Add UMAP PaperAgent MCP server

- MCP server with 4 tools: umap_reduce_dimensions, umap_parameter_sweep,
umap_compare_metrics, umap_with_labels
- Docker deployment for HuggingFace Spaces
- Sample data (iris, digits datasets)
- Based on McInnes et al. (2018) arXiv:1802.03426

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Dockerfile

@@ -0,0 +1,29 @@
+FROM python:3.10-slim
+
+WORKDIR /app
+
+# Install system dependencies
+RUN apt-get update && apt-get install -y \
+    build-essential \
+    && rm -rf /var/lib/apt/lists/*
+
+# Copy requirements and install
+COPY mcp/requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+RUN pip install --no-cache-dir uvicorn starlette
+
+# Copy application code
+COPY . .
+
+# Create data directories
+RUN mkdir -p /data/upload /data/tmp_outputs
+
+# Set environment variables
+ENV UMAP_OUTPUT_DIR=/data/tmp_outputs
+ENV UMAP_INPUT_DIR=/data/upload
+
+# Expose port
+EXPOSE 7860
+
+# Run the application
+CMD ["python", "app.py"]

README.md

@@ -1,12 +1,38 @@
 ---
-title: Umap Agent
-emoji: 🐠
-colorFrom: purple
-colorTo: green
-sdk: gradio
-sdk_version: 6.5.0
-app_file: app.py
-pinned: false
+title: UMAP PaperAgent
+emoji: 🗺️
+colorFrom: blue
+colorTo: purple
+sdk: docker
+app_port: 7860
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# UMAP PaperAgent
+
+A PaperAgent MCP server for the UMAP dimensionality reduction algorithm.
+
+## Paper Reference
+
+McInnes, L., Healy, J., & Melville, J. (2018). **UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction**. [arXiv:1802.03426](https://arxiv.org/abs/1802.03426)
+
+## Available MCP Tools
+
+| Tool | Description |
+|------|-------------|
+| `umap_reduce_dimensions` | Basic UMAP dimensionality reduction |
+| `umap_parameter_sweep` | Grid of UMAP results with different parameters |
+| `umap_compare_metrics` | Compare distance metrics |
+| `umap_with_labels` | UMAP with colored class labels |
+
+## Usage with Claude Code
+
+```bash
+claude mcp add --transport http umap "https://EtashGuha-umap-agent.hf.space/mcp"
+```
+
+## API Endpoints
+
+- `GET /` - Server info
+- `POST /upload` - Upload data files
+- `GET /outputs/{filename}` - Download results
+- `POST /mcp` - MCP protocol endpoint

app.py

@@ -0,0 +1,99 @@
+"""
+UMAP PaperAgent - HuggingFace Spaces Deployment
+
+Based on: McInnes, L., Healy, J., & Melville, J. (2018).
+UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction.
+arXiv:1802.03426
+"""
+
+import os
+import uuid
+from pathlib import Path
+from starlette.applications import Starlette
+from starlette.routing import Route, Mount
+from starlette.responses import JSONResponse, FileResponse
+from starlette.requests import Request
+from fastmcp.server.http import create_sse_app
+
+# Import the MCP server
+import sys
+sys.path.insert(0, str(Path(__file__).parent / "mcp"))
+from umap_mcp import mcp
+
+# Setup directories
+UPLOAD_DIR = Path("/data/upload")
+OUTPUT_DIR = Path("/data/tmp_outputs")
+UPLOAD_DIR.mkdir(parents=True, exist_ok=True)
+OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
+
+# Update MCP paths
+os.environ["UMAP_OUTPUT_DIR"] = str(OUTPUT_DIR)
+os.environ["UMAP_INPUT_DIR"] = str(UPLOAD_DIR)
+
+
+async def root(request: Request):
+    """Root endpoint with API info."""
+    return JSONResponse({
+        "name": "UMAP PaperAgent",
+        "paper": "McInnes et al. (2018) - UMAP: Uniform Manifold Approximation and Projection",
+        "arxiv": "https://arxiv.org/abs/1802.03426",
+        "endpoints": {
+            "/mcp": "MCP protocol endpoint",
+            "/upload": "Upload data files",
+            "/outputs/{filename}": "Download output files"
+        },
+        "tools": [
+            "umap_reduce_dimensions",
+            "umap_parameter_sweep",
+            "umap_compare_metrics",
+            "umap_with_labels"
+        ]
+    })
+
+
+async def upload_file(request: Request):
+    """Upload a file to the server."""
+    form = await request.form()
+    file = form.get("file")
+    if not file:
+        return JSONResponse({"error": "No file provided"}, status_code=400)
+
+    file_id = uuid.uuid4().hex
+    ext = Path(file.filename).suffix if file.filename else ".txt"
+    save_path = UPLOAD_DIR / f"{file_id}{ext}"
+
+    content = await file.read()
+    with open(save_path, "wb") as f:
+        f.write(content)
+
+    return JSONResponse({"path": str(save_path)})
+
+
+async def get_output(request: Request):
+    """Download an output file."""
