From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:470:142:3::10]:41472) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jAuC7-0001kA-Jo for guix-patches@gnu.org; Sun, 08 Mar 2020 07:34:08 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1jAuC6-0007nG-DF for guix-patches@gnu.org; Sun, 08 Mar 2020 07:34:07 -0400 Received: from debbugs.gnu.org ([209.51.188.43]:42000) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1jAuC6-0007mn-A2 for guix-patches@gnu.org; Sun, 08 Mar 2020 07:34:06 -0400 Received: from Debian-debbugs by debbugs.gnu.org with local (Exim 4.84_2) (envelope-from ) id 1jAuC2-0006Vz-0l for guix-patches@gnu.org; Sun, 08 Mar 2020 07:34:02 -0400 Subject: [bug#39258] [PATCH v2 0/3] Xapian for Guix package search Resent-Message-ID: From: Ludovic =?UTF-8?Q?Court=C3=A8s?= References: <20200307133116.11443-1-arunisaac@systemreboot.net> <87sgijgb1v.fsf@gnu.org> Date: Sun, 08 Mar 2020 12:33:44 +0100 In-Reply-To: (Arun Isaac's message of "Sun, 08 Mar 2020 14:31:42 +0530") Message-ID: <875zffcc87.fsf@gnu.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: guix-patches-bounces+kyle=kyleam.com@gnu.org Sender: "Guix-patches" To: Arun Isaac Cc: mail@ambrevar.xyz, 39258@debbugs.gnu.org, zimon.toutoune@gmail.com Hi, Arun Isaac skribis: >>> It turns out that most of the time is spent in printing and texinfo >>> rendering of the search results. > > Also, when we put all package metadata into the Xapian index, we don't > have to look up any of the package variables in (gnu packages *) during > `guix search` time. This also contributes substantially to the speedup. Yup. >> In general, pre-rendering doesn=E2=80=99t seem practical to me: the outp= ut of >> =E2=80=98guix search=E2=80=99 is locale-dependent (it speaks the user=E2= =80=99s language) and > > Note that we already need to index package synopses and descriptions in > all languages. I still haven't implemented this, though. Oh, right. Tricky! >> adjusts to the terminal width (well, this is temporarily broken on >> Guile 3.0.0, but see =E2=80=98%text-width=E2=80=99 in (guix ui)). > > This could be accomplished even with pre-rendering. Xapian provides > "slots" to store arbitrary strings with a document. Instead of storing > the pre-rendered document as a whole, we could store pre-rendered fields > in separate slots. Then, during `guix search` time, we can assemble the > result from these pre-rendered fields. I=E2=80=99m not sure I understand. The index wouldn=E2=80=99t store pre-re= ndered strings for every possible terminal width, right? >> Also, if the 12K+ descriptions need to be rendered at the time the user >> runs =E2=80=98guix pull=E2=80=99, the experience may not be great, becau= se it could take >> a bit of time. > > This is a problem, but I would see it as a necessary "compilation" > step. :-P In fact, this whole patchset speeds up `guix search` by doing > part of the work of `guix search` ahead of time. So, some such cost is > unavoidable. Yeah. I think we need to take the whole user experience into account, not just =E2=80=98guix search=E2=80=99. =E2=80=98guix pull=E2=80=99 alread= y feels very slow, and it=E2=80=99s a fairly common operation. Conversely, =E2=80=98guix search=E2=80=99 takes r= oughly between 0.5 and 2 seconds and is an uncommon operation on a =E2=80=9Cslow p= ath=E2=80=9D (in the sense that when you=E2=80=99re searching for software, you=E2=80=99= ll probably have to spend more than a couple of seconds to find what you=E2=80=99re loo= king for.) >> What I like about the recutils format in this context is that it=E2=80= =99s both >> human- and machine-readable. The examples in the manual show how it can >> be useful to select the information displayed or to refine the search >> (info "(guix) Invoking guix package"). > > Xapian's query language is much more natural (as in natural language) > than the regexp based techniques we need to use with recutils. I have > hardly ever used the regexp based search and I suspect many others > haven't either. Also, refining the search query should be easier to do > with Xapian. We could even use Xapian's query expansion feature to > suggest improved queries to the user. I=E2=80=99m not sufficiently familiar with Xapian=E2=80=99s query language.= The examples I had in mind were: guix search malloc | recsel -p name,version,relevance guix search | recsel -p name -e 'license ~ "LGPL 3"' guix search crypto library | \ recsel -e '! (name ~ "^(ghc|perl|python|ruby)")' -p name,synopsis It=E2=80=99s not so much about regexps than it is about selecting individual fields. >> Were you able to measure the cost of rendering specifically? > > generate-package-search-index takes around 50 seconds. If I modify > generate-package-search-index to not pre-render but simply store the > package description alone, it takes around 20 seconds. That gives us a > rough idea of the cost of pre-rendering. To me, adding 20=E2=80=9350 seconds on =E2=80=98guix pull=E2=80=99 would be= undesirable. :-/ >> I think we should look at a profile of =E2=80=98package->recutils=E2=80= =99, there=E2=80=99s >> probably room for improvement there. > > On quick inspection, most of the time in package->recutils is spent in > texinfo rendering the description. Unless we use the simplified search > results format as discussed above, we cannot avoid it. What I meant was that we could use (statprof) to see whether/how Texinfo rendering/parsing can be optimized. Thanks, Ludo=E2=80=99.