|
Revision 1925, 1.5 kB
(checked in by karpet, 2 years ago)
|
verify locale should also be global
|
| Line | |
|---|
| 1 |
libswish3 |
|---|
| 2 |
============================================================================ |
|---|
| 3 |
libswish3 is a document parser compatible with the Swish-e 2.4 -S prog API. |
|---|
| 4 |
libswish3 is a C library for parsing documents into |
|---|
| 5 |
a data structure that can then be stored and searched with a variety of IR backends. |
|---|
| 6 |
|
|---|
| 7 |
The libswish3 model is simple: you define a handler function and pass |
|---|
| 8 |
a pointer to that function to whichever parser input function you choose (in-memory, |
|---|
| 9 |
stdin, or filesystem). After each document is parsed, your handler function |
|---|
| 10 |
is passed the swish_ParseData object to do with as it pleases, most likely |
|---|
| 11 |
storing the data in an index. |
|---|
| 12 |
|
|---|
| 13 |
See the swish_lint.c program for how you might write a handler function. |
|---|
| 14 |
|
|---|
| 15 |
|
|---|
| 16 |
Differences from Swish-e 2.4 parser |
|---|
| 17 |
-------------------------------------- |
|---|
| 18 |
The following Swish-e configuration options are not supported: |
|---|
| 19 |
|
|---|
| 20 |
# FileRules |
|---|
| 21 |
# FileFilter |
|---|
| 22 |
# FileFilterMatch |
|---|
| 23 |
# FileMatch |
|---|
| 24 |
# FileRules |
|---|
| 25 |
# EquivalentServer |
|---|
| 26 |
# MaxDepth |
|---|
| 27 |
# Delay |
|---|
| 28 |
# TmpDir |
|---|
| 29 |
# ReplaceRules |
|---|
| 30 |
# SpiderDirectory |
|---|
| 31 |
|
|---|
| 32 |
|
|---|
| 33 |
|
|---|
| 34 |
Getting Started |
|---|
| 35 |
----------------- |
|---|
| 36 |
|
|---|
| 37 |
See the INSTALL doc. |
|---|
| 38 |
|
|---|
| 39 |
But basically: |
|---|
| 40 |
|
|---|
| 41 |
./bootstrap (only necessary if you are doing development) |
|---|
| 42 |
./configure |
|---|
| 43 |
make && make test |
|---|
| 44 |
sudo make install |
|---|
| 45 |
|
|---|
| 46 |
|
|---|
| 47 |
|
|---|
| 48 |
Profiling with gprof |
|---|
| 49 |
--------------------------- |
|---|
| 50 |
|
|---|
| 51 |
The default Makefile.am files include the -pg flag in order |
|---|
| 52 |
to get profiling information. You should remove the -pg before |
|---|
| 53 |
compiling for a production (i.e., non-development) system. |
|---|
| 54 |
|
|---|
| 55 |
See the gprof man page. |
|---|
| 56 |
|
|---|
| 57 |
Basically: |
|---|
| 58 |
|
|---|
| 59 |
gprof .libs/swish_lint |
|---|
| 60 |
|
|---|
| 61 |
should give the 'make test' profile. |
|---|
| 62 |
|
|---|
