with
TeleFinder Web Servers
describes the e.g. software and provides details about how to set up
and use the software, specifically with TeleFinder Server 5.7.
This is beta 1.1b3 for PPC Only!
Note: If you have an older version of e.g. you MUST throw away your old indices and reindex your site. The old indices are not compatible with this version of e.g.
e.g. is a full text retrieval common gateway interface (CGI) that allows you to index your Web site and make it searchable to browsers surfing your site. It is a technology demonstration showing some of the features of a search engine developed in Apple's Advanced Technology Group. e.g. is not meant to be a full-featured product and is not supported by Apple Computer. Do NOT call the Apple support hotline if you have a problem with e.g.
The American Heritage Dictionary defines e.g. as follows:
"e.g. Lat. exempli gratia (for example)."
e.g. allows users to specify queries by example. After searching your site the user is presented with a "hit list" of documents that matched the query terms entered. The user may then select documents of interest and submit a new query specifying that the new hit list should contain documents like the selected documents from the previous hit list.
Follow the steps in
this section to set up the e.g. CGI. Once you've set up e.g., you can
modify several options and settings, as described in the next
1. Copy the contents of the e.g. folder to the root of your TeleFinder Web Server folder. The following files must be copied or the CGI will not function properly:
I would suggest you also include my graphics folder and headsearch.html.
You should also put in the EnglishStopwords and English Substitutions to prevent your index from becoming too large. Stopwords keeps e.g. from indexing common words like: the, of, at etc. Substitutions allows the search engine to interpret what the user is asking for. You may view the files in BBEDIT and modify them. Just make sure you don't change the filetype when saving.
2. Launch or Reload
TeleFinder Web Server.
3. Double-click the eg.acgi icon.
4. The first time you launch the CGI you're asked to agree to the terms of the license agreement. Read the agreement carefully and click the Accept button.
5. A dialog box asks you if you want to use the auto-setup feature.
NOTE: DO NOT USE THE AUTOSETUP FEATURE WITH TELEFINDER. Use the manual instructions!
Auto-setup will do
Set the eg root folder to the TeleFinder Web Server root folder.
Set the auto index date to midnight.(You may not want to use this feature at all. Memory requirements may be large. Run it manually).
Index the entire contents of your TeleFinder Web Server folder.
If you want to do all of those things, click the Setup button. If not, you will have to perform each step manually. Instructions for manual setup are contained later in this document in the various menu item descriptions.
6. You may want to customize the way the search form looks by modifying the "EGTemplate.html" file. The template file is divided into 3 sections: the header, the egcontent, and the footer. You have complete control of how the page looks except for the egcontent section.
The file must be called "EGTemplate.html" and must have the <egcontent> tag in it somewhere. Currently the egcontent section is hard-coded into the CGI itself and cannot be modified. This may change in the future.
7. Add a link from your home page to the CGI to allow browsers to search your Web site. The link should look like this: http://your.domain.name/AppleEG.PPC.acgi. Use the headsearch page as an example or simply modify this page for your use. If you dropped it in your web server folder and have completed the preceeding steps you should be able to use your index now by accessing http://yourdomain/headsearch.html
See the next section for details about how to control the performance of the e.g. CGI.
There are several
options and settings that you can modify to suit your site's needs.
All of the settings and options can be modified from the Control
menu. Each menu item is described in the order they appear in the
Update Index Now indexes all files immediately. Searching can continue until the very last phase of indexing, which should only last a few seconds.
Settings will bring up a dialog that contains some settings related to the index. You can modify the time that automatic indexing occurs by using the 24-hour specification. For example, if you want to index your site automatically each night at 8 p.m., change the settings to read 20:00.
You can also store the full URL for each document in the index. This is important if you want to move the index file to another Web site so that people can search the documents on your site without actually being at your site. For example, you index your site and find that your site is getting very busy and would like all searching to be done on a second Web server. You can then copy the CGI and index to the second site and point to that site for all of the searching. The browsers will then use the second site to search the index and will only hit your initial site to view those documents.
In order to add the URL to the index, type the URL in the appropriate field in the dialog. The URL should look something like: http://your.domain.name/. Don't forget that trailing slash!
Edit Folder List allows you to specify which folders of documents should be indexed. You may have 3 folders inside of your TeleFinder Web Server folder. If you only wanted to index 2 of those you would have to specify those 2 using this menu item. Make sure you leave out your log files if you have any of them residing in your web space folder!
The root folder is the Web server's root folder. In the case of TeleFinder Web Server, it allows you to specify a different folder as the root folder. The e.g. CGI needs to know which folder is the root folder so that the URLs that get generated for the hit list will be correct. You can change the root folder by clicking the "Root folder" button and selecting a new root folder.
If you used the auto-setup feature described previously, then the list of folders to index only contains one folder and that is the root folder. If you do not want the entire root folder to be indexed then you will want to remove that folder from the list and add the folders that you do want to be indexed.
You will need to use the Update Index Now menu item to make the changes take effect or wait until automatic indexing occurs.
Edit Omit List lets you specify suffixes of documents not to index. You may not want to index documents that end in .SPML for instance. Since e.g. only indexes text documents you may want to add .pdf or other types of files for security reasons or to speed up the indexing process.
Reload Template will let the CGI know that the EGTemplate.html file has been changed and should be looked at before handling any more requests. You will want to use this after making changes to the template file so that the changes will take effect immediately.
Auto Setup will reset all of the previously set settings and start the index from scratch. If you made any modifications to any of the above settings you will have to make them again after the auto setup has completed. If you receive any errors during the auto-setup procedure you should throw away the egindex file and start over again.
Verbose Messages is a toggle switch to turn on the verbose-message logging. Without verbose messages you will only see error messages in the status window. Having verbose messages turned on will add a line in the status window each time a search request is made.
e.g. maintains several logs that allow you to keep track of its usage:
The main log is called eglog. This is a plain-text file that contains all messages that appear in the status window in verbose message mode.
The eg_query_log.html file contains a log of all searches done using e.g. It contains IP addresses, query parameters, and number of documents found for each query done by e.g.
eg_index_info.html is a file that contains information about the index. You may want to make this file available to those searching your site so that they can see what kind of information is searchable. The file contains information about the number of files in the index, size of the index, when the index was created and updated, and which words appear most often in the documents indexed.
Hopefully you were
able to get your own search engine working. You can try mine at:
© 1996 Apple Computer, Inc. Apple and the Apple logo are trademarks of Apple Computer, Inc., registered in the United States and other countries. This file is modified for use on TeleFinder Web Servers by Bob Nunn, Operator Headgap Systems. firstname.lastname@example.org March 10, 1999.