Merge branch 'main' into 1940-replace_sliders_by_buttons

backend/ml/inputs.py

@@ -4,7 +4,7 @@
 import pandas as pd
 from django.db.models import Case, F, Q, QuerySet, When
 from django.db.models.expressions import RawSQL
-from solidago.pipeline import TournesolInput
+from solidago.pipeline import PipelineInput
 
 from core.models import User
 from tournesol.models import (
@@ -17,7 +17,7 @@
 from vouch.models import Voucher
 
 
-class MlInputFromDb(TournesolInput):
+class MlInputFromDb(PipelineInput):
     SCALING_CALIBRATION_MIN_ENTITIES_TO_COMPARE = 20
 
     def __init__(self, poll_name: str):
@@ -58,38 +58,38 @@ def get_scaling_calibration_users(self) -> QuerySet[User]:
             n_compared_entities__gte=self.SCALING_CALIBRATION_MIN_ENTITIES_TO_COMPARE,
         ).order_by("-n_compared_entities")[: self.MAX_SCALING_CALIBRATION_USERS]
 
-    def get_comparisons(self, criteria=None, user_id=None) -> pd.DataFrame:
+    def get_comparisons(self, criterion=None, user_id=None) -> pd.DataFrame:
         scores_queryset = ComparisonCriteriaScore.objects.filter(
             comparison__poll__name=self.poll_name,
             comparison__user__is_active=True,
         )
-        if criteria is not None:
-            scores_queryset = scores_queryset.filter(criteria=criteria)
+        if criterion is not None:
+            scores_queryset = scores_queryset.filter(criteria=criterion)
 
         if user_id is not None:
             scores_queryset = scores_queryset.filter(comparison__user_id=user_id)
 
         values = scores_queryset.values(
             "score",
             "score_max",
-            "criteria",
             "weight",
+            criterion=F("criteria"),
             entity_a=F("comparison__entity_1_id"),
             entity_b=F("comparison__entity_2_id"),
             user_id=F("comparison__user_id"),
         )
         if len(values) > 0:
             dtf = pd.DataFrame(values)
             return dtf[
-                ["user_id", "entity_a", "entity_b", "criteria", "score", "score_max", "weight"]
+                ["user_id", "entity_a", "entity_b", "criterion", "score", "score_max", "weight"]
             ]
 
         return pd.DataFrame(
             columns=[
                 "user_id",
                 "entity_a",
                 "entity_b",
-                "criteria",
+                "criterion",
                 "score",
                 "score_max",
                 "weight",
@@ -136,7 +136,7 @@ def get_user_scalings(self, user_id=None) -> pd.DataFrame:
         Returns:
         - ratings_df: DataFrame with columns
             * `user_id`: int
-            * `criteria`: str
+            * `criterion`: str
             * `scale`: float
             * `scale_uncertainty`: float
             * `translation`: float
@@ -148,17 +148,18 @@ def get_user_scalings(self, user_id=None) -> pd.DataFrame:
             scalings = scalings.filter(user_id=user_id)
         values = scalings.values(
             "user_id",
-            "criteria",
             "scale",
             "scale_uncertainty",
             "translation",
             "translation_uncertainty",
+            criterion=F("criteria"),
+
         )
         if len(values) == 0:
             return pd.DataFrame(
                 columns=[
                     "user_id",
-                    "criteria",
+                    "criterion",
                     "scale",
                     "scale_uncertainty",
                     "translation",
@@ -168,28 +169,28 @@ def get_user_scalings(self, user_id=None) -> pd.DataFrame:
         return pd.DataFrame(values)
 
     def get_individual_scores(
-        self, criteria: Optional[str] = None, user_id: Optional[int] = None
+        self, user_id: Optional[int] = None, criterion: Optional[str] = None,
     ) -> pd.DataFrame:
         scores_queryset = ContributorRatingCriteriaScore.objects.filter(
             contributor_rating__poll__name=self.poll_name,
             contributor_rating__user__is_active=True,
         )
-        if criteria is not None:
-            scores_queryset = scores_queryset.filter(criteria=criteria)
+        if criterion is not None:
+            scores_queryset = scores_queryset.filter(criteria=criterion)
         if user_id is not None:
             scores_queryset = scores_queryset.filter(contributor_rating__user_id=user_id)
 
         values = scores_queryset.values(
             "raw_score",
-            "criteria",
-            entity=F("contributor_rating__entity_id"),
+            criterion=F("criteria"),
+            entity_id=F("contributor_rating__entity_id"),
             user_id=F("contributor_rating__user_id"),
         )
         if len(values) == 0:
-            return pd.DataFrame(columns=["user_id", "entity", "criteria", "raw_score"])
+            return pd.DataFrame(columns=["user_id", "entity_id", "criterion", "raw_score"])
 
         dtf = pd.DataFrame(values)
-        return dtf[["user_id", "entity", "criteria", "raw_score"]]
+        return dtf[["user_id", "entity_id", "criterion", "raw_score"]]
 
     def get_vouches(self):
         values = Voucher.objects.filter(

backend/ml/management/commands/ml_train.py

@@ -1,5 +1,6 @@
 import os
 from concurrent.futures import ProcessPoolExecutor, ThreadPoolExecutor, as_completed
+from functools import cache
 
 from django import db
 from django.conf import settings
@@ -18,6 +19,7 @@
 from tournesol.models.poll import ALGORITHM_MEHESTAN, DEFAULT_POLL_NAME
 
 
+@cache
 def get_solidago_pipeline(run_trust_propagation: bool = True):
     if run_trust_propagation:
         trust_algo = LipschiTrust()

backend/ml/mehestan/run.py

@@ -4,7 +4,7 @@
 
 from django import db
 from django.conf import settings
-from solidago.pipeline import TournesolInput
+from solidago.pipeline import PipelineInput
 from solidago.pipeline.legacy2023.criterion_pipeline import run_pipeline_for_criterion
 from solidago.pipeline.legacy2023.individual_scores import get_individual_scores
 
@@ -21,15 +21,15 @@
 def update_user_scores(poll: Poll, user: User):
     params = MehestanParameters()
     ml_input = MlInputFromDb(poll_name=poll.name)
-    for criteria in poll.criterias_list:
-        output = TournesolPollOutput(poll_name=poll.name, criterion=criteria)
+    for criterion in poll.criterias_list:
+        output = TournesolPollOutput(poll_name=poll.name, criterion=criterion)
         scores = get_individual_scores(
             ml_input,
-            criteria,
+            criterion,
             parameters=params,
             single_user_id=user.pk,
         )
-        scores["criteria"] = criteria
+        scores["criterion"] = criterion
         scores.rename(
             columns={
                 "score": "raw_score",
@@ -45,7 +45,7 @@ def close_db_connection_callback():
 
 
 def run_mehestan(
-    ml_input: TournesolInput, poll: Poll, parameters: MehestanParameters, main_criterion_only=False
+    ml_input: PipelineInput, poll: Poll, parameters: MehestanParameters, main_criterion_only=False
 ):
     """
     This function use multiprocessing.

backend/ml/outputs.py

@@ -166,7 +166,7 @@ def save_individual_scores(
                         raw_uncertainty=row.raw_uncertainty,
                         voting_right=row.voting_right,
                     )
-                    for _, row in scores.iterrows()
+                    for row in scores.itertuples()
                 ),
                 batch_size=10000,
             )
@@ -257,7 +257,7 @@ def apply_score_scalings(poll: Poll, contributor_scores: pd.DataFrame):
         contributor_scores: DataFrame with columns:
             user_id: int
             entity_id: int
-            criteria: str
+            criterion: str
             raw_score: float
             raw_uncertainty: float
 
@@ -270,14 +270,17 @@ def apply_score_scalings(poll: Poll, contributor_scores: pd.DataFrame):
         return contributor_scores
 
     ml_input = MlInputFromDb(poll_name=poll.name)
-    scalings = ml_input.get_user_scalings().set_index(["user_id", "criteria"])
+    scalings = ml_input.get_user_scalings().set_index(["user_id", "criterion"])
     contributor_scores = contributor_scores.join(
-        scalings, on=["user_id", "criteria"], how="left"
+        scalings, on=["user_id", "criterion"], how="left"
+    ).fillna(
+        {
+            "scale": 1.0,
+            "translation": 0.0,
+            "scale_uncertainty": 0.0,
+            "translation_uncertatinty": 0.0,
+        }
     )
-    contributor_scores["scale"].fillna(1, inplace=True)
-    contributor_scores["translation"].fillna(0, inplace=True)
-    contributor_scores["scale_uncertainty"].fillna(0, inplace=True)
-    contributor_scores["translation_uncertainty"].fillna(0, inplace=True)
 
     # Apply individual scaling
     contributor_scores["uncertainty"] = (

backend/requirements.txt

@@ -37,14 +37,14 @@ PyYAML==6.0.1
 langdetect==1.0.9
 # Pandas is used extensively in the ML algorithms and for some data management
 # tasks such as building the public dataset
-pandas==2.1.2
+pandas==2.2.3
 # Numba provides just-in-time compilation to run optimized machine code
 # for performance-critical functions in Mehestan implementation.
-numba==0.58.1
+numba==0.60.0
 # Numpy is used extensively in the ML algorithms and in some other algorithms
 # such as computing comparison suggestions. See https://numpy.org/
 # Check the compatibility with Numba before upgrading.
-numpy==1.26.1
+numpy==1.26.4
 # Scipy is used in some ML algorithms
 scipy==1.11.3
 # API Youtube data

backend/tournesol/management/commands/load_public_dataset.py

@@ -6,7 +6,7 @@
 from django.core.management import call_command
 from django.core.management.base import BaseCommand
 from django.db import transaction
-from solidago.pipeline.inputs import TournesolInputFromPublicDataset
+from solidago.pipeline.inputs import TournesolDataset
 
 from core.models import User
 from core.models.user import EmailDomain
@@ -27,7 +27,7 @@ def add_arguments(self, parser):
         parser.add_argument("--user-sampling", type=float, default=None)
         parser.add_argument("--dataset-url", type=str, default=PUBLIC_DATASET_URL)
 
