Bump to version 6.0.0

Author: silil

README.md

@@ -28,10 +28,10 @@ Triage is designed to:
 To install Triage locally, you need:
 
 - Ubuntu/RedHat
-- Python 3.9+
-- A PostgreSQL 9.6+ database with your source data (events,
+- Python 3.10+
+- A PostgreSQL 13+ database with your source data (events,
   geographical data, etc) loaded.
-  - **NOTE**: If your database is PostgreSQL 11+ you will get some
+  - **NOTE**: If your database is PostgreSQL 13+ you will get some
     speed improvements. We recommend updating to a recent
     version of PostgreSQL.
 - Ample space on an available disk (or for example in Amazon Web
@@ -90,7 +90,7 @@ There are a plethora of options available for experiment running, affecting thin
 
 ## Development
 
-Triage was initially developed at [University of Chicago's Center For Data Science and Public Policy](http://dsapp.uchicago.edu) and is now being maintained at Carnegie Mellon University.
+Triage was initially developed at [University of Chicago's Center For Data Science and Public Policy](http://dsapp.uchicago.edu) and is now being maintained at [Carnegie Mellon University](https://datasciencepublicpolicy.org/).
 
 To build this package (without installation), its dependencies may
 alternatively be installed from the terminal using `pip`:

example/colab/colab_triage.ipynb

@@ -53,17 +53,35 @@
         "## Getting Set Up\n",
         "\n",
         "We'll need a few dependencies to run triage in a colab notebook:\n",
+        "- Python version 3.10\n",
         "- A local postgresql server (we'll use version 11)\n",
         "- A simplified dataset loaded into this database (we'll use data from DonorsChoose)\n",
         "- Triage and its dependencies (we'll use the current version in pypi)"
       ]
     },
+    {
+      "cell_type": "markdown",
+      "metadata": {},
+      "source": [
+        "### Python version 3.10 \n",
+        "\n",
+        "🛑     Triage requires Python 3.10 in Colab. Colab's default Python version is 3.12+, to change the version click at the bottom right of Colab where it says **Python 3** and select the option **\"Change runtime type\"**. A window with different options will pop up, choose the option **2025.07** on the \"Runtime version\" options section.and select **Save**. "
+      ]
+    },
     {
       "cell_type": "code",
       "execution_count": null,
-      "metadata": {
-        "id": "-htIBoS7N4gK"
-      },
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "# Making sure you have python 3.10 version\n",
+        "!python --version"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
       "outputs": [],
       "source": [
         "# Install and start postgresql-11 server\n",
@@ -144,12 +162,12 @@
       },
       "outputs": [],
       "source": [
-        "from sqlalchemy.engine.url import URL\n",
+        "from sqlalchemy.engine import URL\n",
         "from triage.util.db import create_engine\n",
         "import pandas as pd\n",
         "\n",
-        "db_url = URL(\n",
-        "            'postgres',\n",
+        "db_url = URL.create(\n",
+        "            'postgresql+psycopg2',\n",
         "            host='localhost',\n",
         "            username='postgres',\n",
         "            database='donors_choose',\n",
@@ -874,16 +892,26 @@
       "metadata": {
         "id": "jYzBKFG3qDhQ"
       },
-      "outputs": [],
+      "outputs": [
+        {
+          "ename": "SyntaxError",
+          "evalue": "invalid syntax (136456548.py, line 41)",
+          "output_type": "error",
+          "traceback": [
