churnkit 0.76.1a1__py3-none-any.whl → 0.76.1a2__py3-none-any.whl

{churnkit-0.76.1a1.data → churnkit-0.76.1a2.data}/data/share/churnkit/exploration_notebooks/10_spec_generation.ipynb

@@ -70,10 +70,12 @@
    "outputs": [],
    "source": [
     "from customer_retention.analysis.notebook_progress import track_and_export_previous\n",
+    "\n",
     "track_and_export_previous(\"10_spec_generation.ipynb\")\n",
     "\n",
-    "from pathlib import Path\n",
     "from enum import Enum\n",
+    "from pathlib import Path\n",
+    "\n",
     "\n",
     "class GenerationTarget(Enum):\n",
     "    LOCAL_FEAST_MLFLOW = \"local\"\n",
@@ -99,8 +101,7 @@
     "\n",
     "print(f\"Pipeline: {PIPELINE_NAME}\")\n",
     "print(f\"Target: {GENERATION_TARGET.value}\")\n",
-    "print(f\"Format: {OUTPUT_FORMAT.value}\")\n",
-    "from customer_retention.stages.temporal import TEMPORAL_METADATA_COLS"
+    "print(f\"Format: {OUTPUT_FORMAT.value}\")"
    ]
   },
   {
@@ -143,9 +144,11 @@
    "outputs": [],
    "source": [
     "import yaml\n",
+    "\n",
     "from customer_retention.analysis.auto_explorer import ExplorationFindings\n",
     "from customer_retention.analysis.auto_explorer.layered_recommendations import RecommendationRegistry\n",
-    "from customer_retention.core.config.experiments import FINDINGS_DIR, EXPERIMENTS_DIR, OUTPUT_DIR, setup_experiments_structure\n",
+    "from customer_retention.core.config.experiments import EXPERIMENTS_DIR, FINDINGS_DIR\n",
+    "\n",
     "\n",
     "def load_findings_and_recommendations(findings_dir: Path):\n",
     "    findings_files = sorted(\n",
@@ -154,37 +157,37 @@
     "    )\n",
     "    if not findings_files:\n",
     "        raise FileNotFoundError(f\"No findings in {findings_dir}. Run exploration notebooks first.\")\n",
-    "    \n",
+    "\n",
     "    findings = ExplorationFindings.load(str(findings_files[0]))\n",
-    "    \n",
+    "\n",
     "    # Look for recommendations file matching the findings file pattern\n",
     "    # Step 06 saves as: {name}_recommendations.yaml (matching {name}_findings.yaml)\n",
     "    findings_name = findings_files[0].stem.replace(\"_findings\", \"\")\n",
     "    recommendations_path = findings_dir / f\"{findings_name}_recommendations.yaml\"\n",
-    "    \n",
+    "\n",
     "    # Fallback to generic recommendations.yaml if not found\n",
     "    if not recommendations_path.exists():\n",
     "        recommendations_path = findings_dir / \"recommendations.yaml\"\n",
-    "    \n",
+    "\n",
     "    # Final fallback: find any *_recommendations.yaml\n",
     "    if not recommendations_path.exists():\n",
-    "        rec_files = sorted(findings_dir.glob(\"*_recommendations.yaml\"), \n",
+    "        rec_files = sorted(findings_dir.glob(\"*_recommendations.yaml\"),\n",
     "                          key=lambda f: f.stat().st_mtime, reverse=True)\n",
     "        if rec_files:\n",
     "            recommendations_path = rec_files[0]\n",
-    "    \n",
+    "\n",
     "    registry = None\n",
     "    if recommendations_path.exists():\n",
     "        with open(recommendations_path) as f:\n",
     "            registry = RecommendationRegistry.from_dict(yaml.safe_load(f))\n",
     "        print(f\"Loaded recommendations from: {recommendations_path.name}\")\n",
