Merge pull request #626 from graphistry/dev/dev-skrub

Dev/dev skrub

.github/workflows/ci.yml

@@ -54,6 +54,8 @@ jobs:
         ./bin/lint.sh
 
     - name: Type check
+      env:
+          PYTHON_VERSION: ${{ matrix.python-version }}
       run: |
         source pygraphistry/bin/activate
         ./bin/typecheck.sh
@@ -101,6 +103,8 @@ jobs:
         ./bin/lint.sh
 
     - name: Type check
+      env:
+          PYTHON_VERSION: ${{ matrix.python-version }}
       run: |
         source pygraphistry/bin/activate
         ./bin/typecheck.sh
@@ -143,6 +147,8 @@ jobs:
         python -m pip install -e .[test,pygraphviz]
 
     - name: Type check
+      env:
+          PYTHON_VERSION: ${{ matrix.python-version }}
       run: |
         source pygraphistry/bin/activate
         ./bin/typecheck.sh
@@ -159,8 +165,7 @@ jobs:
 
     strategy:
       matrix:
-        #python-version: [3.8, 3.9, '3.10', 3.11, 3.12]
-        python-version: [3.8, 3.9]
+        python-version: [3.9, '3.10', 3.11, 3.12]
 
     steps:
 
@@ -185,6 +190,8 @@ jobs:
         python -m pip install -e .[test,testai,umap-learn]
 
     - name: Type check
+      env:
+          PYTHON_VERSION: ${{ matrix.python-version }}
       run: |
         source pygraphistry/bin/activate
         ./bin/typecheck.sh
@@ -206,8 +213,7 @@ jobs:
 
     strategy:
       matrix:
-        python-version: [3.8, 3.9]
-        #python-version: [3.8, 3.9, '3.10', 3.11, 3.12]
+        python-version: [3.9, '3.10', 3.11, 3.12]
         #include:
         #  - python-version: 3.12
         #    continue-on-error: true
@@ -233,14 +239,16 @@ jobs:
         source pygraphistry/bin/activate
         python -m pip install --upgrade pip
         python -m pip install -e .[test,testai,ai]
-        echo "dirty-cat: `pip show dirty-cat | grep Version`"
+        echo "skrub: `pip show skrub | grep Version`"
         echo "pandas: `pip show pandas | grep Version`"
         echo "numpy: `pip show numpy | grep Version`"
         echo "scikit-learn: `pip show scikit-learn | grep Version`"
         echo "scipy: `pip show scipy | grep Version`"
         echo "umap-learn: `pip show umap-learn | grep Version`"
 
     - name: Type check
+      env:
+          PYTHON_VERSION: ${{ matrix.python-version }}
       run: |
         source pygraphistry/bin/activate
         ./bin/typecheck.sh
@@ -270,6 +278,11 @@ jobs:
         source pygraphistry/bin/activate
         ./bin/test-embed.sh
 
+    - name: Full DGL tests (rich featurize)
+      run: |
+        source pygraphistry/bin/activate
+        ./bin/test-dgl.sh
+
 
   test-neo4j:
 

CHANGELOG.md

@@ -7,6 +7,43 @@ This project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.htm
 
 ## [Development]
 
+## [0.36.0 - 2025-02-05]
+
+### Breaking
+
+* `from_cugraph` returns using the src/dst bindings of `cugraph.Graph` object instead of base `Plottable`
+* `pip install graphistry[umap-learn]` and `pip install graphistry[ai]` are now Python 3.9+ (was 3.8+)
+* `Plottable`'s fields `_node_dbscan` / `_edge_dbscan` are now `_dbscan_nodes` / `_dbscan_edges`
+
+### Feat
+
+* Switch to `skrub` for feature engineering
+* More AI methods support GPU path
+* Support cugraph 26.10+, numpy 2.0+
+* Add more umap, dbscan fields to `Plottable`
+
+### Infra
+
+* `[umap-learn]` + `[ai]` unpin deps - scikit, scipy, torch (now 2), etc
+
+### Refactor
+
+* Move more type models to models/compute/{feature,umap,cluster}
+* Turn more print => logger
+
+### Fixes
+
+* Remove lint/type ignores and fix root causes
+
+### Tests
+
+* Stop ignoring warnings in featurize and umap
+* python version tests use corresponding python version for mypy
+* ci umap tests: py 3.8, 3.9 => 3.9..3.12
+* ci ai tests: py 3.8, 3.9 => 3.9..3.12
+* ci tests dgl
+* plugin tests check for module imports
+
 ## [0.35.10 - 2025-01-24]
 
 ### Fixes: 

bin/test-dgl.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+set -ex
+
+# Run from project root
+# - Args get passed to pytest phase
+# Non-zero exit code on fail
+
+# Assume [umap-learn,test]
+
+python -m pytest --version
+
+python -B -m pytest -vv \
+    graphistry/tests/test_dgl_utils.py

bin/typecheck.sh

@@ -6,5 +6,9 @@ set -ex
 
 mypy --version
 
-# Check core
-mypy --config-file mypy.ini graphistry
+if [ -n "$PYTHON_VERSION" ]; then
+  SHORT_VERSION=$(echo "$PYTHON_VERSION" | cut -d. -f1,2)
+  mypy --python-version "$SHORT_VERSION" --config-file mypy.ini graphistry
+else
+  mypy --config-file mypy.ini graphistry
+fi

docker/test-cpu-umap-ai.sh

@@ -2,7 +2,9 @@
 set -ex
 
 
-PYTHON_VERSION=${PYTHON_VERSION:-3.8} \
+TEST_FILES=${@:-"graphistry/tests/test_feature_utils.py graphistry/tests/test_umap_utils.py"}
+
+PYTHON_VERSION=${PYTHON_VERSION:-3.10} \
 PIP_DEPS=${PIP_DEPS:--e .[ai,test,testai]} \
 WITH_LINT=${WITH_LINT:-1} \
 WITH_TYPECHECK=${WITH_TYPECHECK:-1} \
@@ -11,6 +13,4 @@ WITH_TEST=${WITH_TEST:-1} \
 SENTENCE_TRANSFORMER=${SENTENCE_TRANSFORMER-average_word_embeddings_komninos} \
 SENTENCE_TRANSFORMER=${SENTENCE_TRANSFORMER} \
     ./test-cpu-local.sh \
-        graphistry/tests/test_feature_utils.py \
-        graphistry/tests/test_umap_utils.py \
-        $@
+        $TEST_FILES

docker/test-cpu-umap.sh

@@ -2,7 +2,9 @@
 set -ex
 
 
-PYTHON_VERSION=${PYTHON_VERSION:-3.8} \
+TEST_FILES=${@:-"graphistry/tests/test_feature_utils.py graphistry/tests/test_umap_utils.py"}
+
+PYTHON_VERSION=${PYTHON_VERSION:-3.9} \
 PIP_DEPS=${PIP_DEPS:--e .[umap-learn,test,testai]} \
 WITH_LINT=${WITH_LINT:-1} \
 WITH_TYPECHECK=${WITH_TYPECHECK:-1} \
@@ -11,6 +13,4 @@ WITH_TEST=${WITH_TEST:-1} \
 SENTENCE_TRANSFORMER=${SENTENCE_TRANSFORMER-average_word_embeddings_komninos} \
 SENTENCE_TRANSFORMER=${SENTENCE_TRANSFORMER} \
     ./test-cpu-local.sh \
-        graphistry/tests/test_feature_utils.py \
-        graphistry/tests/test_umap_utils.py \
-        $@
+        $TEST_FILES

docker/test-gpu-local.sh

@@ -47,5 +47,4 @@ docker run \
     ${NETWORK} \
     graphistry/test-gpu:${TEST_CPU_VERSION} \
         --maxfail=1 \
-        --ignore=graphistry/tests/test_feature_utils.py \
         $@

docs/source/conf.py

@@ -239,10 +239,10 @@
     ('py:class', 'torch'),
     ('py:class', 'umap'),
     ('py:class', 'sentence_transformers'),
-    ('py:class', 'dirty_cat'),
     ('py:class', 'sklearn'),
     ('py:class', 'scipy'),
     ('py:class', 'seaborn'),
+    ('py:class', 'skrub'),
     ('py:class', 'annoy'),
     ('py:class', 'NetworkX graph'),
     ('py:class', 'Pandas dataframe'),

graphistry/Engine.py

@@ -1,9 +1,9 @@
 from inspect import getmodule
+import warnings
 import numpy as np
 import pandas as pd
 from typing import Any, Optional, Union
 from enum import Enum
-from graphistry.utils.lazy_import import lazy_cudf_import
 
 
 class Engine(Enum):
@@ -29,6 +29,8 @@ def resolve_engine(
     g_or_df: Optional[Any] = None,
 ) -> Engine:
 