+    filename = request.path_params["filename"]
+    file_path = OUTPUT_DIR / filename
+    if file_path.exists():
+        return FileResponse(file_path)
+    return JSONResponse({"error": "File not found"}, status_code=404)
+
+
+# Custom routes
+custom_routes = [
+    Route("/", root),
+    Route("/upload", upload_file, methods=["POST"]),
+    Route("/outputs/{filename}", get_output),
+]
+
+# Create the MCP SSE app
+app = create_sse_app(
+    mcp,
+    message_path="/mcp",
+    sse_path="/mcp/sse",
+    routes=custom_routes
+)
+
+
+if __name__ == "__main__":
+    import uvicorn
+    uvicorn.run(app, host="0.0.0.0", port=7860)

mcp/requirements.txt

@@ -0,0 +1,6 @@
+fastmcp>=2.0.0
+umap-learn>=0.5.0
+pandas>=1.5.0
+numpy>=1.21.0
+matplotlib>=3.5.0
+scikit-learn>=1.0.0

mcp/umap_mcp.py

@@ -0,0 +1,378 @@
+"""
+UMAP PaperAgent MCP Server
+
+Based on: McInnes, L., Healy, J., & Melville, J. (2018).
+UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction.
+arXiv:1802.03426
+
+This MCP server provides tools for dimensionality reduction and visualization
+using the UMAP algorithm, following the Paper2Agent framework.
+"""
+
+from fastmcp import FastMCP
+from typing import Annotated, Literal, Optional, List
+import os
+from pathlib import Path
+from datetime import datetime
+
+# Create MCP server
+mcp = FastMCP(name="UMAP")
+
+# Project paths
+PROJECT_ROOT = Path(__file__).parent.parent.resolve()
+DEFAULT_OUTPUT_DIR = PROJECT_ROOT / "tmp" / "outputs"
+DEFAULT_INPUT_DIR = PROJECT_ROOT / "tmp" / "inputs"
+
+OUTPUT_DIR = Path(os.environ.get("UMAP_OUTPUT_DIR", DEFAULT_OUTPUT_DIR))
+INPUT_DIR = Path(os.environ.get("UMAP_INPUT_DIR", DEFAULT_INPUT_DIR))
+
+OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
+INPUT_DIR.mkdir(parents=True, exist_ok=True)
+
+timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+
+
+@mcp.tool()
+def umap_reduce_dimensions(
+    data_path: Annotated[str, "Path to input data file (CSV format with samples as rows, features as columns)"],
+    n_neighbors: Annotated[int, "Number of neighbors for local structure (default: 15). Higher values capture more global structure."] = 15,
+    min_dist: Annotated[float, "Minimum distance between points in embedding (default: 0.1). Lower values create tighter clusters."] = 0.1,
+    n_components: Annotated[int, "Number of dimensions in output embedding (default: 2)"] = 2,
+    metric: Annotated[Literal["euclidean", "manhattan", "cosine", "correlation"], "Distance metric to use"] = "euclidean",
+    random_state: Annotated[int, "Random seed for reproducibility"] = 42,
+    out_prefix: Annotated[Optional[str], "Output file prefix"] = None,
+) -> dict:
+    """
+    Perform UMAP dimensionality reduction on high-dimensional data.
+
+    Input is a CSV file with samples as rows and features as columns.
+    Output is the UMAP embedding coordinates and a visualization.