+            "\u001b[0;36m  Cell \u001b[0;32mIn[1], line 41\u001b[0;36m\u001b[0m\n\u001b[0;31m    cursor.execute(text\"SET ROLE {dbconfig['role']}\")\u001b[0m\n\u001b[0m                       ^\u001b[0m\n\u001b[0;31mSyntaxError\u001b[0m\u001b[0;31m:\u001b[0m invalid syntax\n"
+          ]
+        }
+      ],
       "source": [
         "import yaml\n",
         "import shutil\n",
         "import os\n",
         "import logging\n",
         "\n",
-        "from sqlalchemy.engine.url import URL\n",
+        "from sqlalchemy.engine import URL\n",
         "from sqlalchemy.event import listens_for\n",
         "from sqlalchemy.pool import Pool\n",
+        "from sqlalchemy import event\n",
         "\n",
         "from triage.util.db import create_engine\n",
         "from triage.experiments import MultiCoreExperiment\n",
@@ -910,18 +938,15 @@
         "  #     dbconfig = yaml.safe_load(dbf)\n",
         "\n",
         "  dbconfig = yaml.safe_load(database_yaml)\n",
-        "  print(dbconfig['role'])\n",
         "\n",
         "  # assume group role to ensure shared permissions\n",
-        "  @listens_for(Pool, \"connect\")\n",
-        "  def assume_role(dbapi_con, connection_record):\n",
-        "      logging.debug(f\"setting role {dbconfig['role']};\")\n",
-        "      dbapi_con.cursor().execute(f\"set role {dbconfig['role']};\")\n",
-        "      # logging.debug(f\"setting role postres;\")\n",
-        "      # dbapi_con.cursor().execute(f\"set role postgres;\")\n",
-        "\n",
-        "  db_url = URL(\n",
-        "              'postgres',\n",
+        "  @event.listens_for(Pool, \"connect\")\n",
+        "  def assume_role(dbapi_conn, connection_record):\n",
+        "      with dbapi_conn.cursor() as cursor:\n",
+        "          cursor.execute(\"SET ROLE %s\", (dbconfig['role'],))\n",
+        "\n",
+        "  db_url = URL.create(\n",
+        "              'postgresql+psycopg2',\n",
         "              host=dbconfig['host'],\n",
         "              username=dbconfig['user'],\n",
         "              database=dbconfig['db'],\n",
@@ -1005,15 +1030,15 @@
         "import yaml\n",
         "import pandas as pd\n",
         "\n",
-        "from sqlalchemy.engine.url import URL\n",
+        "from sqlalchemy.engine import URL\n",
         "from triage.util.db import create_engine\n",
         "from triage.component.postmodeling.experiment_summarizer import ExperimentReport\n",
         "from matplotlib import pyplot as plt\n",
         "%matplotlib inline\n",
         "\n",
         "dbconfig = yaml.safe_load(database_yaml)\n",
-        "db_url = URL(\n",
-        "            'postgres',\n",
+        "db_url = URL.create(\n",
+        "            'postgresql+psycopg2',\n",
         "            host=dbconfig['host'],\n",
         "            username=dbconfig['user'],\n",
         "            database=dbconfig['db'],\n",
@@ -1599,8 +1624,8 @@
       "outputs": [],
       "source": [
         "dbconfig = yaml.safe_load(database_yaml)\n",
-        "db_url = URL(\n",
-        "            'postgres',\n",
+        "db_url = URL.create(\n",
+        "            'postgresql+psycopg2',\n",
         "            host=dbconfig['host'],\n",
         "            username=dbconfig['user'],\n",
         "            database=dbconfig['db'],\n",
@@ -2010,7 +2035,7 @@
       "provenance": []
     },
     "kernelspec": {
-      "display_name": "colab_env",
+      "display_name": "triage_6",
       "language": "python",
       "name": "python3"
     },
@@ -2024,7 +2049,7 @@
       "name": "python",
       "nbconvert_exporter": "python",
       "pygments_lexer": "ipython3",
-      "version": "3.11.9"
+      "version": "3.10.6"
     }
   },
   "nbformat": 4,

