Обсуждение: [PATCH] pgweb: Search: bugfix: replace sites.baseurl with sites.hostname
Hello,
There is an issue in the sql files for search in pgweb, as schema.sql does not correspond to what is used in
functions.sql.
functions.sql uses sites.baseurl, which is not a field in schema.sql. I replaced it with sites.hostname, which I guess
isthe targeted field.
Without that patch, I get the following error when performing a search:
Traceback (most recent call last):
File "/srv/pgweb/local/pgweb/search/views.py", line 284, in search
'internal': include_internal,
psycopg2.errors.UndefinedColumn: column sites.baseurl does not exist
LINE 1: SELECT sites.id AS siteid, sites.baseurl, webpages.suburl, t...
^
QUERY: SELECT sites.id AS siteid, sites.baseurl, webpages.suburl, ts_rank_cd(fti,tsq) * relprio AS ts_rank_cd FROM
web
pages INNER JOIN sites ON webpages.site=sites.id WHERE fti @@ tsq AND site=1 AND (includeinternal OR NOT isinternal)
OR
DER BY ts_rank_cd(fti,tsq) * relprio DESC LIMIT 1000
CONTEXT: PL/pgSQL function site_search(text,integer,integer,text,boolean) line 21 at OPEN
Cheers,
--
Célestin Matte
Вложения
Re: [PATCH] pgweb: Search: bugfix: replace sites.baseurl with sites.hostname
От
Magnus Hagander
Дата:
On Fri, Oct 22, 2021 at 4:59 PM Célestin Matte <celestin.matte@cmatte.me> wrote:
Hello,
There is an issue in the sql files for search in pgweb, as schema.sql does not correspond to what is used in functions.sql.
functions.sql uses sites.baseurl, which is not a field in schema.sql. I replaced it with sites.hostname, which I guess is the targeted field.
Without that patch, I get the following error when performing a search:
Traceback (most recent call last):
File "/srv/pgweb/local/pgweb/search/views.py", line 284, in search
'internal': include_internal,
psycopg2.errors.UndefinedColumn: column sites.baseurl does not exist
LINE 1: SELECT sites.id AS siteid, sites.baseurl, webpages.suburl, t...
^
QUERY: SELECT sites.id AS siteid, sites.baseurl, webpages.suburl, ts_rank_cd(fti,tsq) * relprio AS ts_rank_cd FROM web
pages INNER JOIN sites ON webpages.site=sites.id WHERE fti @@ tsq AND site=1 AND (includeinternal OR NOT isinternal) OR
DER BY ts_rank_cd(fti,tsq) * relprio DESC LIMIT 1000
CONTEXT: PL/pgSQL function site_search(text,integer,integer,text,boolean) line 21 at OPEN
Is that really right/enough?
All the instances I run have both fields, and for example hostname is www.postgresql.org and baseurl is https://www.postgresql.org.
Now, today that's the only site we have, so we could just replace it with putting the https:// part into it int he templates, but I think this originally comes from the times when we indexed multiple sites in a previous system -- and it appears to have been incorrectly imported in the very first commit. The search system is really due for a cleanup/replacement since that has been removed, because a lot of things could be simplified around that.
But short term, I believe the correct thing to do is to add the baseurl field to the table in the SQL file. I'll go push a patch taht does that.
Re: [PATCH] pgweb: Search: bugfix: replace sites.baseurl with sites.hostname
От
Célestin Matte
Дата:
Hello, > But short term, I believe the correct thing to do is to add the baseurl field to the table in the SQL file. I'll go pusha patch taht does that. This broke data.sql as baseurl is NOT NULL. Attached patch fixes that. This might have also broken the crawler but I haven't installed it to test that. Cheers, -- Célestin Matte
Вложения
Re: [PATCH] pgweb: Search: bugfix: replace sites.baseurl with sites.hostname
От
Magnus Hagander
Дата:
On Mon, Oct 25, 2021 at 6:19 PM Célestin Matte <celestin.matte@cmatte.me> wrote:
Hello,
> But short term, I believe the correct thing to do is to add the baseurl field to the table in the SQL file. I'll go push a patch taht does that.
This broke data.sql as baseurl is NOT NULL. Attached patch fixes that.
This might have also broken the crawler but I haven't installed it to test that.
Indeed it does. Oops.
That said, most of that file are not-current examples, and should just be cleaned out, so I'll do that while fixing this.
Yeah, another sign that this needs some proper spring cleaning to get rid of configuration of the flexibility that we don't care about. But tihs one for now, to unbreak things short-term.