+
+    Reference: McInnes et al. (2018) arXiv:1802.03426
+    """
+    import pandas as pd
+    import numpy as np
+    import umap
+    import matplotlib
+    matplotlib.use('Agg')
+    import matplotlib.pyplot as plt
+
+    # Load data
+    if not os.path.exists(data_path):
+        return {"error": f"Data file not found: {data_path}"}
+
+    df = pd.read_csv(data_path, index_col=0)
+    X = df.values
+
+    # Run UMAP
+    reducer = umap.UMAP(
+        n_neighbors=n_neighbors,
+        min_dist=min_dist,
+        n_components=n_components,
+        metric=metric,
+        random_state=random_state
+    )
+
+    embedding = reducer.fit_transform(X)
+
+    # Create output prefix
+    if out_prefix is None:
+        out_prefix = f"umap_{timestamp}"
+
+    # Save embedding
+    embedding_df = pd.DataFrame(
+        embedding,
+        columns=[f"UMAP{i+1}" for i in range(n_components)],
+        index=df.index
+    )
+    embedding_path = OUTPUT_DIR / f"{out_prefix}_embedding.csv"
+    embedding_df.to_csv(embedding_path)
+
+    # Create visualization (if 2D)
+    viz_path = None
+    if n_components == 2:
+        fig, ax = plt.subplots(figsize=(10, 8))
+        scatter = ax.scatter(embedding[:, 0], embedding[:, 1],
+                           c=range(len(embedding)), cmap='viridis',
+                           alpha=0.6, s=10)
+        ax.set_xlabel('UMAP1', fontsize=12)
+        ax.set_ylabel('UMAP2', fontsize=12)
+        ax.set_title(f'UMAP Embedding\n(n_neighbors={n_neighbors}, min_dist={min_dist}, metric={metric})',
+                    fontsize=14)
+        plt.colorbar(scatter, ax=ax, label='Sample index')
+
+        viz_path = OUTPUT_DIR / f"{out_prefix}_visualization.png"
+        plt.savefig(viz_path, dpi=150, bbox_inches='tight')
+        plt.close()
+
+    artifacts = [
+        {"description": "UMAP embedding coordinates", "path": str(embedding_path)}
+    ]
+    if viz_path:
+        artifacts.append({"description": "UMAP visualization", "path": str(viz_path)})
+
+    return {
+        "message": f"UMAP reduction completed: {X.shape[0]} samples, {X.shape[1]} features -> {n_components} dimensions",
+        "reference": "https://arxiv.org/abs/1802.03426",
+        "parameters": {
+            "n_neighbors": n_neighbors,
+            "min_dist": min_dist,
+            "n_components": n_components,
+            "metric": metric
+        },
+        "artifacts": artifacts
+    }
+
+
+@mcp.tool()
+def umap_parameter_sweep(
+    data_path: Annotated[str, "Path to input data file (CSV format)"],
+    n_neighbors_values: Annotated[List[int], "List of n_neighbors values to test"] = [5, 15, 50],
+    min_dist_values: Annotated[List[float], "List of min_dist values to test"] = [0.0, 0.1, 0.5],
+    metric: Annotated[Literal["euclidean", "manhattan", "cosine", "correlation"], "Distance metric"] = "euclidean",
+    random_state: Annotated[int, "Random seed"] = 42,
+    out_prefix: Annotated[Optional[str], "Output file prefix"] = None,
+) -> dict:
+    """
+    Perform parameter sweep to visualize how UMAP parameters affect the embedding.
+
+    Creates a grid of UMAP visualizations with different n_neighbors and min_dist values.
+    This helps understand the paper's key insight about balancing local vs global structure.
+
+    Reference: McInnes et al. (2018) arXiv:1802.03426
+    """
+    import pandas as pd
+    import numpy as np
+    import umap
+    import matplotlib
+    matplotlib.use('Agg')
+    import matplotlib.pyplot as plt
+
+    if not os.path.exists(data_path):
+        return {"error": f"Data file not found: {data_path}"}
+
+    df = pd.read_csv(data_path, index_col=0)
+    X = df.values
+
+    if out_prefix is None:
+        out_prefix = f"umap_sweep_{timestamp}"
+
+    n_rows = len(n_neighbors_values)
+    n_cols = len(min_dist_values)
+
+    fig, axes = plt.subplots(n_rows, n_cols, figsize=(4*n_cols, 4*n_rows))
+    if n_rows == 1:
+        axes = axes.reshape(1, -1)
+    if n_cols == 1:
+        axes = axes.reshape(-1, 1)
+
+    for i, n_neighbors in enumerate(n_neighbors_values):
+        for j, min_dist in enumerate(min_dist_values):
+            reducer = umap.UMAP(
+                n_neighbors=n_neighbors,
+                min_dist=min_dist,
+                n_components=2,
+                metric=metric,
+                random_state=random_state
+            )
+            embedding = reducer.fit_transform(X)
+
+            ax = axes[i, j]
+            ax.scatter(embedding[:, 0], embedding[:, 1],
+                      c=range(len(embedding)), cmap='viridis',
+                      alpha=0.6, s=5)
+            ax.set_title(f'n_neighbors={n_neighbors}\nmin_dist={min_dist}', fontsize=10)
+            ax.set_xticks([])
+            ax.set_yticks([])
+
+    fig.suptitle(f'UMAP Parameter Sweep (metric={metric})\n'
+                f'Top to bottom: increasing n_neighbors (more global)\n'
+                f'Left to right: increasing min_dist (more spread)',
+                fontsize=12, y=1.02)
+    plt.tight_layout()
+
+    viz_path = OUTPUT_DIR / f"{out_prefix}_parameter_sweep.png"
+    plt.savefig(viz_path, dpi=150, bbox_inches='tight')
+    plt.close()
+
+    return {
+        "message": f"Parameter sweep completed: {n_rows}x{n_cols} grid",
+        "reference": "https://arxiv.org/abs/1802.03426",
+        "parameters_tested": {
+            "n_neighbors": n_neighbors_values,
+            "min_dist": min_dist_values,
+            "metric": metric
+        },
+        "artifacts": [
+            {"description": "Parameter sweep visualization", "path": str(viz_path)}
+        ]
+    }
+
+
+@mcp.tool()
+def umap_compare_metrics(
+    data_path: Annotated[str, "Path to input data file (CSV format)"],
+    metrics: Annotated[List[str], "List of distance metrics to compare"] = ["euclidean", "cosine", "manhattan", "correlation"],
+    n_neighbors: Annotated[int, "Number of neighbors"] = 15,
+    min_dist: Annotated[float, "Minimum distance"] = 0.1,
+    random_state: Annotated[int, "Random seed"] = 42,
+    out_prefix: Annotated[Optional[str], "Output file prefix"] = None,
+) -> dict:
+    """
+    Compare different distance metrics for UMAP embedding.