-    "    \n",
+    "\n",
     "    multi_dataset_path = findings_dir / \"multi_dataset_findings.yaml\"\n",
     "    multi_dataset = None\n",
     "    if multi_dataset_path.exists():\n",
     "        with open(multi_dataset_path) as f:\n",
     "            multi_dataset = yaml.safe_load(f)\n",
-    "    \n",
+    "\n",
     "    return findings, registry, multi_dataset\n",
     "\n",
     "findings, registry, multi_dataset = load_findings_and_recommendations(FINDINGS_DIR)\n",
@@ -244,7 +247,7 @@
     "    if not registry:\n",
     "        print(\"No recommendations loaded. Run notebooks 02-07 first.\")\n",
     "        return\n",
-    "    \n",
+    "\n",
     "    for layer in [\"bronze\", \"silver\", \"gold\"]:\n",
     "        recs = registry.get_by_layer(layer)\n",
     "        print(f\"\\n{layer.upper()} ({len(recs)} recommendations):\")\n",
@@ -300,7 +303,6 @@
    },
    "outputs": [],
    "source": [
-    "import os\n",
     "\n",
     "output_dir = OUTPUT_BASE_DIR / GENERATION_TARGET.value / PIPELINE_NAME\n",
     "output_dir.mkdir(parents=True, exist_ok=True)\n",
@@ -348,24 +350,24 @@
    "outputs": [],
    "source": [
     "if GENERATION_TARGET == GenerationTarget.LOCAL_FEAST_MLFLOW:\n",
-    "    from customer_retention.generators.spec_generator import MLflowPipelineGenerator, MLflowConfig\n",
     "    from customer_retention.generators.pipeline_generator import PipelineGenerator\n",
-    "    \n",
+    "    from customer_retention.generators.spec_generator import MLflowConfig, MLflowPipelineGenerator\n",
+    "\n",
     "    mlflow_config = MLflowConfig(\n",
     "        tracking_uri=\"./mlruns\",\n",
     "        experiment_name=PIPELINE_NAME,\n",
     "        log_data_quality=True,\n",
     "        nested_runs=True\n",
     "    )\n",
-    "    \n",
+    "\n",
     "    mlflow_gen = MLflowPipelineGenerator(mlflow_config=mlflow_config, output_dir=str(output_dir))\n",
-    "    \n",
+    "\n",
     "    if OUTPUT_FORMAT == OutputFormat.PYTHON:\n",
     "        saved = mlflow_gen.save_all(findings)\n",
     "        print(\"Generated MLflow pipeline files:\")\n",
     "        for f in saved:\n",
     "            print(f\"  {f}\")\n",
-    "    \n",
+    "\n",
     "    pipeline_gen = PipelineGenerator(\n",
     "        findings_dir=str(FINDINGS_DIR),\n",
     "        output_dir=str(output_dir),\n",
@@ -421,7 +423,7 @@
    "source": [
     "if GENERATION_TARGET == GenerationTarget.DATABRICKS:\n",
     "    from customer_retention.generators.spec_generator import DatabricksSpecGenerator, PipelineSpec, SourceSpec\n",
-    "    \n",
+    "\n",
     "    spec = PipelineSpec(\n",
     "        name=PIPELINE_NAME,\n",
     "        version=\"1.0.0\",\n",
@@ -431,7 +433,7 @@
     "            format=findings.source_format\n",
     "        )]\n",
     "    )\n",
-    "    \n",
+    "\n",
     "    if findings.target_column:\n",
     "        from customer_retention.generators.spec_generator import ModelSpec\n",
     "        spec.model_config = ModelSpec(\n",
@@ -439,13 +441,13 @@
     "            model_type=\"gradient_boosting\",\n",
     "            target_column=findings.target_column\n",
     "        )\n",
-    "    \n",
+    "\n",
     "    db_gen = DatabricksSpecGenerator(\n",
