* Update suggestion command for building cornac from source

* Fix install command using escape dependency with specified version

Add Comparative Aspects and Opinions Ranking for Recommendation Explanations (Companion) model (#635)

* Add Companion model

* Add example

* Update docs

* Update Companion model with additional arguments, removing sentiment loss

* Add paper link

* refactor code

* Add Cornac-AB section to Readme.md

* Update README.md

* Update assets directory and downsized images

---------

Co-authored-by: Tuan Truong <tqtg@users.noreply.github.com>

* Add dark theme to readthedocs model table

* Restructure models table into separate section

* Add filter label and changed clear filters button behavior

* Add color highlighting for selected filter

* Fix issue #628 with '.A' stuff on csr_matrix

* Modifying requirements.txt for numpy and another .A

---------

Co-authored-by: tqtg <tuantq.vnu@gmail.com>

* Added static page for model viewing and filtering

* Update model json data

* Update model js table

* Add generate model json python file for js table

* Reset filter input when clear filters button is selected

The new Recommender.rank() function adds k as required value, which breaks some models that do not use k in ranking evaluation (e.g., ComparER, EFM, LRPPM).

This commit updates .rank() for mentioned models with topK option.

Include 'models' and 'datasets' labels to auto-increment minor version.

* Init method

* Example file

* Added HypAR

Added hyper and its functions.

* Added dataset standard files and virtual env names to ignore

* Create graph comments

* Flock comments

* Graph wrapper comments

* ao embedding comments

* Remove learned ui embeddings

* Learned embedding and norm comments

* fit comments

* Removed unused code

* Comments on main file

* Removed more and comments

* More comments

* More comments

* Updated readmes

* Fix dataset name

* Reference fix

* Path fix

* Removed example

* Removed HypAR from examples readme

* Styling

* Added hypar doc string

* Removed spacing

* Added paper link

* Moved docstrings to earlier

* Simplified model type

* Added paper reference

* Added lightgcn as part of HypAR

* Updated url

* Fix double normalize of first layer

* Add git ignore

* Lightgcn fix

Removed normalization for layers, not used for lgcn.
Fixed sum weight constant to num layers instead of cur layer index.
Allow lgcn to take blocks.
Fixed requirement error caused by newer dgl versions.
Moved edge normalization to graph for easier use.

* Lightgcn debug error fix

* Simplified layer normalization and readability

* Easier support of rcuda

* Update link in working_with_auxiliary_data.md

* fixed url

---------

Co-authored-by: Quoc-Tuan Truong <tqtg@users.noreply.github.com>

Change docker registry to ghcr.io

* Include metric_user_results in evaluation response, added eval json endpoint

* Remove query from response

* Utilize mapped inversed user id map to get original id in response

* Update serving test case to remove 'query' and add 'user_result' in response

* simplify user ID mapping

* Combined evaluation and evaluation_json endpoints

* Updated abort responses to show plaintext instead of html

* Added unit test cases

* Updated error response for empty data

* Added unit tests for provided data evaluation

* Update app.py

* Update test_app.py

---------

Co-authored-by: Quoc-Tuan Truong <tqtg@users.noreply.github.com>

* Add colunm model type into table models

* Update README.md

* Shorten requirements to give more space for other columns

* Shorten a few model names

* Update README.md

* Shorten LRPPM model title

* Shorten examples

---------

Co-authored-by: Quoc-Tuan Truong <tqtg@users.noreply.github.com>