+
+    Visualizes how the choice of distance metric affects the embedding structure.
+    This demonstrates the paper's flexibility in supporting various metrics.
+
+    Reference: McInnes et al. (2018) arXiv:1802.03426
+    """
+    import pandas as pd
+    import numpy as np
+    import umap
+    import matplotlib
+    matplotlib.use('Agg')
+    import matplotlib.pyplot as plt
+
+    if not os.path.exists(data_path):
+        return {"error": f"Data file not found: {data_path}"}
+
+    df = pd.read_csv(data_path, index_col=0)
+    X = df.values
+
+    if out_prefix is None:
+        out_prefix = f"umap_metrics_{timestamp}"
+
+    n_metrics = len(metrics)
+    fig, axes = plt.subplots(1, n_metrics, figsize=(4*n_metrics, 4))
+    if n_metrics == 1:
+        axes = [axes]
+
+    for i, metric in enumerate(metrics):
+        try:
+            reducer = umap.UMAP(
+                n_neighbors=n_neighbors,
+                min_dist=min_dist,
+                n_components=2,
+                metric=metric,
+                random_state=random_state
+            )
+            embedding = reducer.fit_transform(X)
+
+            axes[i].scatter(embedding[:, 0], embedding[:, 1],
+                          c=range(len(embedding)), cmap='viridis',
+                          alpha=0.6, s=10)
+            axes[i].set_title(f'Metric: {metric}', fontsize=12)
+            axes[i].set_xlabel('UMAP1')
+            axes[i].set_ylabel('UMAP2')
+        except Exception as e:
+            axes[i].text(0.5, 0.5, f'Error: {str(e)[:30]}',
+                        ha='center', va='center', transform=axes[i].transAxes)
+            axes[i].set_title(f'Metric: {metric} (failed)', fontsize=12)
+
+    fig.suptitle(f'UMAP Distance Metric Comparison\n(n_neighbors={n_neighbors}, min_dist={min_dist})',
+                fontsize=14)
+    plt.tight_layout()
+
+    viz_path = OUTPUT_DIR / f"{out_prefix}_metric_comparison.png"
+    plt.savefig(viz_path, dpi=150, bbox_inches='tight')
+    plt.close()
+
+    return {
+        "message": f"Metric comparison completed for {n_metrics} metrics",
+        "reference": "https://arxiv.org/abs/1802.03426",
+        "metrics_compared": metrics,
+        "artifacts": [
+            {"description": "Metric comparison visualization", "path": str(viz_path)}
+        ]
+    }
+
+
+@mcp.tool()
+def umap_with_labels(
+    data_path: Annotated[str, "Path to input data file (CSV format)"],
+    labels_path: Annotated[str, "Path to labels file (CSV with 'label' column)"],
+    n_neighbors: Annotated[int, "Number of neighbors"] = 15,
+    min_dist: Annotated[float, "Minimum distance"] = 0.1,
+    metric: Annotated[Literal["euclidean", "manhattan", "cosine", "correlation"], "Distance metric"] = "euclidean",
+    random_state: Annotated[int, "Random seed"] = 42,
+    out_prefix: Annotated[Optional[str], "Output file prefix"] = None,
+) -> dict:
+    """
+    Perform UMAP with colored labels to assess cluster separation.
+
+    Useful for evaluating how well UMAP preserves known class structure.
