Re: Search syntax...
> I've made two fixes to the search engine: > - prevented duplicates from appearing in the results > - case-insensitivity on our special field names in searches (date, > from, > subject, message) > > That tidies things up nicely. > > Jeff I'll say, Jeff! Yep, the dupes are gone. Interestingly, I've been doing lots of searches all evening and didn't notice when the change took place! If I capitalize 'From:' for an author search, I see on the results page that the system has "corrected" my search syntax to 'from:' rather than turning up 0 results. Thanks! Mike D. [Mike Devour, Citizen, Patriot, Libertarian] [[email protected]] [Speaking only for myself... ] -- To unsubscribe, send mail to [email protected].
Re: Search syntax...
I've made two fixes to the search engine: - prevented duplicates from appearing in the results - case-insensitivity on our special field names in searches (date, from, subject, message) That tidies things up nicely. Jeff On Thu, Feb 18, 2010 at 10:57 PM, M. G. Devour wrote: > Jeff, > > Thanks for the overview. I'm going to be working on a new web site for > my group, so I may attempt something like that. > > Mike D. > > > An "advanced search" form would basically have a bunch of fields like > > "subject" and "date" and "from" then string them together into a query > > syntax described below. Then redirect that query to the standard search > > URL. Implementation would most likely be in PHP. Hard part isn't the > > programming, it is designing an elegant suer friendly form that matches > > the query syntax. Actually that isn't so hard either. Hard part is > > making time to get it done. > > > > > -- > To unsubscribe, send mail to [email protected]. >
Re: Search syntax...
Jeff, Thanks for the overview. I'm going to be working on a new web site for my group, so I may attempt something like that. Mike D. > An "advanced search" form would basically have a bunch of fields like > "subject" and "date" and "from" then string them together into a query > syntax described below. Then redirect that query to the standard search > URL. Implementation would most likely be in PHP. Hard part isn't the > programming, it is designing an elegant suer friendly form that matches > the query syntax. Actually that isn't so hard either. Hard part is > making time to get it done. -- To unsubscribe, send mail to [email protected].
Re: Search syntax...
An "advanced search" form would basically have a bunch of fields like "subject" and "date" and "from" then string them together into a query syntax described below. Then redirect that query to the standard search URL. Implementation would most likely be in PHP. Hard part isn't the programming, it is designing an elegant suer friendly form that matches the query syntax. Actually that isn't so hard either. Hard part is making time to get it done. http://www.mail-archive.com/faq.html#search .
Re: Search syntax...
Yes, looks like there is a problem where we are displaying double search results. We'll look at it this week and fix that right up. MHonArc is all written in Perl. Lucene is built on Java, but we use it via a Python wrapper. The Mail Archive's wrapping layers and sorting code is written mostly in shell, with some PHP and Python sprinkled about. Thanks On Tue, Feb 16, 2010 at 6:47 PM, M. G. Devour wrote: > > Hi Mike. That's a very good point about the case sensitivity on the > > "from" and "date" directives. I'll see what I can do to fix that. > > Thank you, Jeff. I figure that while it's all fresh for me, the things > I stumble on have the best chance of helping future users. Once I'm > used to it all I'll have a harder time seeing possible improvements. > > I've noticed that in almost every search I do, each message turns up > twice, either right next to each other or separated by one or two other > results. For instance, a search I made that had 70 results represented > 36 unique pages, once I'd been through them all in detail. I couldn't > see a reason for the duplication. Any thoughts as to why? > > More results is better than less, most likely, and the nice contrast > between visited and unvisited page links speeds review, but it's an > interesting effect. > > > It's great to hear you are enjoying the search engine. We use Lucene > > for the per-list search indexes; it's certainly a high quality > > open-source engine. > > Yes, it's a pretty piece of code. Having it not choke on really short > terms is basically unique in what I've seen. I remember reading that > the index file might be a third the size of the database? If so, it's a > worthy investment. I think we're reaping the benefit of really cheap > storage and a text-only format... > > I forget... what language(s) are Mhonarc and Lucene implemented in? > Your sorting code? If I do any coding or tweaking, it'd be nice to be > able to offer it up for use by others. Some kind of search form has to > be do-able, if it's not already out there. > > Thank you, again. Excellent work. > > Mike D. > > [Mike Devour, Citizen, Patriot, Libertarian] > [[email protected]] > [Speaking only for myself... ] > > > > -- > To unsubscribe, send mail to [email protected]. >
Re: Search syntax...
> Hi Mike. That's a very good point about the case sensitivity on the > "from" and "date" directives. I'll see what I can do to fix that. Thank you, Jeff. I figure that while it's all fresh for me, the things I stumble on have the best chance of helping future users. Once I'm used to it all I'll have a harder time seeing possible improvements. I've noticed that in almost every search I do, each message turns up twice, either right next to each other or separated by one or two other results. For instance, a search I made that had 70 results represented 36 unique pages, once I'd been through them all in detail. I couldn't see a reason for the duplication. Any thoughts as to why? More results is better than less, most likely, and the nice contrast between visited and unvisited page links speeds review, but it's an interesting effect. > It's great to hear you are enjoying the search engine. We use Lucene > for the per-list search indexes; it's certainly a high quality > open-source engine. Yes, it's a pretty piece of code. Having it not choke on really short terms is basically unique in what I've seen. I remember reading that the index file might be a third the size of the database? If so, it's a worthy investment. I think we're reaping the benefit of really cheap storage and a text-only format... I forget... what language(s) are Mhonarc and Lucene implemented in? Your sorting code? If I do any coding or tweaking, it'd be nice to be able to offer it up for use by others. Some kind of search form has to be do-able, if it's not already out there. Thank you, again. Excellent work. Mike D. [Mike Devour, Citizen, Patriot, Libertarian] [[email protected]] [Speaking only for myself... ] -- To unsubscribe, send mail to [email protected].
Re: Search syntax...
Hi Mike. That's a very good point about the case sensitivity on the "from" and "date" directives. I'll see what I can do to fix that. It's great to hear you are enjoying the search engine. We use Lucene for the per-list search indexes; it's certainly a high quality open-source engine. Jeff On Tue, Feb 16, 2010 at 6:56 AM, M. G. Devour wrote: > Well, I've been playing with my shiny new archive! Great stuff! > > I stumbled on a less than intuitive detail of the search syntax... > > First I tried: > > "From: M. G. Devour" > > ...and got messages in which I was quoted by others. Okay, not what I > thought I was going to get, but makes sense when you read the > instructions. > > Then I moved the quotation mark and tried: > > From:"M. G. Devour" > > ...which gave zero results. Huh? > > Finally, I tried the *correct* syntax: > > from:"M. G. Devour" > > ...which turns up all of my posts, as intended. > > Same thing is true of date: searches. I'm curious why the from: and > date: directives bother to be case sensitive? It tripped me up there > for a bit! > > Otherwise, I've gotta say, wow, this is a powerful search facility! > > It is just so great to be able to use arbitrarily short search terms. > There are so many acronyms and legitimate items in our group's jargon > that are only 2 or 3 letters long, that the 3 letter minimum of most > forum software as well as the archive that we used to have, is really > crippling sometimes. > > Have people managed to design nice front-ends for this, similar to > Google's advanced search form? Something like that would be extremely > useful on our web site. > > Thanks, guys. This is turning out very nicely. > > Be well! > > Mike D. > > > > > > -- > To unsubscribe, send mail to [email protected]. >
