Re: [MeeGo-dev] Can not tracker *.rm/rmvb files
From the tracker maillist,I know that you are working on this issue. Can you give me more hint how to patch tracker-extract to associate the application/vnd.rn-realmediamimetype with gstreamer extractor Current I can only tracker realmedia video by MIMETYPE. ps: I've installed the gst-plugins-ugly. See http://meego.gitorious.org/tracker/tracker/blobs/harmattan/src/tracker-extract/tracker-extract-gstreamer.c#line152 Not sure why that is not upstream anyway. -- Aleksander ___ MeeGo-dev mailing list MeeGo-dev@meego.com http://lists.meego.com/listinfo/meego-dev http://wiki.meego.com/Mailing_list_guidelines
Re: [MeeGo-dev] Can not tracker *.rm/rmvb files
I encounters a confusing problem that the *.rm/rmvb video files can not be trackered by the following SQL SELECT nie:url(nie:isStoredAs(?video)) WHERE{?video a nmm:Video}; I analysed the /usr/share/mime/packages/freedesktop.org.xml that the mime type [mime-type type=application/vnd.rn-realmedia]is defined. Also,I found only video/* mime type is trackered in tracker-extract-gstreamer.c,is that right? So,does the qttracker only tracker video files which obtain mime type [video/*]? Can anyone give me some suggestion? Yes. Please fill a bug in MeeGo bugzilla: https://bugs.meego.com/enter_bug.cgi Product: OS Middleware Component: Content framework (tracker) And specially, attach a sample test file. Possibly it's just adding a new mimetype to the list of mimetypes extract-able by the gstreamer plugin. Cheers, -- Aleksander ___ MeeGo-dev mailing list MeeGo-dev@meego.com http://lists.meego.com/listinfo/meego-dev http://wiki.meego.com/Mailing_list_guidelines
Re: [MeeGo-dev] Does Meego well support internationalization/localization?
In Tracker we will be using ICU directly to support Chinese Pinyin collations. We'll be getting /meegotouch/i18n/lc_collate from gconf and pass it directly to libicu. This work is already on-going and will be available in the next 2-3 weeks (currently we're assuming en_US.utf8 locale always in collation). Thanks for update. For some app, locale is temporary changed, not saved to gconf. Does tracker have interface to query with a given collation from input? Not right now. There's a branch which implements a COLLATE keyword in SPARQL, but wasn't merged because we didn't find it useful at that time, we preferred to have a single collation implicitly stated in the text columns inside the database, as that would give us the proper orders already in the indexes of the indexed properties. Anyway, if there are real use cases of temporary locales applied in applications, or just different locale needs for different collations, we may end up doing it. Tracker's FTS, if enabled in meego (which I'm not really sure if it is), should already support it. We currently use libicu word-breaking algorithm to properly separate words during FTS parsing. FTS is enabled in MeeGo Tracker. But it uses glib as the default parser. I will submit a patch to enable libicu as default. Oh please, yes, enable libicu. BTW: For the input of tracker-search-tool, its length (g_utf8_strlen) must = 3. But the length of lots lots of Chinese words is 2. Can we make it configurable for different languages? It actually should be configurable in the tracker-fts.cfg configuration file, isn't it? [Indexing] # Set the minimum length of words to index (0-30, default=3) MinWordLength=3 Oh, maybe you mean directly in the t-s-t text input? I'd suggest you drop an email to tracker mailing list about this or fill a bug in bugzilla. We're currently replacing t-s-t with a new tool called needle, still not in the master branch. -- Aleksander ___ MeeGo-dev mailing list MeeGo-dev@meego.com http://lists.meego.com/listinfo/meego-dev
[MeeGo-dev] Does Meego well support internationalization/localization?
Hi Zhenqiang, - are there any libraries available helping easily implement Chinese localization features like Chinese search, sort using Chinese Pinyin? 1 Sort Current Meego content framework does not support Chinese Pinyin sort. But if your order is the same as ICU, it is easy to implement it in app. For handset applications based on libmeegotouch: MLocale supports two kind of Chinese sortings. zh_CN at collation=pinyin sorts Chinese according to the pinyin phonetics, zh_CN at collation=stroke sorts Chinese according to the stroke count of the characters. In Tracker we will be using ICU directly to support Chinese Pinyin collations. We'll be getting /meegotouch/i18n/lc_collate from gconf and pass it directly to libicu. This work is already on-going and will be available in the next 2-3 weeks (currently we're assuming en_US.utf8 locale always in collation). 2 Search (1) Big efforts are required to support search bases on Chinese word. e.g. which text files have word 搜索? For CJK languages, there is no ' ' to separate words in a sentence, some word segmentation libraries are required. Tracker's FTS, if enabled in meego (which I'm not really sure if it is), should already support it. We currently use libicu word-breaking algorithm to properly separate words during FTS parsing. (3) If you want to search Chinese with Pinyin, application should translate Chinese to Pinyin and save the maps. This is currently not supported in Tracker when doing FTS queries. This is, we do not convert all Chinese text to Pinyin during FTS automatically. But, if not using FTS and looking for exact matches of property values, it could probably work just if the proper pinyin collation is enabled (didn't test that anyway). If applications are the ones inserting both Chinese and Pinyin strings in full-text-indexed properties in tracker, FTS searches should just work properly right now. Cheers, -- Aleksander ___ MeeGo-dev mailing list MeeGo-dev@meego.com http://lists.meego.com/listinfo/meego-dev