+
+    Reference: McInnes et al. (2018) arXiv:1802.03426
+    """
+    import pandas as pd
+    import numpy as np
+    import umap
+    import matplotlib
+    matplotlib.use('Agg')
+    import matplotlib.pyplot as plt
+
+    if not os.path.exists(data_path):
+        return {"error": f"Data file not found: {data_path}"}
+    if not os.path.exists(labels_path):
+        return {"error": f"Labels file not found: {labels_path}"}
+
+    df = pd.read_csv(data_path, index_col=0)
+    labels_df = pd.read_csv(labels_path, index_col=0)
+
+    X = df.values
+    labels = labels_df['label'].values if 'label' in labels_df.columns else labels_df.iloc[:, 0].values
+
+    if out_prefix is None:
+        out_prefix = f"umap_labeled_{timestamp}"
+
+    reducer = umap.UMAP(
+        n_neighbors=n_neighbors,
+        min_dist=min_dist,
+        n_components=2,
+        metric=metric,
+        random_state=random_state
+    )
+    embedding = reducer.fit_transform(X)
+
+    # Save embedding with labels
+    embedding_df = pd.DataFrame(
+        embedding,
+        columns=['UMAP1', 'UMAP2'],
+        index=df.index
+    )
+    embedding_df['label'] = labels
+    embedding_path = OUTPUT_DIR / f"{out_prefix}_embedding.csv"
+    embedding_df.to_csv(embedding_path)
+
+    # Create visualization
+    fig, ax = plt.subplots(figsize=(10, 8))
+    unique_labels = np.unique(labels)
+    colors = plt.cm.tab10(np.linspace(0, 1, len(unique_labels)))
+
+    for label, color in zip(unique_labels, colors):
+        mask = labels == label
+        ax.scatter(embedding[mask, 0], embedding[mask, 1],
+                  c=[color], label=str(label), alpha=0.6, s=15)
+
+    ax.set_xlabel('UMAP1', fontsize=12)
+    ax.set_ylabel('UMAP2', fontsize=12)
+    ax.set_title(f'UMAP Embedding with Labels\n(n_neighbors={n_neighbors}, min_dist={min_dist})', fontsize=14)
+    ax.legend(title='Label', bbox_to_anchor=(1.05, 1), loc='upper left')
+
+    viz_path = OUTPUT_DIR / f"{out_prefix}_visualization.png"
+    plt.savefig(viz_path, dpi=150, bbox_inches='tight')
+    plt.close()
+
+    return {
+        "message": f"UMAP with labels completed: {len(unique_labels)} unique labels",
+        "reference": "https://arxiv.org/abs/1802.03426",
+        "labels_found": list(map(str, unique_labels)),
+        "artifacts": [
+            {"description": "UMAP embedding with labels", "path": str(embedding_path)},
+            {"description": "Labeled UMAP visualization", "path": str(viz_path)}
+        ]
+    }
+
+
+if __name__ == "__main__":
+    mcp.run()

tmp/inputs/digits_labels.csv

@@ -0,0 +1,1798 @@
+,label
+digit_0,0
+digit_1,1
+digit_2,2
+digit_3,3
+digit_4,4
+digit_5,5
+digit_6,6
+digit_7,7
+digit_8,8
+digit_9,9
+digit_10,0
+digit_11,1
+digit_12,2
+digit_13,3
+digit_14,4
+digit_15,5
+digit_16,6
+digit_17,7
+digit_18,8
+digit_19,9
+digit_20,0
+digit_21,1
+digit_22,2
+digit_23,3
+digit_24,4
+digit_25,5
+digit_26,6
+digit_27,7
+digit_28,8
+digit_29,9
+digit_30,0
+digit_31,9
+digit_32,5
+digit_33,5
+digit_34,6
+digit_35,5
+digit_36,0
+digit_37,9
+digit_38,8
+digit_39,9
+digit_40,8
+digit_41,4
+digit_42,1
+digit_43,7
+digit_44,7
+digit_45,3
+digit_46,5
+digit_47,1
+digit_48,0
+digit_49,0
+digit_50,2
+digit_51,2
+digit_52,7
+digit_53,8
+digit_54,2
+digit_55,0
+digit_56,1
+digit_57,2
+digit_58,6
+digit_59,3
+digit_60,3
+digit_61,7
+digit_62,3
+digit_63,3
+digit_64,4
+digit_65,6
+digit_66,6
+digit_67,6
+digit_68,4
+digit_69,9
+digit_70,1
+digit_71,5
+digit_72,0
+digit_73,9
+digit_74,5
+digit_75,2
+digit_76,8
+digit_77,2
+digit_78,0
+digit_79,0
+digit_80,1
+digit_81,7
+digit_82,6
+digit_83,3
+digit_84,2
+digit_85,1
+digit_86,7
+digit_87,4
+digit_88,6
+digit_89,3
+digit_90,1
+digit_91,3
+digit_92,9
+digit_93,1
+digit_94,7
+digit_95,6
+digit_96,8
+digit_97,4
+digit_98,3
+digit_99,1
+digit_100,4
+digit_101,0
+digit_102,5
+digit_103,3
+digit_104,6
+digit_105,9
+digit_106,6
+digit_107,1
+digit_108,7
+digit_109,5
+digit_110,4
+digit_111,4
+digit_112,7
+digit_113,2
+digit_114,8
+digit_115,2
+digit_116,2
+digit_117,5
+digit_118,7
+digit_119,9
+digit_120,5
+digit_121,4
+digit_122,8
+digit_123,8
+digit_124,4
+digit_125,9
+digit_126,0
+digit_127,8
+digit_128,9
+digit_129,8
+digit_130,0
+digit_131,1
+digit_132,2
+digit_133,3
+digit_134,4
+digit_135,5
+digit_136,6
+digit_137,7
+digit_138,8
+digit_139,9
+digit_140,0
+digit_141,1
+digit_142,2
+digit_143,3
+digit_144,4
+digit_145,5
+digit_146,6
+digit_147,7
+digit_148,8
+digit_149,9
+digit_150,0
+digit_151,1
+digit_152,2
+digit_153,3
+digit_154,4
+digit_155,5
+digit_156,6
+digit_157,7
+digit_158,8
+digit_159,9
+digit_160,0
+digit_161,9
+digit_162,5
+digit_163,5
+digit_164,6
+digit_165,5
+digit_166,0
+digit_167,9
+digit_168,8
+digit_169,9
+digit_170,8
+digit_171,4
+digit_172,1
+digit_173,7
+digit_174,7
+digit_175,3
+digit_176,5
+digit_177,1
+digit_178,0
+digit_179,0
+digit_180,2
+digit_181,2
+digit_182,7
+digit_183,8
+digit_184,2
+digit_185,0
+digit_186,1
+digit_187,2
+digit_188,6
+digit_189,3
+digit_190,3
+digit_191,7
+digit_192,3
+digit_193,3
+digit_194,4
+digit_195,6
+digit_196,6
+digit_197,6
+digit_198,4
+digit_199,9
+digit_200,1
+digit_201,5
+digit_202,0
+digit_203,9
+digit_204,5
+digit_205,2
+digit_206,8
+digit_207,2
+digit_208,0
+digit_209,0
+digit_210,1
+digit_211,7
+digit_212,6
+digit_213,3
+digit_214,2
+digit_215,1
+digit_216,7
+digit_217,3
+digit_218,1
+digit_219,3
+digit_220,9
+digit_221,1
+digit_222,7
+digit_223,6
+digit_224,8
+digit_225,4
+digit_226,3
+digit_227,1
+digit_228,4
+digit_229,0
+digit_230,5
+digit_231,3
+digit_232,6
+digit_233,9
+digit_234,6
+digit_235,1
+digit_236,7
+digit_237,5
+digit_238,4
+digit_239,4
+digit_240,7
+digit_241,2
+digit_242,8
+digit_243,2
+digit_244,2
+digit_245,5
+digit_246,5
+digit_247,4
+digit_248,8
+digit_249,8
+digit_250,4
+digit_251,9
+digit_252,0
+digit_253,8
+digit_254,9
+digit_255,8
+digit_256,0
+digit_257,1
+digit_258,2
+digit_259,3
+digit_260,4
+digit_261,5
+digit_262,6
+digit_263,7
+digit_264,8
+digit_265,9
+digit_266,0
+digit_267,1
+digit_268,2
+digit_269,3
+digit_270,4
+digit_271,5
+digit_272,6
+digit_273,7
+digit_274,8
+digit_275,9
+digit_276,0
+digit_277,1
+digit_278,2
+digit_279,3
+digit_280,4
+digit_281,5
+digit_282,6
+digit_283,7
+digit_284,8
+digit_285,9
+digit_286,0
+digit_287,9
+digit_288,5
+digit_289,5
+digit_290,6
+digit_291,5
+digit_292,0
+digit_293,9
+digit_294,8
+digit_295,9
+digit_296,8
+digit_297,4
+digit_298,1
+digit_299,7
+digit_300,7
+digit_301,3
+digit_302,5
+digit_303,1
+digit_304,0
+digit_305,0
+digit_306,2
+digit_307,2
+digit_308,7
+digit_309,8
+digit_310,2
+digit_311,0
+digit_312,1
+digit_313,2
+digit_314,6
+digit_315,3
+digit_316,3
+digit_317,7
+digit_318,3
+digit_319,3
+digit_320,4
+digit_321,6
+digit_322,6
+digit_323,6
+digit_324,4
+digit_325,9
+digit_326,1
+digit_327,5
+digit_328,0
+digit_329,9
+digit_330,5
+digit_331,2
+digit_332,8
+digit_333,2
+digit_334,0
+digit_335,0
+digit_336,1
+digit_337,7
+digit_338,6
+digit_339,3
+digit_340,2
+digit_341,1
+digit_342,7
+digit_343,4
+digit_344,6
+digit_345,3
+digit_346,1
+digit_347,3
+digit_348,9
+digit_349,1
+digit_350,7
+digit_351,6
+digit_352,8
+digit_353,4
+digit_354,3
+digit_355,1
+digit_356,4
+digit_357,0
+digit_358,5
+digit_359,3
+digit_360,6
+digit_361,9
+digit_362,6
+digit_363,1
+digit_364,7
+digit_365,5
+digit_366,4
+digit_367,4
+digit_368,7
+digit_369,2
+digit_370,8
+digit_371,2
+digit_372,2
+digit_373,5
+digit_374,7
+digit_375,9
+digit_376,5
+digit_377,4
+digit_378,8
+digit_379,8
+digit_380,4
+digit_381,9
+digit_382,0
+digit_383,8
+digit_384,9
+digit_385,3
+digit_386,0
+digit_387,1
+digit_388,2
+digit_389,3
+digit_390,4
+digit_391,5
+digit_392,6
+digit_393,7
+digit_394,8
+digit_395,9
+digit_396,0
+digit_397,1
+digit_398,2
+digit_399,3
+digit_400,4
+digit_401,5
+digit_402,6
+digit_403,7
+digit_404,8
+digit_405,9
+digit_406,0
+digit_407,1
+digit_408,2
+digit_409,3
+digit_410,4
+digit_411,5
+digit_412,6
+digit_413,7
+digit_414,8
+digit_415,9
+digit_416,0
+digit_417,9
+digit_418,5
+digit_419,5
+digit_420,6
+digit_421,5
+digit_422,0
+digit_423,9
+digit_424,8
+digit_425,9
+digit_426,8
+digit_427,4
+digit_428,1
+digit_429,7
+digit_430,7
+digit_431,3
+digit_432,5
+digit_433,1
+digit_434,0
+digit_435,0
+digit_436,2
+digit_437,2
+digit_438,7
+digit_439,8
+digit_440,2
+digit_441,0
+digit_442,1
+digit_443,2
+digit_444,6
+digit_445,3
+digit_446,3
+digit_447,7
+digit_448,3
+digit_449,3
+digit_450,4
+digit_451,6
+digit_452,6
+digit_453,6
+digit_454,4
+digit_455,9
+digit_456,1
+digit_457,5
+digit_458,0
+digit_459,9
+digit_460,5
+digit_461,2
+digit_462,8
+digit_463,2
+digit_464,0
+digit_465,0
+digit_466,1
+digit_467,7
+digit_468,6
+digit_469,3
+digit_470,2
+digit_471,1
+digit_472,7
+digit_473,4
+digit_474,6
+digit_475,3
+digit_476,1
+digit_477,3
+digit_478,9
+digit_479,1
+digit_480,7
+digit_481,6
+digit_482,8
+digit_483,4
+digit_484,3
+digit_485,1
+digit_486,4
+digit_487,0
+digit_488,5
+digit_489,3
+digit_490,6
+digit_491,9
+digit_492,6
+digit_493,1
+digit_494,7
+digit_495,5
+digit_496,4
+digit_497,4
+digit_498,7
+digit_499,2
+digit_500,8
+digit_501,2
+digit_502,2
+digit_503,5
+digit_504,7
+digit_505,9
+digit_506,5
+digit_507,4
+digit_508,8
+digit_509,8
+digit_510,4
+digit_511,9
+digit_512,0
+digit_513,8
+digit_514,9
+digit_515,8
+digit_516,0
+digit_517,1
+digit_518,2
+digit_519,3
+digit_520,4
+digit_521,5
+digit_522,6
+digit_523,7
+digit_524,8
+digit_525,9
+digit_526,0
+digit_527,1
+digit_528,2
+digit_529,3
+digit_530,4
+digit_531,5
+digit_532,6
+digit_533,7
+digit_534,8
+digit_535,9
+digit_536,0
+digit_537,1
+digit_538,2
+digit_539,3
+digit_540,4
+digit_541,5
+digit_542,6
+digit_543,7
+digit_544,8
+digit_545,9
+digit_546,0
+digit_547,9
+digit_548,5
+digit_549,5
+digit_550,6
+digit_551,5
+digit_552,0
+digit_553,9
+digit_554,8
+digit_555,9
+digit_556,8
+digit_557,4
+digit_558,1
+digit_559,7
+digit_560,7
+digit_561,3
+digit_562,5
+digit_563,1
+digit_564,0
+digit_565,0
+digit_566,2
+digit_567,2
+digit_568,7
+digit_569,8
+digit_570,2
+digit_571,0
+digit_572,1
+digit_573,2
+digit_574,6
+digit_575,3
+digit_576,3
+digit_577,7
+digit_578,3
+digit_579,3
+digit_580,4
+digit_581,6
+digit_582,6
+digit_583,6