+    from graphistry.utils.lazy_import import lazy_cudf_import
+
     if isinstance(engine, str):
         engine = EngineAbstract(engine)
 
@@ -42,7 +44,8 @@ def resolve_engine(
         if isinstance(g_or_df, Plottable):
             if g_or_df._nodes is not None and g_or_df._edges is not None:
                 if not isinstance(g_or_df._nodes, type(g_or_df._edges)):
-                    raise ValueError(f'Edges and nodes must be same type for auto engine selection, got: {type(g_or_df._edges)} and {type(g_or_df._nodes)}')
+                    #raise ValueError(f'Edges and nodes must be same type for auto engine selection, got: {type(g_or_df._edges)} and {type(g_or_df._nodes)}')
+                    warnings.warn(f'Edges and nodes must be same type for auto engine selection, got: {type(g_or_df._edges)} and {type(g_or_df._nodes)}')                
             g_or_df = g_or_df._edges if g_or_df._edges is not None else g_or_df._nodes
 
     if g_or_df is not None:

graphistry/Plottable.py

@@ -2,7 +2,10 @@
 from typing_extensions import Literal
 import pandas as pd
 
+from graphistry.models.ModelDict import ModelDict
 from graphistry.models.compute.chain_remote import FormatType, OutputTypeAll, OutputTypeDf, OutputTypeGraph
+from graphistry.models.compute.dbscan import DBSCANEngine
+from graphistry.models.compute.umap import UMAPEngineConcrete
 from graphistry.plugins_types.cugraph_types import CuGraphKind
 from graphistry.Engine import Engine, EngineAbstract
 from graphistry.utils.json import JSONVal
@@ -72,11 +75,13 @@ class Plottable(object):
     _node_embedding : Optional[pd.DataFrame]
     _node_encoder : Optional[Any]
     _node_features : Optional[pd.DataFrame]
+    _node_features_raw: Optional[pd.DataFrame]
     _node_target : Optional[pd.DataFrame]
 
     _edge_embedding : Optional[pd.DataFrame]
     _edge_encoder : Optional[Any]
     _edge_features : Optional[pd.DataFrame]
+    _edge_features_raw: Optional[pd.DataFrame]
     _edge_target : Optional[pd.DataFrame]
 
     _weighted_adjacency: Optional[Any]
@@ -88,10 +93,27 @@ class Plottable(object):
     _xy: Optional[pd.DataFrame]
 
     _umap : Optional[UMAP]
-    _umap_params: Optional[Dict[str, Any]]
+    _umap_engine: Optional[UMAPEngineConcrete]
+    _umap_params: Optional[Union[ModelDict, Dict[str, Any]]]
     _umap_fit_kwargs: Optional[Dict[str, Any]]
     _umap_transform_kwargs: Optional[Dict[str, Any]]
 
+    # extra umap
+    _n_components: int
+    _metric: str
+    _n_neighbors: int
+    _min_dist: float
+    _spread: float
+    _local_connectivity: int
+    _repulsion_strength: float
+    _negative_sample_rate: float
+    _suffix: str
+
+    _dbscan_engine: Optional[DBSCANEngine]
+    _dbscan_params: Optional[ModelDict]
+    _dbscan_nodes: Optional[Any]  # fit model
+    _dbscan_edges: Optional[Any]  # fit model
+
     _adjacency : Optional[Any]
     _entity_to_index : Optional[dict]
     _index_to_entity : Optional[dict]

graphistry/PlotterBase.py

@@ -203,10 +203,16 @@ def __init__(self, *args: Any, **kwargs: Any) -> None:
 
         # the fit umap instance
         self._umap = None
+        self._umap_engine = None
         self._umap_params : Optional[Dict[str, Any]] = None
         self._umap_fit_kwargs : Optional[Dict[str, Any]] = None
         self._umap_transform_kwargs : Optional[Dict[str, Any]] = None
 
+        self._dbscan_engine = None
+        self._dbscan_params = None
+        self._dbscan_nodes = None  # fit model
+        self._dbscan_edges = None  # fit model
+
         self._adjacency = None
         self._entity_to_index = None
         self._index_to_entity = None
@@ -216,11 +222,7 @@ def __init__(self, *args: Any, **kwargs: Any) -> None:
         self._use_feat: bool = False
         self._triplets: Optional[List] = None 
         self._kg_embed_dim: int = 128
-
-        # Dbscan
-        self._node_dbscan = None  # the fit dbscan instance
-        self._edge_dbscan = None
-
+
         # DGL
         self.DGL_graph = None  # the DGL graph