version 1.17, 2004/04/03 17:19:33
|
version 1.20, 2004/08/16 22:34:04
|
Line 155 or from a remote location
|
Line 155 or from a remote location
|
bash; export CVS_RSH=ssh; cvs -d :ext:myusername@141.14.236.86:/perseus/cvsroot co /perseus/cvsroot/mpitexts/perl/perllib |
bash; export CVS_RSH=ssh; cvs -d :ext:myusername@141.14.236.86:/perseus/cvsroot co /perseus/cvsroot/mpitexts/perl/perllib |
\end{verbatim} |
\end{verbatim} |
|
|
\subsubsection{Indexing} |
\input{soft-search} |
\label{sec:indexing} |
|
|
|
\paragraph{Status quo ECHO} |
|
Currently indexing is not implemented on the ECHO server. |
|
|
|
\paragraph{Plan ECHO} |
|
|
|
\begin{enumerate} |
|
\item construct remote (141.14.236.86) index for each file at |
|
per-change or daily intervals |
|
\item store indices locally in |
|
\url{archimedes/data/db/PROJECT_NAME/CORPUS_NAME/WORK} |
|
\item 2 progs on server 1. cgi: \url{indexer} 2. backend \url{da_remote} |
|
\item 2 progs on client 1. cgi: \url{sendindex} 2. backend \url{getindex} |
|
\item indexing transaction handled by two cgi scripts, one on the |
|
server the other on the client [this is the 1st implementation bcs |
|
its easiest and there are no port issues, but probably it'd be |
|
better to have a separate port]. |
|
\item client cgi: getindex -- sends 1. list of files to index |
|
2. uri to which xml notification of completion is to be sent. Upon |
|
notification, activates backend prog that fetches and installs the |
|
indices. |
|
\item server cgi: indexer receives filelist and notification |
|
addess. Activates backend that fetches files, indexes, places |
|
completed indexes in a networked location, then sends xml |
|
notification back to client. |
|
\item single script provides backend access to indices |
|
\item leave front-end issues like display, collection and navigation |
|
to web-design programmers. Do only a sample for now. |
|
\end{enumerate} |
|
|
|
\subsubsection{Morphology} |
\subsubsection{Morphology} |
\label{sec:morphology} |
\label{sec:morphology} |
Line 215 No parameter--update all lemmatization i
|
Line 185 No parameter--update all lemmatization i
|
\paragraph{makefast.pl ARCHIMEDES} |
\paragraph{makefast.pl ARCHIMEDES} |
|
|
Updates the toc.cgi morphology indices |
Updates the toc.cgi morphology indices |
Parameters |
Parameters: |
No parameter--update all lemmatization indices |
No parameter--update all lemmatization indices |
[latin | ital | greek | en | nl | de]-- update this language |
[latin | ital | greek | en | nl | de]-- update this language |
|
|
\subsubsection{summary of differences btwn the archimedes toc.cgi |
The indices are produced from the corpus word index 'xml:raw:norm', |
implementation and the echo toc.cgi impelementation (toc.x.cgi)} |
which correlates raw forms to normalized forms, and |
|
'\$lang:inc_lemma', which correlates incidentia to lemmata. The basic |
|
rule is, if exists \$raw->\$norm->\$inc_lemma, then \$raw is included |
|
in the 'fast' index for that language. |
|
|
|
Currently stores the indices with the name xml:hit:\$lang, where |
|
\$lang is one of [ital,greek,latin,de,en,fr,nl] in the directory |
|
/usr/share/perlobjects/wordindex in Archim::Object::Depot format |
|
(Storable). Access to these indices is provided by |
|
Archim::Toc::Utils->get_hits_hash(\$lang) . |
|
|
|
The functionality of makefast.pl is duplicated by Archim::Toc::Index->make_fast_lemma(\$lang); |
|
|
|
|
|
\subsubsection{summary of differences btwn the archimedes toc.cgi implementation and the echo toc.cgi impelementation (toc.x.cgi)} |
|
|
\paragraph{missing in archimedes} |
\paragraph{missing in archimedes} |
\begin{enumerate} |
\begin{enumerate} |
Line 235 No parameter--update all lemmatization i
|
Line 219 No parameter--update all lemmatization i
|
\item remote text method may work differently |
\item remote text method may work differently |
|
|
|
|
|
|
\end{enumerate} |
\end{enumerate} |
\paragraph{differences} |
\paragraph{differences} |
\begin{enumerate} |
\begin{enumerate} |