Selection and batch operations in search#1447
Draft
alethiophile wants to merge 45 commits intoDifegue:devfrom
Draft
Selection and batch operations in search#1447alethiophile wants to merge 45 commits intoDifegue:devfrom
alethiophile wants to merge 45 commits intoDifegue:devfrom
Conversation
If the grouptanks parameter is true, then search results may include a mix of archives and tankoubons. Tankoubons have different metadata handling than archives, and so the existing sorting code didn't handle it correctly. There's a question here about how sorting by a prefix should be handled. In some places, tankoubons are considered to have a tagset equal to the union of all their constituent archives. This is sensible for bare tags, but problematic for metadata tags with prefixes. In particular, if a tankoubon has multiple tags prefixed "artist" etc, then a search sorting on the artist prefix is ambiguous. (This problem also applies to archives that simply have more than one artist: tag in their native tagset.) Currently, the handling of this case for archives is that the system picks an arbitrary one of the matching tags to use for the sort key. (Specifically, it's the first one in the serialized tag string.) Tag string serialization has no defined order, so which tag is selected will depend on the source of the tag and the order in which operations are executed. It's not clear how this problem is correctly handled, in the general case. For this particular bug fix, I just punt: when sorting by a tag prefix, tankoubons use only the tank's native set of tags, ignoring any tags on the constituent archives.
this just uses the thumbnail of the first archive in the tankoubon
This route is used to view tankoubons. In the future, it will display all tank metadata, allow editing that metadata, and allow reordering/removing/adding archives to the tankoubon. This commit just views the current contents. This does not use DataTables, but builds a similar UI manually. All tankoubon info is fetched to the client, and UI handling done client-side. Some functions from index.js and index_datatables.js have been moved to common.js in order to make them available here. Tankoubon results in searches are updated to link to this viewing page, instead of the reader mode (which won't function with tank IDs anyway).
This adds Sortable.js to the frontend dependencies, then uses it to enable manually reordering archive entries from the tankoubon view.
previously, the sortedset results of a tankoubon query were ordered in the result array by a string compare, not a numeric compare. this would order results as [1, 10, 11, 12, 2, 3, 4...] rather than the correct [1, 2, 3, 4, 10, 11, 12]. This caused the order to display incorrectly on view. It also caused instability when editing the order of tankoubon entries, because the frontend would fetch the incorrect order, then set the new order to be the incorrect one. Round-tripping the order via GET -> PUT would therefore shuffle the entries.
This changes the data model in the Redis mock to model sorted sets properly, adds a few missing methods, and adds some new data necessary for full tankoubon testing.
Bug the first: when a tank with a mixed-case title was created, it didn't appear in title searches, because the title was stored in original case but searches coerce to lower case for case-insensitivity. Fix: store the title in lower case, like archives are. Bug the second: when a tank was deleted, archives in it didn't appear in tankgrouped searches despite no longer being in a tank, because they weren't added back to the LRR_TANKGROUPED index. Fix: add them to that index.
previously, tankoubon tags in search results were created from the union of all member tags, without including the tank's own tags and without deduplicating this includes the tank's tags and deduplicates the whole set
A tankoubon has both its own tags on the tank itself, and imputed tags from the union of the tags on the contained archives. Tag searches should use both tagsets. This requires updating the tag index sets for the imputed tags on a given tankoubon, after any operation that could change those imputed tags — i.e. any operation that changes the set of archives in a tankoubon, or changes the set of tags on any archive in a tankoubon. The new utility function update_tank_imputed_indexes handles this. update_tank_imputed_indexes is called in update_archive_list, add_to_tankoubon, remove_from_tankoubon, and set_tags. Note that for the last of these, it must call it for every tankoubon that contains a given archive; get_tankoubons_containing_archive is used for this. There's potentially a performance issue, because that function is O(n) in the number of tankoubons in the DB. I have deferred optimizations until the issue is actually demonstrated. Note that the $removed_tags parameter to update_tank_imputed_indexes should contain every tag that this operation theoretically might have removed from the tank. The function will check all these against the actual new full tagset before removing any.
this is redundant with the archive get_categories endpoint, which works for tankoubons as well
This function was present in the original commit (9a746e86) but was silently dropped during a cherry-pick. The call site in common.js (handleContextMenu delete case for TANK_ IDs) references it.
alethiophile
force-pushed
the
search-select-batch
branch
from
b1c594a to
36b3649
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This is my UI for doing batch operations on archives from the search page. Currently, this is only used for adding archives to tankoubons, but I intend to add a few more operations on tags and categories.
Each result in a search has a checkbox. When checkboxes are selected, a bar pops up allowing you to do bulk operations on the selection. (Basic pattern inspired by similar operations in file managers.) The selection is maintained across page navigation. Because this is in the search result view, it gives you an easy way to operate on the results of a search: just run the search and select everything.
It's based against my tankoubon-search-and-view tree, since it's intended to work with that code, but is technically separable.