Re: [xwiki-devs] Map Application - GSoC 19

That is much like the default view of popups in Leaflet. This kind of popup supports very little information, that's why I made a dedicated space for popups. However, we could go with your suggestion of "view more". We could either open the parent page with "view more" or fill the search widget as you suggested. I would go with displaying more information in the search widget. Is that fine with you? Along this line, another improvement (you probably have it in mind) would I do not think this is required. If a developer wants a custom display for the popup information, he can create a class sheet and make pages with that sheet and it will become a map item after adding location object to the page. Ok, we need to investigate this. I have a preliminary question about this I did not go with this flow because of better performance since a separate async request will be made for change in each state. What I am doing now is that I take the map state on search or other events that reload the map asynchronously.Thanks for your suggestion Stephane. I could update the page by observing a change in each state. This is a little slow because the map will have to be reloaded for each state but still a good option. Best, Fawad On Thu, Jul 11, 2019 at 3:26 PM Stéphane Laurière &lt;slauriere(a)xwiki.com&gt; wrote: > Hi Fawad, Hi all, > > > Hi all, > > > > However, as a user I would expect the list to be present also when > the search input is empty, with the possibility to reduce the list by > hitting the link "Show result items", what do you think? > > > > > > By this do you mean that all the results should show in the "Show result > items" when the search is empty? Thereby decreasing the number of result > items when an actual search is made. > > Yes indeed, the number of items would then decrease either when a text > search is made or a facet filter is checked. > > > Speaking of the list, that'd be great imho that the popups can be > displayed while the list is active, so that the user can browse content via > the list, wouldn't it? > > > > > > I am not too sure where we could place them, since both the list and > popups take a large amount of space. The popups are build to support page > data, so they are also big. We could do it so the upper space belongs to > the popup and the lower space belongs to the list. Though it will cramp the > space a little imo. We could also increase the overall height of the map > (which is 400px for now). WDYT? > > < > https://www.airbnb.com/s/soweto/homes?refinement_paths%5B%5D=%2Fhomes&q… > > > > The ability to highlight a marker when hovering the list is also quite > handy. > > (note also the "search as I move the map feature") > > Imho the popups could be reasonably small by default, and later on we may > consider various improvements such as 1) the ability to increase the popup > side dynamically like what happens on the map below when you hit the link > "Plus d'information" at the bottom of the panel: > http://carte.preference-commerce.fr/cci-fr/ 2) the ability to choose > where this contextual information should show up. For instance, for users > having large content to get displayed, it could make sense to let the > content replace entirely the list one, just like what Google does in this > example (which is close to what we have at the moment, except that I > wouldn't hide the search input, and I would add a "Back to results" link > just like what Google does, and keep the same container so that the user > understands better that the two panels are complementary, I find it easier > to grasp the underlying behaviour): > > < > https://www.google.com/maps/place/Ch%C3%A2teau+De+Monteton/@44.6171828,0.13… > > > > Along this line, another improvement (you probably have it in mind) would > > Restoring the full screen is working in most cases, that's cool, > however if I run a search, then enter full screen, then refine the search, > it seems that the full screen state is not preserved, is it? > > > > > > That's one use case which leads to a loss of the previous map state and > the one before that is used. Any state before the facet click is not > preserved except the search widget and "Refine your search" area which is > more or less forced. As I said before, the problem is that we cannot pass > the state anywhere when a facet is clicked. I will look into cumulative > onclick events but I think that's not going to work. > > Ok, we need to investigate this. I have a preliminary question about this > > > I agree LeafletMap would be closer to what it does, but contrarily > to the Macro Map, if I'm not mistaken, the page currently named > Maps.Code.Leaflet does not create a JavaScript LeafletMap class, so > technically it's more a LeafletUtils than a LeafletMap, isn't it? > > > > > > You are right, it does not create any js class and since css is also > inside the same page the name should be LeafletUtils or LeafletService. I > will change it accordingly. > > Ok cool. > > > Also, I looked into it and the idea of custom class property for getting > the location is interesting indeed. I will see how I should prioritize it > and implement it then since we have other major features to look into. > > > > I would also like to ask; what do you foresee as the basic elements of > the application? We have map configuration, map query, search and filter > widget, markers and popups so far. Do you have anything other than this in > mind that could be thought of as the basis of the application? I think we > should completely stabilize the basics by the end of this week so I can > work on other features like map paths, indoor maps and custom shapes. > > I'll think about it and will give you my feedback on this as soon as > possible. > > Cheers > > Stéphane > > > > Best, > > Fawad > > > > On Thu, Jul 11, 2019 at 1:50 PM Stéphane Laurière &lt;slauriere(a)xwiki.com > <mailto:slauriere@xwiki.com>> wrote: > > > > Hi Fawad, all, > > > > > I have fixed the errors you mentioned. It appears one of the > variables had a bad name which was causing the error. > > The error does not show anymore indeed on my end, thanks. > > > > > On the UI side, possibly as a consequence, I can't see a list > of results anymore, and the search widget state is not restored. > > > > > > > > > The list of results is just fine on my end, I am not sure what > could be wrong. > > > > In the meantime, I figured out that the list shows up indeed when > issuing a query in the search input, cool. However, as a user I would > expect the