requirement/main.txt

@@ -1,9 +1,9 @@
 polars==0.18.2
 pyarrow>=12.0.1
 numpy==1.26.0
-pandas==1.5.0
-alembic==1.7.7
-SQLAlchemy==1.3.18 # pyup: ignore
+pandas==2.0.3
+alembic==1.17.2
+SQLAlchemy==2.0
 PyYAML==6.0.2
 psycopg2-binary==2.9.3
 boto3==1.22.4
@@ -19,15 +19,14 @@ sqlparse==0.4.4
 pebble==4.6.3
 adjustText==0.7.3
 graphviz==0.20
-requests==2.31.0
+requests==2.32.3
 coloredlogs==15.0.1
 verboselogs==1.7
 s3fs==0.4.2 # pyup: ignore
 scikit-learn==1.6.1
-matplotlib==3.5.1
-matplotlib-inline==0.1.3
-seaborn==0.11.2
+matplotlib==3.8.1
+seaborn==0.13.2
 ohio==0.5.0
-aequitas==0.42.0
+aequitas==1.0.0
 plotly==5.24.0
 jupyter==1.0.0

setup.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 5.5.4
+current_version = 5.5.5
 commit = True
 tag = True
 

setup.py

@@ -44,7 +44,7 @@ def stream_requirements(fd):
 
 setup(
     name='triage',
-    version='5.5.5',
+    version='6.0.0',
     description="Risk modeling and prediction",
     long_description=README_PATH.read_text(),
     long_description_content_type="text/markdown",
@@ -73,10 +73,10 @@ def stream_requirements(fd):
         "License :: OSI Approved :: MIT License",
         "Natural Language :: English",
         "Programming Language :: Python :: 3",
-        "Programming Language :: Python :: 3.9",
         "Programming Language :: Python :: 3.10",
+        "Programming Language :: Python :: 3.11",
     ],
-    python_requires=">=3.8",
+    python_requires=">=3.10",
     test_suite="tests",
     tests_require=REQUIREMENTS_TEST,
 )

src/tests/architect_tests/test_builders.py

@@ -4,7 +4,7 @@
 import testing.postgresql
 
 from contextlib import contextmanager
-
+from sqlalchemy import text
 from triage import create_engine
 from triage.component.catwalk.utils import filename_friendly_hash
 from triage.component.architect.feature_group_creator import FeatureGroup
@@ -280,7 +280,9 @@ def test_make_entity_date_table():
                 experiment_hash=experiment_hash,
                 engine=engine,
             )
-            engine.execute("CREATE TABLE features.tmp_entity_date (a int, b date);")
+            with engine.begin() as conn:
+                conn.execute(text("CREATE TABLE features.tmp_entity_date (a int, b date);"))
+
             # call the function to test the creation of the table
             entity_date_table_name = builder.make_entity_date_table(
                 as_of_times=dates,
@@ -293,12 +295,11 @@ def test_make_entity_date_table():
             )
 
             # read in the table
-            result = pd.read_sql(
-                "select * from features.{} order by entity_id, as_of_date".format(
-                    entity_date_table_name
-                ),
-                engine,
-            )
+            with engine.connect() as conn:
+                result = pd.read_sql(
+                    f"select * from features.{entity_date_table_name} order by entity_id, as_of_date",
+                    conn.connection,
+                )
             # compare the table to the test dataframe
             test = result == ids_dates
             assert test.all().all()
@@ -327,8 +328,11 @@ def test_make_entity_date_table_include_missing_labels():
         label_timespan="1 month",
     )
     # this line adds the new entity-date combo as an expected one
-    ids_dates = ids_dates.append(
-        {"entity_id": 0, "as_of_date": datetime.date(2016, 6, 1)}, ignore_index=True
+    ids_dates = pd.concat([
+                            ids_dates, 
+                            pd.DataFrame({"entity_id": [0], "as_of_date": [datetime.date(2016, 6, 1)]})
+                           ], 
+                           ignore_index=True
     )
 
     with testing.postgresql.Postgresql() as postgresql:
@@ -347,7 +351,8 @@ def test_make_entity_date_table_include_missing_labels():
                 include_missing_labels_in_train_as=False,
                 engine=engine,
             )
-            engine.execute("CREATE TABLE features.tmp_entity_date (a int, b date);")
+            with engine.begin() as conn:
+                conn.execute(text("CREATE TABLE features.tmp_entity_date (a int, b date);"))
             # call the function to test the creation of the table
             entity_date_table_name = builder.make_entity_date_table(
                 as_of_times=dates,
@@ -360,12 +365,11 @@ def test_make_entity_date_table_include_missing_labels():
             )
 
