Open neilferreira opened 8 years ago
@neilferreira Good points and relevant issues in the current implementation.
Now when you said it, it could be possible to also to render the final HTML, run that through portal transforms and index it in c.dxtextindexer adapter:
For record keeping
This will be fixed by finalising https://github.com/plone/plone.app.blocks/pull/21
https://github.com/plone/plone.app.mosaic/issues/54 is "closed" but not fixed
Aha, that's the one that I thought on https://github.com/plone/plone.app.blocks/pull/21 :-)
So @neilferreira could you give us an update on your CLA status?
CLA has been signed. I'm also part of the Plone foundation on Github.
@datakurre Did you want me to commit this code anywhere to have it merged in? https://github.com/neilferreira/plone.app.blocks/commit/57b40acbb29a84f1ca051ea16a5767160abbd237
It seems to have gotten lost in between #21 and this issue if I'm not mistaken
Please, could you do a new pull. I probably thought that Nathan's indexing pull covered this, but it was still missing portal transforms call to clean html tags.
Btw, does it make sense to index external content tiles? (Code which resolves relation value to get title). I'd also like to keep relationvalue optional dependency, similar to dxtextindexer.
On 19. huhtikuuta 2016 klo 8.55 +0300, Neil Ferreiranotifications@github.com, wrote:
@datakurre(https://github.com/datakurre)Did you want me to commit this code anywhere to have it merged in?neilferreira@57b40ac(https://github.com/neilferreira/plone.app.blocks/commit/57b40acbb29a84f1ca051ea16a5767160abbd237)
It seems to have gotten lost in between#21(https://github.com/plone/plone.app.blocks/pull/21)and this issue if I'm not mistaken
— You are receiving this because you were mentioned. Reply to this email directly orview it on GitHub(https://github.com/plone/plone.app.blocks/issues/25#issuecomment-211740845)
I believe, this is implemented.
Apologies for re-opening this for a second time but I believe I'll be re-visiting this on a project quite shortly, and will likely give it some attention then.
The current implementation still indexes the object's id in the SearchableText (unsure why) and makes the assumption that only title, label and content will be indexed on the tiles. If you use RichText fields in your tiles (which we do all the time) these are not indexed at all, making the indexing quite useless.
Todo list, which most of which is already done on https://github.com/neilferreira/plone.app.blocks/commit/57b40acbb29a84f1ca051ea16a5767160abbd237
I actually really like what @hvelarde has created on https://github.com/collective/collective.cover/pull/556/files so now I'm almost hesitant to be automatically doing any sort of schema indexing, and rather leaving it to the developer to specify an indexer for their tile themselves.
@datakurre now that you've seen these two approaches how did you want to approach this?
Enhancing the current indexers with adapter support would be nice, but instead of inventing new interfaces, I'd reuse collective.dexteritytextindexers interfaces.
On 4. lokakuuta 2016 klo 4.35 +0300, Neil Ferreira notifications@github.com, wrote:
I actually really like what @hvelarde (https://github.com/hvelarde) has created on https://github.com/collective/collective.cover/pull/556/files so now I'm almost hesitant to be automatically doing any sort of schema indexing, and rather leaving it to the developer to specify an indexer for their tile themselves.
@datakurre (https://github.com/datakurre) now that you've seen these two approaches how did you want to approach this?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub (https://github.com/plone/plone.app.blocks/issues/25#issuecomment-251273851), or mute the thread (https://github.com/notifications/unsubscribe-auth/AAJyv8IW_UKLYk0XRu-ur64r2J7Hk2oIks5qwa1YgaJpZM4Hjcn-).
Btw., if someone takes the time to do the implementation and a PLIP to merge collective.dexteritytextindexers in plone.app.dexterity this would be awesome.
Hi all
I have numerous mosaic pages that I need to have indexed for my site search. I've been following the trail and noticed that the layout content can be indexed via this package from following this trail by @datakurre : https://github.com/plone/plone.app.mosaic/issues/54 https://github.com/plone/plone.app.blocks/commit/d6cbc0a8c1689336332b61d376b3e7f205a68de8
I have followed the steps outlined in the comment on the mosaic issue, but am now getting these values indexed in my "SearchableText" index
As you can see, the full HTML is being indexed, rather than a plaintext version of the tiles being indexed. It also has the issue where the <div data-tile objects are being indexed, instead of a rendered version of them. Meaning you won't actually have the Persistent tiles's content being indexed.
A snippet of the 'value' being indexed in https://github.com/collective/collective.dexteritytextindexer/blob/master/collective/dexteritytextindexer/indexer.py#L83
Is this the planned behaviour? If I added a tile onto my page such as a "Heading" tile with the content "Hello world" I would presume my SearchableText to become ['hello', 'world']
Any pointers to resolving this? Is my expectation the same as you intended for this to work? OR should I be looking implementing this myself?
Thankyou.