list to be present also when the search input is empty, with the > possibility to reduce the list by hitting the link "Show result items", > what do you think? Speaking of the list, that'd be great imho that the > popups can be displayed while the list is active, so that the user can > browse content via the list, wouldn't it? This should not prevent us from > release version 1.0 though. > > > > > About the map state, it does not work well with facets. Since > facets have a separate code we cannot apply custom code when a facet is > selected thus limiting our ability to pass the map state through js. I > tried looking around for related js events but could not find one through > which we can pass the map state after a facet is clicked. Do you have > anything in mind for this? > > > > Actually, with the new version, the facet state is restored as I > would expect it, sorry for the confusion. Restoring the full screen is > working in most cases, that's cool, however if I run a search, then enter > full screen, then refine the search, it seems that the full screen state is > not preserved, is it? > > > > A note about demos: as far as I can see, the museum map demo does > not showcase the "country" facet by default, don't you think that'd be > worth adding it so that the user has a simple demo with facets working with > custom fields? > > > > > I think the name "Maps.Code.Leaflet" might be misleading for > potential developers: this could mean this page is providing Leaflet, while > it does not. I would suggest to choose a name that is closer to what the > JavaScript really provides, from a functional point of view, what do you > think? > > > > > > > > > I think its fine as is. Since it is placed inside the Code space, > the developers will immediately be able to know that all the Leaflet > related code resides inside it. But to be on the safe side, we can rename > it to LeafletMap as in the macro-map. > > > > I agree LeafletMap would be closer to what it does, but contrarily > to the Macro Map, if I'm not mistaken, the page currently named > Maps.Code.Leaflet does not create a JavaScript LeafletMap class, so > technically it's more a LeafletUtils than a LeafletMap, isn't it? I > acknowledge I'm picky with names... I noticed that leaflet-main requires > leaflet-commons but I'm actually wondering how to make sure that > leaflet-commons is known from RequireJS before leaflet-main gets loaded. I > have not faced any error in practice, but I'm wondering if you tackled the > issue already or if we need to figure it out. > > > > > Ludovic just suggested an improvement (for the next > versions): let the user configure which existing field could be used in an > existing class for retrieving geographical information, that could be > interesting indeed, to be discussed. The calendar application works this > way already as far as I understood: it lets the user define the date / time > field to be used. > > > > > > > > > Thanks Ludovic for your suggestion. I would look into the > calendar application to have a better understanding of this and let you > know my thoughts. > > > > Ok cool. > > > > Cheers > > > > Stéphane > > > > > > > Best, > > > Fawad > > > > > > > > > On Wed, Jul 10, 2019 at 2:55 PM Stéphane Laurière < > slauriere(a)xwiki.com <mailto:slauriere@xwiki.com> <mailto: > slauriere(a)xwiki.com <mailto:slauriere@xwiki.com>>> wrote: > > > > > > Hi Fawad, > > > > > > Good to hear from you again, I hope things are fine on your > end as well. Thanks for the update. Sorry for the delay, we were traveling > yesterday. Releasing the application soon sounds good. I'm facing a few > issues though, they may be related to an installation issue on my side, not > sure. I grabbed the latest code and imported as a XAR over the existing > pages in my 11.x wiki without error, and I notice the following (I'll > consider posting some Jira issues if needed): > > > > > > - catalina.out errors (not sure if they were present with > previous version): > > > > > > 2019-07-10 11:34:30,349 [ > http://ako:8090/xwiki-11.2/wiki/abraca/jsx/Maps/Code/Leaflet?language=en&am…] > WARN c.x.x.w.s.JsExtension - Error at line 203, column 85: > missing variable name. Caused by: [ var index = 0, lat = 0, lng = 0, > coordinates = [], shift = 0, result = 0, byte = null, latitude_change, > longitude_change, factor = Math.pow(10, Number.isInteger(precision) ? > precision : 5);] > > > 2019-07-10 11:34:30,350 [ > http://ako:8090/xwiki-11.2/wiki/abraca/jsx/Maps/Code/Leaflet?language=en&am…] > WARN c.x.x.w.s.JsExtension - Error at line 206, column 13: > identifier is a reserved word. Caused [...] > > > 2019-07-10 11:35:02,841 [ > http://ako:8090/xwiki-11.2/wiki/abraca/jsx/Maps/Code/Leaflet?language=en&am…] > ERROR c.x.x.w.s.JsExtension - Runtime error minimizing JSX object: > Compilation produced 8 syntax errors. > > > 2019-07-10 11:35:02,841 [ > http://ako:8090/xwiki-11.2/wiki/abraca/jsx/Maps/Code/Leaflet?language=en&am…] > WARN c.x.x.w.s.JsExtension - Failed to compress JS extension: null > > > > > > - On the UI side, possibly as a consequence, I can't see a > list of results anymore, and the search widget state is not restored. > > > > > > - I notice there is no default radio button checked in the > search form: I think either "location" or "item" should be checked, to let > the user know what's the default (I'd say "item"). > > > > > > - I think the name "Maps.Code.Leaflet" might be misleading > for potential developers: this could mean this page is providing Leaflet, > while it does not. I would suggest to choose a name that is closer to what > the JavaScript really provides, from a functional point of view, what do > you think? > > > > > > - Ludovic just suggested an improvement (for the next > versions): let the user configure which existing field could be used in an > existing class for retrieving geographical information, that could be > interesting indeed, to be discussed. The calendar application works this > way already as far as I understood: it lets the user define the date / time > field to be used. > > > > > > Cheers > > > > > > Stéphane > > > > > > > > > Fawad Ali: > > > > Hi all, > > > > Hope everyone is well. > > > > > > > > Please review the application developed so far. I have > included a UI test and map states. I think we should release the > application as soon as we can so that user reviews can be gathered. > > > > > > > > Best, > > > > Fawad > > > > > > > > > -- > > Stéphane Laurière > > XWiki – https://xwiki.com > > > > > -- > Stéphane Laurière > XWiki – https://xwiki.com > >