     "        catalog=DATABRICKS_CATALOG,\n",
     "        schema=DATABRICKS_SCHEMA,\n",
     "        output_dir=str(output_dir)\n",
     "    )\n",
-    "    \n",
+    "\n",
     "    saved = db_gen.save_all(spec)\n",
     "    print(\"Generated Databricks artifacts:\")\n",
     "    for f in saved:\n",
@@ -495,15 +497,15 @@
    "source": [
     "if GENERATION_TARGET == GenerationTarget.LLM_DOCS:\n",
     "    from customer_retention.analysis.auto_explorer import RecommendationEngine\n",
-    "    \n",
+    "\n",
     "    recommender = RecommendationEngine()\n",
     "    target_rec = recommender.recommend_target(findings)\n",
     "    feature_recs = recommender.recommend_features(findings)\n",
     "    cleaning_recs = recommender.recommend_cleaning(findings)\n",
-    "    \n",
+    "\n",
     "    docs_dir = output_dir / \"docs\"\n",
     "    docs_dir.mkdir(parents=True, exist_ok=True)\n",
-    "    \n",
+    "\n",
     "    # 1. Overview\n",
     "    overview = f\"\"\"# {PIPELINE_NAME} Pipeline Overview\n",
     "\n",
@@ -526,7 +528,7 @@
     "    for name, col in list(findings.columns.items())[:20]:\n",
     "        overview += f\"| {name} | {col.inferred_type.value} | {col.null_percentage:.1f}% | {col.unique_count} |\\n\"\n",
     "    (docs_dir / \"01_overview.md\").write_text(overview)\n",
-    "    \n",
+    "\n",
     "    # 2. Bronze layer - separate file per source\n",
     "    if registry and registry.sources:\n",
     "        for source_name, bronze_recs in registry.sources.items():\n",
@@ -539,38 +541,38 @@
     "\"\"\"\n",
     "            for rec in bronze_recs.null_handling:\n",
     "                bronze_doc += f\"- `{rec.target_column}`: {rec.action} ({rec.parameters.get('strategy', '')}) - {rec.rationale}\\n\"\n",
-    "            \n",
+    "\n",
     "            bronze_doc += \"\\n## Outlier Handling\\n\"\n",
     "            for rec in bronze_recs.outlier_handling:\n",
     "                bronze_doc += f\"- `{rec.target_column}`: {rec.action} - {rec.rationale}\\n\"\n",
-    "            \n",
+    "\n",
     "            bronze_doc += \"\\n## Type Conversions\\n\"\n",
     "            for rec in bronze_recs.type_conversions:\n",
     "                bronze_doc += f\"- `{rec.target_column}`: {rec.action} - {rec.rationale}\\n\"\n",
-    "            \n",
+    "\n",
     "            bronze_doc += \"\\n## Deduplication\\n\"\n",
     "            for rec in bronze_recs.deduplication:\n",
     "                bronze_doc += f\"- `{rec.target_column}`: {rec.action} - {rec.rationale}\\n\"\n",
-    "            \n",
+    "\n",
     "            bronze_doc += \"\\n## Filtering\\n\"\n",
     "            for rec in bronze_recs.filtering:\n",
     "                bronze_doc += f\"- `{rec.target_column}`: {rec.action} - {rec.rationale}\\n\"\n",
-    "            \n",
+    "\n",
     "            bronze_doc += \"\\n## Text Processing\\n\"\n",
     "            for rec in bronze_recs.text_processing:\n",
     "                bronze_doc += f\"- `{rec.target_column}`: {rec.action} - {rec.rationale}\\n\"\n",
-    "            \n",
+    "\n",
     "            safe_name = source_name.replace(\" \", \"_\").lower()\n",
     "            (docs_dir / f\"02_bronze_cleaning_{safe_name}.md\").write_text(bronze_doc)\n",
     "    else:\n",
-    "        bronze_doc = f\"\"\"# Bronze Layer - Data Cleaning\n",
+    "        bronze_doc = \"\"\"# Bronze Layer - Data Cleaning\n",
     "\n",
     "## Cleaning Recommendations\n",
     "\"\"\"\n",
     "        for rec in cleaning_recs:\n",
     "            bronze_doc += f\"\\n### {rec.column_name}\\n- **Strategy**: {rec.strategy}\\n- **Severity**: {rec.severity}\\n- **Rationale**: {rec.rationale}\\n\"\n",
     "        (docs_dir / \"02_bronze_cleaning.md\").write_text(bronze_doc)\n",
-    "    \n",
+    "\n",
     "    # 3. Silver layer\n",
     "    silver_doc = \"\"\"# Silver Layer - Feature Engineering\n",
     "\n",
@@ -580,18 +582,18 @@
     "        silver_doc += \"\\n### Joins\\n\"\n",
     "        for rec in registry.silver.joins:\n",
     "            silver_doc += f\"- {rec.parameters.get('left_source', '')} ⟷ {rec.parameters.get('right_source', '')} on `{rec.parameters.get('join_keys', [])}`\\n\"\n",
-    "        \n",
+    "\n",
     "        silver_doc += \"\\n### Aggregations\\n\"\n",
     "        for rec in registry.silver.aggregations:\n",
     "            silver_doc += f\"- `{rec.target_column}`: {rec.action} - windows: {rec.parameters.get('windows', [])}\\n\"\n",
-    "        \n",
+    "\n",
     "        silver_doc += \"\\n### Derived Columns\\n\"\n",
     "        for rec in registry.silver.derived_columns:\n",
     "            silver_doc += f\"- `{rec.target_column}`: {rec.parameters.get('expression', rec.action)}\\n\"\n",
     "    else:\n",
     "        silver_doc += \"\\nNo silver-layer recommendations found.\\n\"\n",
     "    (docs_dir / \"03_silver_features.md\").write_text(silver_doc)\n",
-    "    \n",
+    "\n",
     "    # 4. Gold layer\n",
     "    gold_doc = \"\"\"# Gold Layer - ML Features\n",
     "\n",
@@ -599,25 +601,25 @@
     "\"\"\"\n",
     "    for rec in feature_recs[:15]:\n",
     "        gold_doc += f\"\\n### {rec.feature_name}\\n- **Source**: {rec.source_column}\\n- **Type**: {rec.feature_type}\\n- **Description**: {rec.description}\\n\"\n",
-    "    \n",
+    "\n",
     "    if registry and registry.gold:\n",
     "        gold_doc += \"\\n## Encoding\\n\"\n",
     "        for rec in registry.gold.encoding:\n",
     "            gold_doc += f\"- `{rec.target_column}`: {rec.parameters.get('method', rec.action)}\\n\"\n",
-    "        \n",
+    "\n",
     "        gold_doc += \"\\n## Scaling\\n\"\n",
     "        for rec in registry.gold.scaling:\n",
     "            gold_doc += f\"- `{rec.target_column}`: {rec.parameters.get('method', rec.action)}\\n\"\n",
-    "        \n",
+    "\n",
     "        gold_doc += \"\\n## Feature Selection\\n\"\n",
     "        for rec in registry.gold.feature_selection:\n",
     "            gold_doc += f\"- `{rec.target_column}`: {rec.action} - {rec.rationale}\\n\"\n",
-    "        \n",
+    "\n",
     "        gold_doc += \"\\n## Transformations\\n\"\n",
     "        for rec in registry.gold.transformations:\n",
     "            gold_doc += f\"- `{rec.target_column}`: {rec.action} - {rec.parameters}\\n\"\n",
     "    (docs_dir / \"04_gold_ml_features.md\").write_text(gold_doc)\n",
-    "    \n",
+    "\n",
     "    # 5. Training\n",
     "    training_doc = f\"\"\"# Model Training\n",
     "\n",
@@ -636,7 +638,7 @@
     "- F1 Score\n",
     "\"\"\"\n",
     "    (docs_dir / \"05_training.md\").write_text(training_doc)\n",
-    "    \n",
+    "\n",
     "    print(\"Generated LLM documentation:\")\n",
     "    for f in sorted(docs_dir.glob(\"*.md\")):\n",
     "        print(f\"  {f.name}\")\n",
@@ -687,11 +689,12 @@
    "source": [
     "import json\n",
     "\n",
+    "\n",
     "def py_to_notebook(py_path: Path):\n",
     "    content = py_path.read_text()\n",
     "    cells = []\n",
     "    current_lines = []\n",
-    "    \n",
+    "\n",
     "    for line in content.split(\"\\n\"):\n",
     "        if line.startswith(\"# %% \") or line.startswith(\"# %%\\n\"):\n",
     "            if current_lines:\n",
@@ -702,16 +705,16 @@
     "                cells.append({\"cell_type\": \"markdown\", \"metadata\": {}, \"source\": [f\"## {title}\"]})\n",
     "        else:\n",
     "            current_lines.append(line + \"\\n\")\n",
-    "    \n",
+    "\n",
     "    if current_lines:\n",
     "        cells.append({\"cell_type\": \"code\", \"metadata\": {}, \"source\": current_lines, \"outputs\": [], \"execution_count\": None})\n",
-    "    \n",
+    "\n",
     "    notebook = {\n",
     "        \"cells\": cells,\n",
     "        \"metadata\": {\"kernelspec\": {\"display_name\": \"Python 3\", \"language\": \"python\", \"name\": \"python3\"}},\n",
     "        \"nbformat\": 4, \"nbformat_minor\": 4\n",
     "    }\n",
-    "    \n",
+    "\n",
     "    out_path = py_path.with_suffix(\".ipynb\")\n",
     "    out_path.write_text(json.dumps(notebook, indent=1))\n",
     "    return out_path\n",
@@ -783,12 +786,12 @@
     "        print(\"Pipeline will run Bronze → Silver → Gold → Training...\")\n",
     "        subprocess.run([\"python\", \"pipeline_runner.py\"], cwd=str(output_dir.resolve()))\n",
     "    else:\n",
-    "        print(f\"pipeline_runner.py not found. Generate first by running cells above.\")\n",
+    "        print(\"pipeline_runner.py not found. Generate first by running cells above.\")\n",
     "else:\n",
     "    print(\"To run the complete pipeline:\")\n",
     "    print(f\"\\n  cd {output_dir}\")\n",
-    "    print(f\"  python pipeline_runner.py\")\n",
-    "    print(f\"\\nThis will:\")\n",
+    "    print(\"  python pipeline_runner.py\")\n",
+    "    print(\"\\nThis will:\")\n",
     "    print(\"  1. Run Landing layers (event sources)\")\n",
     "    print(\"  2. Run Bronze layers (parallel)\")\n",
     "    print(\"  3. Run Silver merge\")\n",
@@ -916,7 +919,7 @@
     "    print(f\"  - Scalings: {len(registry.gold.scaling) if registry.gold else 0}\")\n",
     "    print(f\"  - Transformations: {len(registry.gold.transformations) if registry.gold else 0}\")\n",
     "    print(f\"  - Feature selections: {len(registry.gold.feature_selection) if registry.gold else 0}\")\n",
-    "    \n",
+    "\n",
     "    # Show what's in each layer for debugging\n",
     "    print()\n",
     "    print(\"Recommendations by layer:\")\n",
@@ -928,13 +931,13 @@
     "                print(f\"    - [{rec.category}] {rec.target_column}: {rec.action}\")\n",
     "            if len(recs) > 3:\n",
     "                print(f\"    ... and {len(recs) - 3} more\")\n",
-    "    \n",
+    "\n",
     "    # Check if gold layer exists but is empty\n",
     "    if registry.gold:\n",
     "        print(f\"\\n✓ Gold layer initialized (target: {registry.gold.target_column})\")\n",
     "    else:\n",
     "        print(\"\\n⚠ Gold layer not initialized - run step 06 first\")\n",
-    "    \n",
+    "\n",
     "    print()\n",
     "    print(\"Use this hash to:\")\n",
     "    print(\"  - Track MLflow experiments (tag: recommendations_hash)\")\n",
@@ -990,6 +993,7 @@
    "source": [
     "# Inspect Feast Feature Store contents\n",
     "import warnings\n",
+    "\n",
     "warnings.filterwarnings(\"ignore\", category=DeprecationWarning, module=\"feast\")\n",
     "\n",
     "feast_repo_path = output_dir / \"feature_repo\"\n",
@@ -998,15 +1002,15 @@
     "    try:\n",
     "        from feast import FeatureStore\n",
     "        store = FeatureStore(repo_path=str(feast_repo_path))\n",
-    "        \n",
+    "\n",
     "        print(\"Feast Feature Store Contents\")\n",
     "        print(\"=\" * 60)\n",
-    "        \n",
+    "\n",
     "        # List entities\n",
     "        entities = store.list_entities()\n",
     "        feature_views = store.list_feature_views()\n",
     "        data_sources = store.list_data_sources()\n",
-    "        \n",
+    "\n",
     "        # Check if registry is empty (feast apply not run yet)\n",
     "        if not entities and not feature_views:\n",
     "            print(\"\\n⚠️  Feature store registry is empty.\")\n",
@@ -1021,7 +1025,7 @@
     "            print(f\"\\n📦 Entities ({len(entities)}):\")\n",
     "            for entity in entities:\n",
     "                print(f\"   - {entity.name} (join_key: {entity.join_keys})\")\n",
-    "            \n",
+    "\n",
     "            print(f\"\\n📊 Feature Views ({len(feature_views)}):\")\n",
     "            for fv in feature_views:\n",
     "                print(f\"   - {fv.name}: {len(fv.features)} features\")\n",
@@ -1029,13 +1033,13 @@
     "                    print(f\"      • {feat.name} ({feat.dtype})\")\n",
     "                if len(fv.features) > 5:\n",
     "                    print(f\"      ... and {len(fv.features) - 5} more\")\n",
-    "            \n",
+    "\n",
     "            print(f\"\\n💾 Data Sources ({len(data_sources)}):\")\n",
     "            for ds in data_sources:\n",
     "                print(f\"   - {ds.name}\")\n",
-    "        \n",
+    "\n",
     "        # Try to show sample data from parquet files\n",
-    "        print(f\"\\n📄 Sample Feature Data:\")\n",
+    "        print(\"\\n📄 Sample Feature Data:\")\n",
     "        data_dir = feast_repo_path / \"data\"\n",
     "        if data_dir.exists():\n",
     "            parquet_files = list(data_dir.glob(\"*.parquet\"))\n",
@@ -1043,14 +1047,14 @@
     "                sample_df = pd.read_parquet(parquet_files[0])\n",
     "                print(f\"   Source: {parquet_files[0].name}\")\n",
     "                print(f\"   Shape: {sample_df.shape[0]:,} rows x {sample_df.shape[1]} columns\")\n",
-    "                print(f\"\\n   Head (first 5 rows):\")\n",
+    "                print(\"\\n   Head (first 5 rows):\")\n",
     "                display(sample_df.head())\n",
     "            else:\n",
     "                print(\"   No parquet files found yet in data/ directory.\")\n",
     "                print(\"   Features will be materialized when you run the pipeline.\")\n",
     "        else:\n",
     "            print(\"   Data directory not created yet.\")\n",
-    "            \n",
+    "\n",
     "    except ImportError:\n",
     "        print(\"Feast not installed. Install with: pip install feast\")\n",
     "    except Exception as e:\n",

{churnkit-0.76.1a1.data → churnkit-0.76.1a2.data}/data/share/churnkit/exploration_notebooks/11_scoring_validation.ipynb

@@ -54,12 +54,11 @@
    "outputs": [],
    "source": [
     "from customer_retention.analysis.notebook_progress import track_and_export_previous\n",
+    "\n",
     "track_and_export_previous(\"11_scoring_validation.ipynb\")\n",
     "\n",
     "import sys\n",
-    "from pathlib import Path\n",
-    "\n",
-    "from customer_retention.core.config.experiments import EXPERIMENTS_DIR, FINDINGS_DIR"
+    "from pathlib import Path\n"
    ]
   },
   {
@@ -95,10 +94,19 @@
     "sys.path.insert(0, str(PIPELINE_DIR))\n",
     "\n",
     "from config import (\n",
-    "    PIPELINE_NAME, TARGET_COLUMN, RECOMMENDATIONS_HASH, MLFLOW_TRACKING_URI,\n",
-    "    FEAST_REPO_PATH, FEAST_FEATURE_VIEW, FEAST_ENTITY_KEY, FEAST_TIMESTAMP_COL,\n",
-    "    PRODUCTION_DIR, EXPERIMENTS_DIR as GEN_EXPERIMENTS_DIR,\n",
-    "    get_feast_data_path, get_gold_path, ARTIFACTS_PATH,\n",
+    "    ARTIFACTS_PATH,\n",
+    "    FEAST_ENTITY_KEY,\n",
+    "    FEAST_FEATURE_VIEW,\n",
+    "    FEAST_REPO_PATH,\n",
+    "    FEAST_TIMESTAMP_COL,\n",
+    "    MLFLOW_TRACKING_URI,\n",
+    "    PIPELINE_NAME,\n",
+    "    PRODUCTION_DIR,\n",
+    "    RECOMMENDATIONS_HASH,\n",
+    "    TARGET_COLUMN,\n",
+    ")\n",
+    "from config import (\n",
+    "    EXPERIMENTS_DIR as GEN_EXPERIMENTS_DIR,\n",
     ")\n",
     "\n",
     "print(f\"Pipeline: {PIPELINE_NAME}\")\n",
@@ -146,18 +154,15 @@
    },
    "outputs": [],
    "source": [
-    "import numpy as np\n",
-    "import pandas as pd\n",
     "import mlflow\n",
     "import mlflow.sklearn\n",
     "import mlflow.xgboost\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
     "import xgboost as xgb\n",
     "from feast import FeatureStore\n",
-    "from customer_retention.transforms import TransformExecutor, ArtifactStore\n",
-    "from customer_retention.generators.pipeline_generator.models import (\n",
-    "    PipelineTransformationType, TransformationStep,\n",
-    ")\n",
-    "from config import EXCLUDED_SOURCES\n",
+    "\n",
+    "from customer_retention.transforms import ArtifactStore, TransformExecutor\n",
     "\n",
     "_registry = ArtifactStore.from_manifest(Path(ARTIFACTS_PATH) / \"manifest.yaml\")\n",
     "_executor = TransformExecutor()\n",
@@ -279,7 +284,7 @@
     "y_pred = (y_proba >= 0.5).astype(int)\n",
     "\n",
     "# --- Metrics ---\n",
-    "from sklearn.metrics import roc_auc_score, accuracy_score, precision_score, recall_score, f1_score\n",
+    "from sklearn.metrics import accuracy_score, f1_score, precision_score, recall_score, roc_auc_score\n",
     "\n",
     "metrics = {\n",
     "    \"accuracy\": accuracy_score(y_true, y_pred),\n",
@@ -346,8 +351,12 @@
    "outputs": [],
    "source": [
     "from sklearn.metrics import (\n",
