gengeneeval 0.1.1__tar.gz → 0.2.1__tar.gz

{gengeneeval-0.1.1 → gengeneeval-0.2.1}/PKG-INFO

@@ -1,10 +1,10 @@
 Metadata-Version: 2.4
 Name: gengeneeval
-Version: 0.1.1
-Summary: Comprehensive evaluation of generated gene expression data. Computes metrics between real and generated datasets with support for condition matching, train/test splits, and publication-quality visualizations.
+Version: 0.2.1
+Summary: Comprehensive evaluation of generated gene expression data. Computes metrics between real and generated datasets with support for condition matching, train/test splits, memory-efficient lazy loading, and publication-quality visualizations.
 License: MIT
 License-File: LICENSE
-Keywords: gene expression,evaluation,metrics,single-cell,generative models,benchmarking
+Keywords: gene expression,evaluation,metrics,single-cell,generative models,benchmarking,memory-efficient
 Author: GenEval Team
 Author-email: geneval@example.com
 Requires-Python: >=3.8,<4.0
@@ -42,11 +42,11 @@ Description-Content-Type: text/markdown
 [![PyPI version](https://badge.fury.io/py/gengeneeval.svg)](https://badge.fury.io/py/gengeneeval)
 [![Python 3.8+](https://img.shields.io/badge/python-3.8+-blue.svg)](https://www.python.org/downloads/)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
-[![Tests](https://github.com/AndreaRubbi/GenGeneEval/actions/workflows/tests.yml/badge.svg)](https://github.com/AndreaRubbi/GenGeneEval/actions)
+[![Tests](https://github.com/AndreaRubbi/GenGeneEval/actions/workflows/test.yml/badge.svg)](https://github.com/AndreaRubbi/GenGeneEval/actions)
 
 **Comprehensive evaluation of generated gene expression data against real datasets.**
 
-GenEval is a modular, object-oriented Python framework for computing metrics between real and generated gene expression datasets stored in AnnData (h5ad) format. It supports condition-based matching, train/test splits, and generates publication-quality visualizations.
+GenEval is a modular, object-oriented Python framework for computing metrics between real and generated gene expression datasets stored in AnnData (h5ad) format. It supports condition-based matching, train/test splits, memory-efficient lazy loading for large datasets, and generates publication-quality visualizations.
 
 ## Features
 
@@ -55,6 +55,10 @@ All metrics are computed **per-gene** (returning a vector) and **aggregated**:
 
 | Metric | Description | Direction |
 |--------|-------------|-----------|
+| **MSE** | Mean Squared Error | Lower is better |
+| **RMSE** | Root Mean Squared Error | Lower is better |
+| **MAE** | Mean Absolute Error | Lower is better |
+| **R²** | Coefficient of Determination | Higher is better |
 | **Pearson Correlation** | Linear correlation between expression profiles | Higher is better |
 | **Spearman Correlation** | Rank correlation (robust to outliers) | Higher is better |
 | **Wasserstein-1** | Earth Mover's Distance (L1) | Lower is better |
@@ -73,6 +77,8 @@ All metrics are computed **per-gene** (returning a vector) and **aggregated**:
 - ✅ Condition-based matching (perturbation, cell type, etc.)
 - ✅ Train/test split support
 - ✅ Per-gene and aggregate metrics
+- ✅ **Memory-efficient lazy loading** for large datasets
+- ✅ **Batched evaluation** to avoid OOM errors
 - ✅ Modular, extensible architecture
 - ✅ Command-line interface
 - ✅ Publication-quality visualizations
@@ -136,6 +142,37 @@ geneval --real real.h5ad --generated generated.h5ad \
         --output results/
 ```
 
+### Memory-Efficient Mode (for Large Datasets)
+
+For datasets too large to fit in memory, use the lazy evaluation API:
+
+```python
+from geneval import evaluate_lazy, load_data_lazy
+
+# Memory-efficient evaluation (streams data one condition at a time)
+results = evaluate_lazy(
+    real_path="large_real.h5ad",
+    generated_path="large_generated.h5ad",
+    condition_columns=["perturbation"],
+    batch_size=256,                    # Process in batches
+    use_backed=True,                   # Memory-mapped file access
+    output_dir="eval_output/",
+    save_per_condition=True,           # Save each condition to disk
+)
+
+# Get summary statistics
+print(results.get_summary())
+
+# Or use the lazy loader directly for custom workflows
+with load_data_lazy("real.h5ad", "gen.h5ad", ["perturbation"]) as loader:
+    print(f"Memory estimate: {loader.estimate_memory_usage()}")
+    
+    # Process one condition at a time
+    for key, real, gen, info in loader.iterate_conditions():
+        # Your custom evaluation logic
+        pass
+```
+
 ## Expected Data Format
 
 GenEval expects AnnData (h5ad) files with:

{gengeneeval-0.1.1 → gengeneeval-0.2.1}/README.md

@@ -3,11 +3,11 @@
 [![PyPI version](https://badge.fury.io/py/gengeneeval.svg)](https://badge.fury.io/py/gengeneeval)
 [![Python 3.8+](https://img.shields.io/badge/python-3.8+-blue.svg)](https://www.python.org/downloads/)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
-[![Tests](https://github.com/AndreaRubbi/GenGeneEval/actions/workflows/tests.yml/badge.svg)](https://github.com/AndreaRubbi/GenGeneEval/actions)
+[![Tests](https://github.com/AndreaRubbi/GenGeneEval/actions/workflows/test.yml/badge.svg)](https://github.com/AndreaRubbi/GenGeneEval/actions)
 
 **Comprehensive evaluation of generated gene expression data against real datasets.**
 
-GenEval is a modular, object-oriented Python framework for computing metrics between real and generated gene expression datasets stored in AnnData (h5ad) format. It supports condition-based matching, train/test splits, and generates publication-quality visualizations.
+GenEval is a modular, object-oriented Python framework for computing metrics between real and generated gene expression datasets stored in AnnData (h5ad) format. It supports condition-based matching, train/test splits, memory-efficient lazy loading for large datasets, and generates publication-quality visualizations.
 
 ## Features
 
@@ -16,6 +16,10 @@ All metrics are computed **per-gene** (returning a vector) and **aggregated**:
 
 | Metric | Description | Direction |
 |--------|-------------|-----------|
+| **MSE** | Mean Squared Error | Lower is better |
+| **RMSE** | Root Mean Squared Error | Lower is better |
+| **MAE** | Mean Absolute Error | Lower is better |
+| **R²** | Coefficient of Determination | Higher is better |
 | **Pearson Correlation** | Linear correlation between expression profiles | Higher is better |
 | **Spearman Correlation** | Rank correlation (robust to outliers) | Higher is better |
 | **Wasserstein-1** | Earth Mover's Distance (L1) | Lower is better |
@@ -34,6 +38,8 @@ All metrics are computed **per-gene** (returning a vector) and **aggregated**:
 - ✅ Condition-based matching (perturbation, cell type, etc.)
 - ✅ Train/test split support
 - ✅ Per-gene and aggregate metrics
+- ✅ **Memory-efficient lazy loading** for large datasets
+- ✅ **Batched evaluation** to avoid OOM errors
 - ✅ Modular, extensible architecture
 - ✅ Command-line interface
 - ✅ Publication-quality visualizations
@@ -97,6 +103,37 @@ geneval --real real.h5ad --generated generated.h5ad \
         --output results/
 ```
 
+### Memory-Efficient Mode (for Large Datasets)
+
+For datasets too large to fit in memory, use the lazy evaluation API:
+
+```python
+from geneval import evaluate_lazy, load_data_lazy
+
+# Memory-efficient evaluation (streams data one condition at a time)
+results = evaluate_lazy(
+    real_path="large_real.h5ad",
+    generated_path="large_generated.h5ad",
+    condition_columns=["perturbation"],
+    batch_size=256,                    # Process in batches
+    use_backed=True,                   # Memory-mapped file access
+    output_dir="eval_output/",
+    save_per_condition=True,           # Save each condition to disk
+)
+
+# Get summary statistics
+print(results.get_summary())
+
+# Or use the lazy loader directly for custom workflows
+with load_data_lazy("real.h5ad", "gen.h5ad", ["perturbation"]) as loader:
+    print(f"Memory estimate: {loader.estimate_memory_usage()}")
+    
+    # Process one condition at a time
+    for key, real, gen, info in loader.iterate_conditions():
+        # Your custom evaluation logic
+        pass
+```
+
 ## Expected Data Format
 
 GenEval expects AnnData (h5ad) files with:

{gengeneeval-0.1.1 → gengeneeval-0.2.1}/pyproject.toml

@@ -1,13 +1,13 @@
 [tool.poetry]
 name = "gengeneeval"
-version = "0.1.1"
-description = "Comprehensive evaluation of generated gene expression data. Computes metrics between real and generated datasets with support for condition matching, train/test splits, and publication-quality visualizations."
+version = "0.2.1"
+description = "Comprehensive evaluation of generated gene expression data. Computes metrics between real and generated datasets with support for condition matching, train/test splits, memory-efficient lazy loading, and publication-quality visualizations."
 authors = ["GenEval Team <geneval@example.com>"]
 license = "MIT"
 readme = "README.md"
 homepage = "https://github.com/AndreaRubbi/GenGeneEval"
 repository = "https://github.com/AndreaRubbi/GenGeneEval"
-keywords = ["gene expression", "evaluation", "metrics", "single-cell", "generative models", "benchmarking"]
+keywords = ["gene expression", "evaluation", "metrics", "single-cell", "generative models", "benchmarking", "memory-efficient"]
 classifiers = [
     "Development Status :: 4 - Beta",
     "Intended Audience :: Science/Research",

{gengeneeval-0.1.1 → gengeneeval-0.2.1}/src/geneval/__init__.py

@@ -8,6 +8,7 @@ Features:
 - Multiple distance and correlation metrics (per-gene and aggregate)
 - Condition-based matching (perturbation, cell type, etc.)
 - Train/test split support
+- Memory-efficient lazy loading for large datasets
 - Publication-quality visualizations
 - Command-line interface
 
@@ -20,12 +21,22 @@ Quick Start:
     ...     output_dir="output/"
     ... )
 
+Memory-Efficient Mode (for large datasets):
+    >>> from geneval import evaluate_lazy
+    >>> results = evaluate_lazy(
+    ...     real_path="real.h5ad",
+    ...     generated_path="generated.h5ad",
+    ...     condition_columns=["perturbation"],
+    ...     batch_size=256,
+    ...     use_backed=True,  # Memory-mapped access
+    ... )
+
 CLI Usage:
     $ geneval --real real.h5ad --generated generated.h5ad \\
               --conditions perturbation cell_type --output results/
 """
 
-__version__ = "0.1.1"
+__version__ = "0.2.1"
 __author__ = "GenEval Team"
 
 # Main evaluation interface
@@ -35,12 +46,26 @@ from .evaluator import (
     MetricRegistry,
 )
 
+# Memory-efficient evaluation
+from .lazy_evaluator import (
+    evaluate_lazy,
+    MemoryEfficientEvaluator,
+    StreamingEvaluationResult,
+)
+
 # Data loading
 from .data.loader import (
     GeneExpressionDataLoader,
     load_data,
 )
 
+# Memory-efficient data loading
+from .data.lazy_loader import (
+    LazyGeneExpressionDataLoader,
+    load_data_lazy,
+    ConditionBatch,
+)
+
 # Results
 from .results import (
     EvaluationResult,
@@ -69,6 +94,12 @@ from .metrics.distances import (
     MultivariateWasserstein,
     MultivariateMMD,
 )
+from .metrics.reconstruction import (
+    MSEDistance,
+    RMSEDistance,
+    MAEDistance,
+    R2Score,
+)
 
 # Visualization
 from .visualization.visualizer import (
@@ -93,9 +124,17 @@ __all__ = [
     "evaluate",
     "GeneEvalEvaluator",
     "MetricRegistry",
+    # Memory-efficient evaluation
+    "evaluate_lazy",
+    "MemoryEfficientEvaluator",
+    "StreamingEvaluationResult",
     # Data loading
     "GeneExpressionDataLoader",
     "load_data",
+    # Memory-efficient data loading
+    "LazyGeneExpressionDataLoader",
+    "load_data_lazy",
+    "ConditionBatch",
     # Results
     "EvaluationResult",
     "SplitResult", 
@@ -117,6 +156,11 @@ __all__ = [
     "EnergyDistance",
     "MultivariateWasserstein",
     "MultivariateMMD",
+    # Reconstruction metrics
+    "MSEDistance",
+    "RMSEDistance",
+    "MAEDistance",
+    "R2Score",
     # Visualization
     "EvaluationVisualizer",
     "visualize",

{gengeneeval-0.1.1 → gengeneeval-0.2.1}/src/geneval/data/__init__.py

@@ -2,6 +2,7 @@
 Data loading module for gene expression evaluation.
 
 Provides data loaders for paired real and generated datasets.
+Includes both standard and memory-efficient lazy loading options.
 """
 
 from .loader import (
@@ -9,15 +10,28 @@ from .loader import (
     load_data,
     DataLoaderError,
 )
+from .lazy_loader import (
+    LazyGeneExpressionDataLoader,
+    load_data_lazy,
+    LazyDataLoaderError,
+    ConditionBatch,
+)
 from .gene_expression_datamodule import (
     GeneExpressionDataModule,
     DataModuleError,
 )
 
 __all__ = [
+    # Standard loader
     "GeneExpressionDataLoader",
     "load_data",
     "DataLoaderError",
+    # Lazy loader (memory-efficient)
+    "LazyGeneExpressionDataLoader",
+    "load_data_lazy",
+    "LazyDataLoaderError",
+    "ConditionBatch",
+    # DataModule
     "GeneExpressionDataModule",
     "DataModuleError",
 ]