* Store all schema_org from webpages in a field
* flatten json tokenizer
* rename goggles -> optics
* update optics syntax
* cargo workspace
* very simple lsp wasm connection
* optics as separate package
* hover stuff
* optics vscode extension published
* syntax errors on-save and begin schema-field
* Use separate targets for LSP and rest (#68)
By moving the different targets into separate workspaces, we avoid some
of the issues where rust-analyzer might just stop working.
By adding the two projects to .vscode/settings.json we keep the ability
to get completions, goto definitions, rename, and such operations.
This requires us to specify the dependency versions in the LSP crate, as
we can no longer refer to them by the workspace version. The positive of
this is that the WASM/LSP dependent crates are now moved to the LSP crate.
* schema.org syntax in optic
* optic can now perform schema searches
* simplified schema_org flattening
* wrote new quickstart.optic
* update like-text
Co-authored-by: Oliver Bøving <oliver@bvng.dk>
* parse single microdata item
* parse microdata from entire website
* convert microdata to schema.org
* Refactor schema.org to have a json_ld parse module
* a lot of schema.org types
* Less types for schema.org
It would currently require waaaaaaaay too much work to define all the types for schema.org compared to the benefits we would get from having them defined.
* test with stackoverflow question and a recipe
* Parse DMOZ data
* index topics as facets
* calculate topic centrality
* fix serious bug in webgraph where some nodes dissapeared (there is still a bug somewhere, but waaaay less nodes are missing now)
* apply topic centrality during search
* Add autosuggest scrape as a separate command
* Save queries continuously
* Save images as they get downloaded (way lower memory usage)
* Created configure subcommand
* Updated justfile and setup documentation
* refactor harmonic centrality into separate centrality module
* betweenness centrality
* betweenness speedup by not using hashmaps
* [WIP] trust centrality
* more robust warc download?
* unit test for trust centrality calculation
* refactor centrality store and also save trust centrality in store
* approx harmonic centrality working
* Order trusted nodes by betweenness, if the user specified too many, merge the worse trusted nodes into some of the better and update their weight
* dislike sites
* re-enable harmonic centrality calculation
* calculate betweenness on full graph
* added nodeid to schema
* use personal centrality during search
* sort centrality values in csv files
* liked and disliked sites in goggles syntax
* Move settings into a subfolder
This ensures that the paths are layed out the same on the frontend and
on the backend.
* Format goggles and sites .astro
* Add Alpine.js and use `ServeDir` instead of `Spa`
The `ServeDir` allows us the use the public dir and remove the extra
/assets/ also solving the issue with loading scripts without inline.
Currently only a few computes are ported to use Alpine.js, but the rest
should be doable!
