Autodocs, index cleanup, linked scores bug fixes, linked structures optimizations (#53)

* refactor: relax strictness

* feat: add docbasemodel

* feat: add new functons to rdkit_utils (for next rerun)

* fix/feat: fix classification and improve annotation table generation

* feat: also return types

* feat: document and improve generation of protein_utils

* chore: remove unused functions

* feat: finish fixing all the formats and document

* chore: remove unused functions and tests - added TODOs

* feat: add rdkit tests + ligand TODO

* feat: add proper ligand stuff, remove neighboring, rename interacting

* chore: typing and linting

* fix: tests and annotation table columns

* chore: remove unused

* fix: structure_qc try except

* chore: linting for tests

* fix all tests and add clusters etc. to plindex

* chore: type and lint

* fix: remove get_extended_plindex

* fix: make sure cluster column exists

* fix: return nan for all nan arrays

* fix: cluster make optional

* docs: make doc descriptions

* lint

* refactor: require no warnings in documentation build

* fix column names

* fix: docs

* lint

* type

* fix missing column descriptions

* fix: use rigid=False and full_report=True

* feat: break up make links and linked structures, superposition and scoring, add tasks and pipeline steps

* fix: just make it run

* fix: ost can't serialize in mp

* chore: flow bump

* style: ruff

* fix: system_dir includes system

* fix: support correct parquet dataset layout

* chore: get source linked structures instead

* fix: test sample

* fix: get_linked_structure for new linked structures

* chore: make descriptions usable

* fix: type and description order in similarity clusters

* chore: vendor column_descriptions for progammatic usage

* fix: make column descriptions now idempotent

* test: add a test for doc base

* style: ruff

* fix: posebusters results for linked structures

* fix: weighted average posebusters results

* feat: add dedicated score_linked_structures and fix create_index

* chore: boilerplate score linked structures

* chore: flow bump

* fix: final fixes for metadata in the pipeline

* style: ruff

* style: mypy

* test: update links data

* test: fix get column descriptions test

* style: ruff

* fix: use latest scores when packing linked scores

* feat: support predictions_dir in evaluation

* fix: column description path

* chore: paranoid include of tsvs

* chore: work down docs discrepancies

* fix: docs as close as possible

* style: ruff

* test: fix test plindex for new columns

* docs: fix docgen pls

* docs: last bits for system pocket

* docs: whitelist ligand_rdkit_validation away from warning

* docs: example for exception

* fix: force anonymous client when bucket is plinder

* docs: notebooks

* test: don't delete the test file

* fix: use anon client for both methods

* fix: lazy download linked structures in PlinderSystem

* fix: postpone model input files to data loader pr

* fix: more ramifications of tutorial iteration

---------

Co-authored-by: yusuf1759 <mr.adeshina.yusuf@gmail.com>
Co-authored-by: Patrick Kunzmann <padix.key@gmail.com>
Co-authored-by: Ninjani <janani.durairaj@gmail.com>
Co-authored-by: Tom Duignan <tjduigna@users.noreply.github.com>
Co-authored-by: Thomas Duignan <thomas.j.duignan@gmail.com>

Author: 5 people

.pre-commit-config.yaml

@@ -18,11 +18,11 @@ repos:
     hooks:
       - id: ruff-format
         name: ruff-format
-        entry: bash -c 'ruff format --force-exclude --preview src'
+        entry: bash -c 'ruff format --force-exclude --preview src tests'
         language: system
         types: [python]
       - id: ruff-linter
         name: ruff-linter
-        entry: bash -c 'ruff check --fix src'
+        entry: bash -c 'ruff check --fix src tests'
         language: system
         types: [python]