-    def create_user(self, username: str, ml_input: TournesolInputFromPublicDataset):
+    def create_user(self, username: str, ml_input: TournesolDataset):
         user = ml_input.users.loc[ml_input.users.public_username == username].iloc[0]
         is_pretrusted = user.trust_score > 0.5
         email = f"{username}@trusted.example" if is_pretrusted else f"{username}@example.com"
@@ -66,7 +66,7 @@ def create_test_user(self):
         )
 
     def handle(self, *args, **options):
-        public_dataset = TournesolInputFromPublicDataset(options["dataset_url"])
+        public_dataset = TournesolDataset(options["dataset_url"])
         nb_comparisons = 0
 
         with transaction.atomic():
@@ -108,7 +108,7 @@ def handle(self, *args, **options):
                 for values in rows.itertuples(index=False):
                     ComparisonCriteriaScore.objects.create(
                         comparison=comparison,
-                        criteria=values.criteria,
+                        criteria=values.criterion,
                         score=values.score,
                         score_max=values.score_max,
                     )

backend/tournesol/tests/test_api_exports.py

@@ -17,7 +17,7 @@
 from django.test import TransactionTestCase, override_settings
 from rest_framework import status
 from rest_framework.test import APIClient
-from solidago.pipeline.inputs import TournesolInputFromPublicDataset
+from solidago.pipeline.inputs import TournesolDataset
 
 from core.models import User
 from core.tests.factories.user import UserFactory
@@ -542,14 +542,14 @@ def test_use_public_export_as_ml_input(self):
         self.assertEqual(response.status_code, status.HTTP_200_OK)
         zip_content = io.BytesIO(response.content)
 
-        ml_input = TournesolInputFromPublicDataset(zip_content)
+        ml_input = TournesolDataset(zip_content)
         comparisons_df = ml_input.get_comparisons()
         rating_properties = ml_input.ratings_properties
 
         self.assertEqual(len(comparisons_df), 1)
         self.assertEqual(
             list(comparisons_df.columns),
-            ["user_id", "entity_a", "entity_b", "criteria", "score", "score_max", "weight"],
+            ["user_id", "entity_a", "entity_b", "criterion", "score", "score_max", "weight"],
         )
 
         self.assertEqual(len(rating_properties), 2)

solidago/.gitignore

@@ -1,3 +1,4 @@
+/devenv/
 /dist/
 **__pycache__/
 .flake8/

solidago/experiments/data_analysis.py

@@ -1,10 +1,10 @@
-from solidago.pipeline.inputs import TournesolInputFromPublicDataset
+from solidago.pipeline.inputs import TournesolDataset
 import numpy as np
 import pandas as pd
 import matplotlib.pyplot as plt
 import scipy
 
-data = TournesolInputFromPublicDataset.download()
+data = TournesolDataset.download()
 
 criteria = {
     "reliability": "Reliable and not misleading",

solidago/experiments/tournesol.py

@@ -5,7 +5,7 @@
 
 from threading import Thread
 
-from solidago.pipeline.inputs import TournesolInputFromPublicDataset
+from solidago.pipeline.inputs import TournesolDataset
 
 from solidago.trust_propagation import LipschiTrust
 from solidago.voting_rights import AffineOvertrust
@@ -32,7 +32,7 @@
     info_logger.addHandler(ch)
 
 logger.info("Retrieve public dataset")
-inputs = TournesolInputFromPublicDataset.download()
+inputs = TournesolDataset.download()
 video_id_to_entity_id = { 
     video_id: entity_id 
     for entity_id, video_id in enumerate(inputs.entity_id_to_video_id)

solidago/pyproject.toml

@@ -20,7 +20,7 @@ keywords = ["tournesol", "collaborative recommendations", "judgement aggregation
 dependencies = [
     "pandas>=1.5.3,<3.0",
     "numpy>=1.24.3,<1.27",
-    "numba==0.58.1",
+    "numba==0.60.0",
 ]
 dynamic = ["version"]
 

solidago/src/solidago/__version__.py

@@ -1,4 +1,4 @@
 # Changing the version will automatically publish a new version on PyPI.
 # (see /.github/workflows/solidago-publish.yml)
 
-__version__ = "0.2.0"
+__version__ = "0.3.0"

solidago/src/solidago/pipeline/__init__.py

@@ -1,5 +1,5 @@
-from .inputs import TournesolInput
+from .inputs import PipelineInput
 from .outputs import PipelineOutput
 from .pipeline import DefaultPipeline, Pipeline
 
-__all__ = ["TournesolInput", "DefaultPipeline", "Pipeline", "PipelineOutput"]
+__all__ = ["PipelineInput", "DefaultPipeline", "Pipeline", "PipelineOutput"]