Search Service

[Forumadmin]

Two new services will appear in the menu of club members when they next start a session in Jowett Member Services.
After 4 months of research and development a search server now delivers a service tailored for club members. This will develop over time as there are many facilities which can be switched on, but for now it provides a simple way to search the many Jowett digital resources.
First to be indexed were the 1044 webpages in the old website comprising contributions from members from 1999 to 2006.
Next were the 1371 pdf files initially stored in the Gallery which then became the JowettTalk Library. So you can now search and view all the Jowetteers as well as the other club's magazines currently in the Library. I will shortly be adding other types of resources, such as Word documents, that are in the library and also be adding many more documents. The system is capable of Optical Character Reading images, but that has not been tested or implemented so they are not included in the currently available index.

SearchServer01.PNG

Some of the documents contain information that may be sensitive so please treat with respect and do not distribute or communicate. It is for this reason that the resources are protected by the Member Services and JowettTalk access control. However, text can be copied and documents can be referred to in JowettTalk posts some of which may be in the Public domain (rather than in a protected part of JowettTalk or on jowett.org or jowett.net). It is usually better to put such composed topics and posts in the Cooperative Space or Personal Album sections of JowettTalk (rather than in other sections of JowettTalk) as these are only viewable by Club Members, since this protects the sensitivity of any information copied. Referenced information that is in JowettTalk will still be protected by the JowettTalk access control. Composed articles can be moved to a more appropriate place in the Library by Curator.

I am hoping the search capability will enable you to compose articles for the Jowetteer, JowettTalk and the general press. I have researched and tested extensions to the system that further assists journalism. This advanced functionality includes: faceted search, clustering, filters, snippets, synonyms, stopwords, highlighting, categorization, “find similar”, automatic thumbnail screenshot inclusion, boost/reduce relevance. I have also looked at providing semantic grouping, word clouds, connections and networks in a visual graph view. But do not let this complexity put you off. The current system is as simple as Google. I may offer the extended system to those who would like to try it. So please contact me if you would.

The session with the search server is independent of the session with Member Services, so has a different timeout. If either times out then you will need to restart the Member Services session and click on the Menu Search button again. Hopefully you have stored the Member Services activation link in your browser so this will be a quick process.

If you have never registered for a session of Member services then click here.
Please read the help guide linked in the first sentence on that page so you know what to do.

Please try it out and contact me with any issues or suggestions.

[Forumadmin]

I am still battling with the software trying to get what I want to out of it.
Some issues so far on the Archive search.
These issues are mainly caused by the need to copy the attachments in JowettTalk and rename them with a unique filename whilst still keeping them secure.
I have now added some other document types to the pdfs. But these are proving troublesome so I may well convert them all to pdfs. Browsers do not display Microsoft documents and neither does JAVA (in which this system is written) unless I buy a service or software. This is not a big issue but some Microsoft docs might not convert perfectly. The alternative is to convert to HTML which is what I did with the Technical Notes from Mike Allfrey. This is a much better solution but means reorganising the Library. I have converted all to pdf.

There is another route that I will investigate which is using the search system to directly access the documents via the JT database. But that requires extensive investigation.
There is a viewer for pdf files so you can click the view link to view the pages and see the highlighted search finds. There is also a download button.
There is no viewer other than one for pdf files.
I have not enabled linking back to the source document.
The Open folder link does not work so you cannot go directly to the document, but you can find the document by going to the post that contains it, for documents that do not have a 'View' button.

This is an index entry when searching for the 'doc' type and search criteria 'Jowett'
The first line is the document title, the second the date the document was added to the index, the third a snippet that contains the searched for text (if it was not in the title), the fourth are the tools that can be used on the document (only 'view' works ), the last line is the file path on the search server. (A separate server has to be used as the hosting service for the Archive does not allow JAVA).

Code: Select all

 JOWETT CAR CLUB Ltd
Sep 16, 2020
You can enter this form on-line where your membership details will be filled in by the system and you can pay by BACS: See details in Jowetteer or on-line at  HYPERLINK "http://jowett.net"...
Open folder
home/jowett/…/JowettTalk/25674_2013 Registration Form -v10.doc 

You will see the file name at the bottom and the number (25674) preceding the actual file name, before the underscore, is the post identifier.
So you can put this in your browser address bar to get to the post with the attachment.

Code: Select all

https://jowett.net/forum/viewtopic.php?p=25674

I hope to make files viewable and directly downloadable as they are for pdfs. But there are only so many hours in the day.


The system should be pretty quick at searching but may be slow at delivering whole documents due to slow network speed.

but no doubt there will be many more issues.
This is day one of testing and development!

The good points are that it does highlight the search criteria in the documents as well as highlighting in the snippets.

[Forumadmin]

The jowett.org search allows you to click on the link to go directly to the source document. So as an example, if I search for my car registration number (NKD258), it will find occurrences not only of NKD258 but also NKD as it uses a fuzzy search technique and weights the find and promotes it in the order on the displayed list. Click on the document title to view the document. You can then use the browser find function (CTRL+F) to find the occurrences on the page.

[Forumadmin]

18:09:2020.
All non pdf documents in JT have been converted to pdf so now can be viewed with the VIEWER and so do not need to be downloaded to your device, saving its precious memory.
I think the presentation is good with text having the searched for text highlighted. Pages are neatly displayed and easily turned over. Powerpoints and pdf image collections can be flipped through making reading and viewing much easier.
If you notice any conversions that have issues please send me the offending file name.
The system automatically removed references to the deleted rtf, doc, xls and ppt files.
Note you can still use the JowettTalk post identifier in the file name to view the topic containing the document and thus view the original document.
e.g.

Code: Select all

https://jowett.net/forum/viewtopic.php?p=25674

I have noticed the positioning of some of the highlighting is not accurate. This is not a high priority issue!

When you are viewing a document you will see your search criteria in the Search box. You can change this to search within the document and highlight occurrences of the new search term. Unfortunately it does not tell you if it found it or allow you to go to the next occurrence so you have to page through the whole document to find out!

[Forumadmin]

Search weighting.PNG

Two 'words' are searched but the fuzzy search will also find parts of words but it will weight the results based on algorithms I set.
For instance the words occuring in the Title have more weight than those in the document text.

Search highlight.PNG

Clicking on the Viewer link brings up the first page of the document so you have to page through to find the highlighted search result.

[Forumadmin]

The Jowetteers from 1960 to 1982 have now been added so you have 60 years of them to search through and peruse.

[Forumadmin]

All the Jowett Jive magazines published by Ted Miller from 1986 (No 1) to 1996 (No41) have been added. Note I was missing no 3 and 4 so if you have these please contact me. There are some fantastic articles in these pages. So search away!

[Srenner]

Will get 3 and 4 so that the set is complete.

[Keith Clements]

Searching the Jowett Archive.

It was good to read the article from Tony Pluckrose in this month's Jowetteer and his intention to preserve the Jowett history, followed by his East Anglian officer report. This was something Mike Allfrey and I discussed in 2006 in Melbourne and so started the Jowett Gallery and Mike's amazing collation of Jowett Technical Notes. Since then Mike has collated a hundred sets of notes. These are available in the JowettTalk Technical Library along with over 1500 technical items contributed by members of all the Jowett Clubs.

The Historical section of the Library has about 400 items with most of the press cuttings having been contributed by members. So if you have any not in there, or a better copy, please scan or photograph and either attach to a post in JT or send to me.

The JowettTalk Library has virtually all the magazines from all 5 Jowett clubs currently totalling over 1500 magazines thanks to Ian AItken in the UK and Bryan Walker in NZ. If you have any that are not in there then please let me know.

All documents have been Optically Character Read so are searchable with the new search system which is available to members and is accessible from JowettTalk or Member Services after you login.

It would be great if the JCC documents in the Bradford Archive could be scanned, or those in the libraries of the other clubs and individuals, so we save the marque's history AND make it accessible to all.

Tony was relating the schism with SJCC which is documented in the Jowetteers of 1964 and 1965. Please read the month by month developments about it. There is an active discussion on bearings on JT. The search system found 27 documents about Glacier Bearings. So you can find out all that has ever been written on the subject! I searched for my car registration number and the search system found 72 results. Try it on yours. One day I may implement the OCR facility on the 15000 images in JowettTalk so that pictures of cars can be found easily.

Hopefully the system will inspire you to do a bit of research and write an article for the club magazine or generate a topic on JowettTalk.

[Keith Clements]

The search archive underwent a complete renewal this morning to sanitize all the file names. You may have experienced a glitch whilst this was happening. Over the twenty years that the website has been collecting documents many peculiar file names have been input. Some are a good sentence long, others have many spaces and special characters. Although the system stored these it does make management difficult, and sometimes they are difficult to read. So a script was written to sanitize them. The copy of the document in JT will remain the same for now.