             # read in the table
-            result = pd.read_sql(
-                "select * from features.{} order by entity_id, as_of_date".format(
-                    entity_date_table_name
-                ),
-                engine,
-            )
+            with engine.connect() as conn:
+                result = pd.read_sql(
+                    f"select * from features.{entity_date_table_name} order by entity_id, as_of_date",
+                    conn.connection,
+                )
 
             # compare the table to the test dataframe
             assert sorted(result.values.tolist()) == sorted(ids_dates.values.tolist())
@@ -484,33 +488,31 @@ def test_stitch_csvs(self):
 
                 matrix_store = matrix_storage_engine.get_store("1234")
 
-                result = builder.stitch_csvs(
+                result, labels_ = builder.stitch_csvs(
                     features_queries=feature_queries,
                     label_query=label_query,
                     matrix_store=matrix_store,
                     matrix_uuid="1234"
                 )
 
-                # chekc if entity_id and as_of_date are as index 
+                # check if entity_id and as_of_date are as index 
                 should_be = ['entity_id', 'as_of_date']
                 actual_indices = result.index.names
-
                 TestCase().assertListEqual(should_be, actual_indices)
 
-                # last element in the DF should be the label
+                # check that last element is not the label
                 last_col = 'booking'
                 output = result.columns.values[-1] # label name
+                TestCase().assertNotEqual(last_col, output)
 
-                TestCase().assertEqual(last_col, output)
-
-                # number of columns must be the sum of all the columns on each feature table + 1 for the label 
-                TestCase().assertEqual(result.shape[1], 4+1, 
-                                       "Number of features and label doesn't match")
+                # number of columns must be the sum of all the columns on each feature table 
+                TestCase().assertEqual(result.shape[1], 4, 
+                                       "Number of features don't match")
 
                 # number of rows 
                 assert result.shape[0] ==  5
                 TestCase().assertEqual(result.shape[0], 5, 
-                                       "Number of rows doesn't match")
+                                       "Number of rows don't match")
 
                 # types of the final df should be float32
                 types = set(result.apply(lambda x: x.dtype == 'float32').values)
@@ -688,17 +690,22 @@ def test_nullcheck(self):
                 }
 
                 uuid = filename_friendly_hash(good_metadata)
-                with self.assertRaises(ValueError):
-                    builder.build_matrix(
-                        as_of_times=dates,
-                        label_name="booking",
-                        label_type="binary",
-                        feature_dictionary=feature_dictionary,
-                        matrix_metadata=good_metadata,
-                        matrix_uuid=uuid,
-                        matrix_type="other",
-                    )
-                
+                # check if an exception was raised (but swallowed in the original code)
+                # so looking for a None data frame. Verify code in builders.py/build_matrix
+                # when we catch a ValueError we just do a return, not raising an exception 
+                # maybe we should consider to actually raise an exception? (2026-01-04)
+                # that will stop the run though!
+        
+                result = builder.build_matrix(
+                                    as_of_times=dates,
+                                    label_name="booking",
+                                    label_type="binary",
+                                    feature_dictionary=feature_dictionary,
+                                    matrix_metadata=good_metadata,
+                                    matrix_uuid=uuid,
+                                    matrix_type="other",
+                                )
+                self.assertIsNone(result)
 
     def test_replace_false_rerun(self):
         with testing.postgresql.Postgresql() as postgresql: