Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    yonatanLehman
    @yonatanLehman
    Its starting to work now. I'm fidning the UI vry bugy - is this a new UI ?
    Jaroslav Vaňkát
    @JVankat
    @yonatanLehman great! What exactly do you find buggy (except of the root url issue we can't reproduce anymore)? We're still working on the new control panel, so all feedback is appreaciated.
    yonatanLehman
    @yonatanLehman
    I'll try and collect it and report later - but it's a pity that I have to deal with the bugs while I'm evaluating the tool...... How do I control the test snippet that is displayed with the result ? I want to be abled to 1) specitfy which text shoudl be displayed (e.g. give a div ID) and ideally to add some HTML/CSS to style it. This needs to be page dependent (e.g API pages give different previes than articles)
    yonatanLehman
    @yonatanLehman
    My API pages can be filtered based on two critea language (java,javascript, c# etc) and platform (selenium, appium, imagbes, etc) - I defined search data paoints that picked out this informatio from the URL. When I select say a language (java) it hides some options (other languages - good) - but it doesn't show all the independent options (platforms) some are missing - maybe its my problem maybe its a bug - how can I debug this?
    Jaroslav Vaňkát
    @JVankat

    @yonatanLehman the search snippet can be defined under Search Settings > Search Snippet (with the Use content behind search snippet XPath) setting (but we only take the content of the element, you can't add some html,...).

    Regarding the platlang filter, the behavior depends on the selected logic, currently the Results must match all selected values option is selected, meaning that only options that are still available after selecting the filter (e.g. java) will be displayed.

    yonatanLehman
    @yonatanLehman
    It's finding data points that I removed (even though I reindexed) and when I select the first data point its not showing the second set of data points that match it - how can I see all the urls that the search indexed (maybe I reached the limit and it didn't see the urls I'm expecting to see).
    yonatanLehman
    @yonatanLehman
    Hi Jarsoslav. Still stuck... Here is my setup now. Our URLs have the form such as "https://applitools-docs-review1.herokuapp.com/docs/api/eyes-sdk/classes-gen/class_eyes/method-eyes-eyes-selenium-java.html". I defined one data point "lang" that defines multiple patterns for the language e.g. /api/eyes-sdk/.-(java). and /api/eyes-sdk/.-(javascript). and /api/eyes-sdk/.-(python). I defined a second data point "platform" that picks out the platform, e.g. /api/eyes-sdk/.-(appium)- and /api/eyes-sdk/.*-(selenium)- I then defined a multiple choice filter that is based on these two data sources with "Results must match all selected values selected". I reindexed and the filter says "0 unique values extracted". If I understand correctly by defining multiple URL patterns for a data source I am saying that it should pick up all occurrences of the pattern. So since each URL has two parts ...platform-language.html once data source picks up the platform and the 2nd the language. What am I doing wrong?
    yonatanLehman
    @yonatanLehman
    OK solved. I define one data point that selects the language and define a filter that uses that. I define a second data point that selects the language and define a filter that selects that. Now I have two different filters and (it seems)that if you select from both filters it only shows results that match both filters. Looks like mission accomplished - The documentation is not very clear on this....
    Olesya
    @lesyaelf_gitlab
    Hi @yonatanLehman, you're currently on the Holmes plan and it's limited to 3K pages. We've indexed a bit more than that but not all URLs that are present in your sitemap. You can check what's indexed in the Index. I can see the filter values extracted though. Could you rather give a sample query and describe the result/filter behavior you'd expect? Can it be indeed cause by some entries missing from the index due to the page limit reached?
    Looks like I was just a bit late with my previous message :) Yes, now I see that you needed to define two separate filters with two separate data points as sources. We'll keep that in mind to improve the docs, thanks for the feedback!
    yonatanLehman
    @yonatanLehman
    What does it mean when a filter relies on two data sources. What is the difference between defining a single datasource with multiple patterns and a filter that depends on multiple data sources?
    Olesya
    @lesyaelf_gitlab
    I guess the confusion comes from how the logic is described: "Results must match all selected values within every filter source (data point)". If more than one data point (with all their values combined) define the same facet/filter, then you can basically pull them in together and merge into a single filter.
    So you define a datasource (data point) first, and it can have one or multiple patterns. And then in turn, one filter can rely on multiple data points (but doesn't have to).
    yonatanLehman
    @yonatanLehman
    OK. So my next step is to put this on my draft site for review by Product. My problem is that on the Installation page it tells me that my site is a Wordpress site. True - but not quite - our main site (applitools.com) is indeed a Wordpress site and already has sitesearch - I am working on the search for my part of the site (applitools.com/docs) - and I need to set this up on my Heroku site (see the URL above) which is not Wordpress. I then need to workout how to combine the main site search with my docs - that's a separate issue I'll discuss next
    Olesya
    @lesyaelf_gitlab
    Right, I see - we tried to simplify it for not-so-techy WordPress users that we also have as customers and overlooked more complex cases, it seems. So the bare minimum code snippet you'd need would be this one, described in the Installation article:
    <script>
       /* Create a configuration object */
       var ss360Config = {
          /* Your Site ID that usually equals the domain you signed up with */
          siteId: 'applitools.com_1',
          /* A CSS selector that points to your search box */
          searchBox: {selector: '#searchBox'},
       };
    </script>
    <script src="https://cdn.sitesearch360.com/v13/sitesearch360-v13.min.js" async></script>
    All additional configuration parameters are listed on the same page (under Advanced configuration). A shortcut would be using our interactive Search Designer to customize the settings and get a code snippet at the bottom
    Olesya
    @lesyaelf_gitlab

    Combining multiple domains in one search comes down to providing multiple sitemaps and/or multiple URLs to crawl under Crawler settings. To break them down (so you can show results from one domain and hide results from the other, for example), you can use Result Groups.

    To show/hide groups you can either adjust your ss360Config code like this (group names must match the names you've defined, this is just an example):

    var ss360Config = { contentGroups: { include: ['Docs','Heroku'] } }

    Or you can add data-ss360-include or data-ss360-exclude attributes to your search box HTML markup:
    <input type="search" data-ss360-include="['Docs','Heroku']">

    yonatanLehman
    @yonatanLehman
    NIce answers! thanks
    Olesya
    @lesyaelf_gitlab
    @yonatanLehman happy to help! Both Jaroslav and I are done for today but if you have further questions, feel free to leave a message here or email us, and we'll answer as soon as we can tomorrow.
    yonatanLehman
    @yonatanLehman
    image.png
    Hi. Where in the Control panel do I see the filter ID? In https://docs.sitesearch360.com/filters it is mentioned but I can't see where the image shown is in the control panel. The edit filter pages looks like this
    Jaroslav Vaňkát
    @JVankat
    Hi @yonatanLehman - this seems to be currently missing in the new interface, we'll take care of adding the ids. The id of your language filter is fid#5 and the platform is fid#6.
    yonatanLehman
    @yonatanLehman
    Thanks.
    Bong-Chu Papke
    @Bongmaster407
    Hi @JVankat and thank you for your response. Just one more question... maybe ;) isn't it possible to limit the results for Smart 404 to any number, e.g. 4? It seems that ss360Config.results.moreResultsPagingSize doesn't affect Smart 404.
    Jaroslav Vaňkát
    @JVankat
    @Bongmaster407 we could make it configurable, let me check whether we can squeeze it into this week, I'll get back to you later today.
    Bong-Chu Papke
    @Bongmaster407
    @JVankat - our client just doesn't want to overload their 404 page with too many Smart 404 suggestions, you know? They would like to have 3 or 4 suggestions and a button to display more if available. But please, I don't want to rush you. For them it's fine to release Smart 404 on their site with 12 suggestions, that's not a problem. We can adapt when you're fine with it. Maybe it would make sense to make it configurable similar to the site search, e.g. max. number of suggestions and paging size, maybe even more.
    Ah, and you removed filters and sorting from Smart 404, right? So, can I safely remove the CSS to hide them?
    Jaroslav Vaňkát
    @JVankat
    @Bongmaster407 yes, I did remove those two, removing the CSS should be safe. I think we should be able to make this configurable until Friday, will keep you posted on this.
    Bong-Chu Papke
    @Bongmaster407
    @JVankat
    Perfect, thanks a lot for both, filters and sorting, and Smart 404. I'd be happy to hear from you soon. As I said, don't rush the changes on Smart 404. It's really a cool feature. Cheers!
    Zhang Weixiang
    @that_wei_twitter
    Is there a way to prioritize the indexing for Header and then keyword density?
    Bong-Chu Papke
    @Bongmaster407
    @JVankat - a big thank you from our client, your support is highly appreciated!
    Az Troenkrasnow
    @troenkrasnow_twitter
    Will the features in the new update be available to Lightspeed users?
    Jaroslav Vaňkát
    @JVankat
    @troenkrasnow_twitter we are currently working on some last tweaks, we should be able to switch Lightspeed customers within the next two weeks.
    Olesya
    @lesyaelf_gitlab
    @that_wei_twitter it's the result title that gets prioritized the most and you can define what content exactly is used for your search result titles with the Title XPaths, e.g., //h1 for H1 headings, or //h2[@class='main-heading'] for H2 tags with the CSS class "main-heading".
    Keyword density is also taken into account but has lower priority.
    We are looking into the examples you've sent via email and will get back to you with more details shortly.
    Bong-Chu Papke
    @Bongmaster407
    Hi @JVankat - I wanted to ask if you have any news regarding the Smart 404 configuration possibilities we talked about (paging size, maybe max. number of suggestion).
    Jaroslav Vaňkát
    @JVankat
    @Bongmaster407 I've just deployed the changes, they are available in v13.3.41 and the setting is ss360Config.smart404.num (max 12). There's no pagination setting, so the pagination and max. number of suggestions are the same.
    Bong-Chu Papke
    @Bongmaster407
    @JVankat - thanks a lot. I'll take a look into it a bit later today as I'm currently working on something else.
    Bong-Chu Papke
    @Bongmaster407
    Hi @JVankat - unfortunately the changes for Smart 404 are not exactly what we expected. The additional config to limit the number of suggestions is useful after all. But we hope to be able to limit the number of visible suggestions, like in the search results, get like 4 suggestions and a button to show more if available. Does this make sense for Smart 404 from your perspective? For our client it would in fact make a lot of sense.
    Jaroslav Vaňkát
    @JVankat
    Hey @Bongmaster407 - okay, I guess there was some misunderstanding. So you would like to load e.g. 12 results, if avaiable, but would like to display the first 4 and hide the rest behind a more results button, is that correct?
    Bong-Chu Papke
    @Bongmaster407
    That's correct, @JVankat - maybe really with a config option like for the search results (ss360Config.results.moreResultsPagingSize).
    Jaroslav Vaňkát
    @JVankat
    @Bongmaster407 got it, let me try to take a look later this week, this won't be that easy to implement. Will keep you posted.
    Bong-Chu Papke
    @Bongmaster407
    Thank you, @JVankat, looking forward to hearing from you.
    yonatanLehman
    @yonatanLehman
    Hi. Suddenly my search results don't show any text snippets - even in the control panel search preview - yesterday it was working fine. I'm using "Use content behind search snippet XPath" with a rather long XPATH which is designed to more or less capture all text (//div[@id='mc-main-content']//p|//div[@id='mc-main-content']//li|//div[@id='mc-main-content']//dt|//div[@id='mc-main-content']//dd|//div[@id='mc-main-content']//div|//div[@id='mc-main-content']//a|//div[@class='mpld-descr']/p|//div[@id='mc-main-content']//h4|//div[@id='mc-main-content']//h3|//div[@id='mc-main-content']//h2|//div[@id='mc-main-content']//h1|//div[@id='mc-main-content']//meta[@name="description"]/@content) - any clues why this might be happening ?
    I did try a smaller XPATH an reindexing but it didn't help
    yonatanLehman
    @yonatanLehman
    I changed the XPATH to something simpler '//div[@id='mc-main-content']//*' and it started working - although I need to narrow it down a bit.Any tips on a good XPATH for general documentation?
    yonatanLehman
    @yonatanLehman
    Can someone tell me how I can set the snippet XPATH to express "all textual elements under div with class "C" excluding anything under div with class "D" or div with class "E". For context extraction this is easy because there are multiple XPATHs and both include and exclude but for results snippets it all needs to be expressed as single XPATH
    Jaroslav Vaňkát
    @JVankat
    @yonatanLehman maybe merging multiple xpaths with the pipe operator would be a simpler solution? For example //body[@class='docs']//div[@class='description'] | //body[@class='blog']//div[@class='excerpt']
    yonatanLehman
    @yonatanLehman
    It's easier to say "everything except" because I might miss something and not notice - it's easy to spot something and eliminate then test 400 pages of content and ensure everything is indexed. - Is there no way to say "except" in XPATH ? Is there a reason why context extraction has include and exclude xpath and snippet doesn't ?
    Jaroslav Vaňkát
    @JVankat
    @yonatanLehman do you maybe have a more specific example?
    yonatanLehman
    @yonatanLehman
    Sure. I need to capture all the text under the div with the class with id="mc-main-content" which appears on all my pages. But for example on the page https://applitools-docs-review1.herokuapp.com/docs/api/eyes-sdk/index-gen/class-eyes-espresso-java.html I don't want to see the language menu (under the h1 header) in the results snippet - this is a div with class="mpld-select-language"
    At the moment the snippet looks like this "Eyes method Class: EyesPlatform: Selenium 3Language: Java SDK: Latest other languages This is the constructor for the Eyes class. This is the main class used for executing checkpoints using Eyes. Syntax Eyes eyes = new Eyes(); Eyes eyes =…" - it should start with "This is the constructor" - the stuff before that should not be there