Add Lanman to simple-search

[funderburkjim]

Two additional steps required to make Lanman dictionary available in simple-search:

• add LAN to dictnames.js
• Regenerate hwnorm1c.sqlite (see Add Lanman (to 7 places) hwnorm1#17)

For example, now this simple-search url works:

https://www.sanskrit-lexicon.uni-koeln.de/simple/lan/shiva

[gasyoun]

https://www.sanskrit-lexicon.uni-koeln.de/simple/lan/shiva

Hurray. But how one can know that such nice URLs exist from homepage?

[funderburkjim]

Obviously, they can't know that.

I still don't think the simple search is quite ready for full disclosure.

For instance, I ran into some problems such as

• https://www.sanskrit-lexicon.uni-koeln.de/simple/lan/भद्र
  • Problem is that the Devanagari is not properly handled: citation =
    %E0%A4%AD%E0%A4%A6%E0%A5%8D%E0%A4%B0
• Similarly with IAST:
  https://www.sanskrit-lexicon.uni-koeln.de/simple/lan/rāma
  • citation = r%C4%81ma
• Even for standard HK input, the search sometimes fails. I've
  noticed this for 'long' words:
  https://www.sanskrit-lexicon.uni-koeln.de/simple/lan/pratyakzadarSana
  In this example, 'working' never finishes, or you get 'error Internal Server Error'
  which is probably due to a timeout.

I'm dubious about wide usage of the simple link with such problems still unsolved.
And I currently think they are hard problems, especially the 'long word' problem.
Probably the Devanagari/IAST problems are not too hard.

And there are still a couple of problems you have pointed out (such as use of capital letters).

[gasyoun]

I still don't think the simple search is quite ready for full disclosure.

Let's fill the gap, so it can be finally done after three years of development.

%E0%A4%AD%E0%A4%A6%E0%A5%8D%E0%A4%B0

So it's a code issue, not Unicode. Same with r%C4%81ma

In this example, 'working' never finishes, or you get 'error Internal Server Error'
which is probably due to a timeout.

This one is harder - any clue?

Probably the Devanagari/IAST problems are not too hard.

Exactly, I'm even eager to hire a developer to solve them, because simple is really important for me.

[funderburkjim]

Use of Devanagari (or IAST) in url now works properly.

The need was to do a php 'uridecode' on the thing with '%' in the encoding.

Examples:

• https://www.sanskrit-lexicon.uni-koeln.de/simple/lan/bhagnāśa
• https://www.sanskrit-lexicon.uni-koeln.de/simple/lan/भद्र

[gasyoun]

Use of Devanagari (or IAST) in url now works properly.

Perfect.
Can I ask again for

https://www.sanskrit-lexicon.uni-koeln.de/s/lan/bhagnāśa

instead of:

https://www.sanskrit-lexicon.uni-koeln.de/simple/lan/bhagnāśa

[funderburkjim]

I prefer to keep 'simple' only.

• 'simple' is intuitive -- 's' is not intuitive
• to make 's' an additional alternative to 'simple' would require
  • adding a line to .htaccess
  • modify php parsing of url which currently happens in
    • list-0.2s_rw.php for cologne
    • list-0.2s_xampp_rw.php for local installations
  • code would need to be refactored to avoid duplication, and
    then modified in regard to the url parsing itself.

Don't want to do this now. You can make a separate 'enhancement' issue request if this detail is important to you.

[gasyoun]

'simple' is intuitive -- 's' is not intuitive

It's 5 letters shorter. Longer URLs break.

You can make a separate 'enhancement' issue request if this detail is important to you.

It is. Because it could become the default way of quoting Cologne URLs.

[funderburkjim]

Revised simple-search algorithm. It is now much quicker, although I think it always will provide the same answers.

In particular, the 'long' word example now is quite speedy:

https://www.sanskrit-lexicon.uni-koeln.de/simple/lan/pratyakzadarSana

[gasyoun]

In particular, the 'long' word example now is quite speedy

So now we can make the simple URLs public?

[gasyoun]

In book there is a white tab before, guess we can replicate that, @funderburkjim

[funderburkjim]

Given the current markup of lan.txt, this would be fairly difficult.

Currently, the things that look like paragraphs are, in lan.txt, preceded by an empty div: <div n="1"/>.

So consecutive paragraphs look like

<div n="1"/>blah1 blah1 blah1
<div  n="2"/>blah2 blah2 blah2
<div n="1"/>blah3 blah3 blah3

We would have to change these to

<div n="1">blah1 blah1 blah1</div>
<div  n="2">blah2 blah2 blah2</div>
<div n="1">blah3 blah3 blah3</div>

and then we could add css text-indent for these divs

div {
  text-indent: 50px;
}

The hard part is closing the divs. The example above is over-simplified, as the 'blah...' part is more complicated.

Current opinion: Could be done along the lines just mentioned, but not worth the trouble.

[gasyoun]

Current opinion: Could be done along the lines just mentioned, but not worth the trouble.

Agree, thanks for the detailed layout analysis.