-    "    accuracy_score, precision_score, recall_score,\n",
-    "    f1_score, roc_auc_score, confusion_matrix,\n",
+    "    accuracy_score,\n",
+    "    confusion_matrix,\n",
+    "    f1_score,\n",
+    "    precision_score,\n",
+    "    recall_score,\n",
+    "    roc_auc_score,\n",
     ")\n",
     "\n",
     "y_true = predictions_df[\"actual\"]\n",
@@ -367,7 +376,7 @@
     "    print(f\"  {name}: {value:.4f}\")\n",
     "\n",
     "cm = confusion_matrix(y_true, y_pred)\n",
-    "print(f\"\\nConfusion Matrix:\")\n",
+    "print(\"\\nConfusion Matrix:\")\n",
     "print(f\"  TN={cm[0,0]:,}  FP={cm[0,1]:,}\")\n",
     "print(f\"  FN={cm[1,0]:,}  TP={cm[1,1]:,}\")"
    ]
@@ -467,12 +476,18 @@
    },
    "outputs": [],
    "source": [
+    "from IPython.display import display\n",
     "from sklearn.metrics import (\n",
-    "    roc_curve, precision_recall_curve, average_precision_score,\n",
-    "    confusion_matrix, roc_auc_score, f1_score, precision_score,\n",
-    "    recall_score, accuracy_score,\n",
+    "    accuracy_score,\n",
+    "    average_precision_score,\n",
+    "    confusion_matrix,\n",
+    "    f1_score,\n",
+    "    precision_recall_curve,\n",
+    "    precision_score,\n",
+    "    recall_score,\n",
+    "    roc_auc_score,\n",
+    "    roc_curve,\n",
     ")\n",
-    "from IPython.display import display, HTML\n",
     "\n",
     "mlflow.set_tracking_uri(MLFLOW_TRACKING_URI)\n",
     "client = mlflow.tracking.MlflowClient()\n",
@@ -1059,7 +1074,7 @@
     ")\n",
     "\n",
     "print(f\"Customer browser ready with {len(browser_df):,} records\")\n",
-    "print(f\"\\nPrediction Distribution:\")\n",
+    "print(\"\\nPrediction Distribution:\")\n",
     "print(f\"  Predicted Positive: {(browser_df['prediction'] == 1).sum():,}\")\n",
     "print(f\"  Predicted Negative: {(browser_df['prediction'] == 0).sum():,}\")\n",
     "print(f\"\\nCorrect Predictions: {browser_df['correct'].sum():,}/{len(browser_df):,} ({browser_df['correct'].mean():.1%})\")"

{churnkit-0.76.1a1.data → churnkit-0.76.1a2.data}/data/share/churnkit/exploration_notebooks/12_view_documentation.ipynb

@@ -44,6 +44,7 @@
    "outputs": [],
    "source": [
     "from customer_retention.analysis.notebook_progress import track_and_export_previous\n",
+    "\n",
     "track_and_export_previous(\"12_view_documentation.ipynb\")"
    ]
   },
@@ -70,8 +71,9 @@
    "outputs": [],
    "source": [
     "from pathlib import Path\n",
-    "from customer_retention.core.config.experiments import get_notebook_experiments_dir\n",
+    "\n",
     "from customer_retention.analysis.notebook_html_exporter import check_exported_html\n",
+    "from customer_retention.core.config.experiments import get_notebook_experiments_dir\n",
     "\n",
     "docs_dir = get_notebook_experiments_dir() / \"docs\"\n",
     "notebook_dir = Path(\"exploration_notebooks\")\n",

{churnkit-0.76.1a1.dist-info → churnkit-0.76.1a2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: churnkit
-Version: 0.76.1a1
+Version: 0.76.1a2
 Summary: Structured ML framework for customer churn prediction -- from exploration notebooks to production pipelines, locally or on Databricks.
 Project-URL: Homepage, https://github.com/aladjov/CR
 Project-URL: Documentation, https://github.com/aladjov/CR/wiki