+digit_584,4
+digit_585,9
+digit_586,1
+digit_587,5
+digit_588,0
+digit_589,9
+digit_590,5
+digit_591,2
+digit_592,8
+digit_593,2
+digit_594,0
+digit_595,0
+digit_596,1
+digit_597,7
+digit_598,6
+digit_599,3
+digit_600,2
+digit_601,1
+digit_602,7
+digit_603,4
+digit_604,6
+digit_605,3
+digit_606,1
+digit_607,3
+digit_608,9
+digit_609,1
+digit_610,7
+digit_611,6
+digit_612,8
+digit_613,4
+digit_614,3
+digit_615,1
+digit_616,4
+digit_617,0
+digit_618,5
+digit_619,3
+digit_620,6
+digit_621,9
+digit_622,6
+digit_623,1
+digit_624,7
+digit_625,5
+digit_626,4
+digit_627,4
+digit_628,7
+digit_629,2
+digit_630,8
+digit_631,2
+digit_632,2
+digit_633,5
+digit_634,7
+digit_635,9
+digit_636,5
+digit_637,4
+digit_638,8
+digit_639,8
+digit_640,4
+digit_641,9
+digit_642,0
+digit_643,8
+digit_644,9
+digit_645,8
+digit_646,0
+digit_647,1
+digit_648,2
+digit_649,3
+digit_650,4
+digit_651,5
+digit_652,6
+digit_653,7
+digit_654,8
+digit_655,9
+digit_656,0
+digit_657,1
+digit_658,2
+digit_659,3
+digit_660,4
+digit_661,5
+digit_662,6
+digit_663,7
+digit_664,8
+digit_665,9
+digit_666,0
+digit_667,1
+digit_668,2
+digit_669,3
+digit_670,4
+digit_671,5
+digit_672,6
+digit_673,7
+digit_674,8
+digit_675,9
+digit_676,0
+digit_677,9
+digit_678,5
+digit_679,5
+digit_680,6
+digit_681,5
+digit_682,0
+digit_683,9
+digit_684,8
+digit_685,9
+digit_686,8
+digit_687,4
+digit_688,1
+digit_689,7
+digit_690,7
+digit_691,3
+digit_692,5
+digit_693,1
+digit_694,0
+digit_695,0
+digit_696,2
+digit_697,2
+digit_698,7
+digit_699,8
+digit_700,2
+digit_701,0
+digit_702,1
+digit_703,2
+digit_704,6
+digit_705,3
+digit_706,3
+digit_707,7
+digit_708,3
+digit_709,3
+digit_710,4
+digit_711,6
+digit_712,6
+digit_713,6
+digit_714,4
+digit_715,9
+digit_716,1
+digit_717,5
+digit_718,0
+digit_719,9
+digit_720,5
+digit_721,2
+digit_722,8
+digit_723,2
+digit_724,0
+digit_725,0
+digit_726,1
+digit_727,7
+digit_728,6
+digit_729,3
+digit_730,2
+digit_731,1
+digit_732,7
+digit_733,4
+digit_734,6
+digit_735,3
+digit_736,1
+digit_737,3
+digit_738,9
+digit_739,1
+digit_740,7
+digit_741,6
+digit_742,8
+digit_743,4
+digit_744,3
+digit_745,1
+digit_746,4
+digit_747,0
+digit_748,5
+digit_749,3
+digit_750,6
+digit_751,9
+digit_752,6
+digit_753,1
+digit_754,7
+digit_755,5
+digit_756,4
+digit_757,4
+digit_758,7
+digit_759,2
+digit_760,8
+digit_761,2
+digit_762,2
+digit_763,5
+digit_764,7
+digit_765,9
+digit_766,5
+digit_767,4
+digit_768,8
+digit_769,8
+digit_770,4
+digit_771,9
+digit_772,0
+digit_773,8
+digit_774,9
+digit_775,8
+digit_776,0
+digit_777,1
+digit_778,2
+digit_779,3
+digit_780,4
+digit_781,5
+digit_782,6
+digit_783,7
+digit_784,8
+digit_785,9
+digit_786,0
+digit_787,1
+digit_788,2
+digit_789,3
+digit_790,4
+digit_791,5
+digit_792,6
+digit_793,7
+digit_794,8
+digit_795,9
+digit_796,0
+digit_797,1
+digit_798,2
+digit_799,3
+digit_800,4
+digit_801,5
+digit_802,6
+digit_803,7
+digit_804,8
+digit_805,9
+digit_806,0
+digit_807,9
+digit_808,5
+digit_809,5
+digit_810,6
+digit_811,5
+digit_812,0
+digit_813,9
+digit_814,8
+digit_815,9
+digit_816,8
+digit_817,4
+digit_818,1
+digit_819,7
+digit_820,7
+digit_821,3
+digit_822,5
+digit_823,1
+digit_824,0
+digit_825,0
+digit_826,2
+digit_827,2
+digit_828,7
+digit_829,8
+digit_830,2
+digit_831,0
+digit_832,1
+digit_833,2
+digit_834,6
+digit_835,3
+digit_836,3
+digit_837,7
+digit_838,3
+digit_839,3
+digit_840,4
+digit_841,6
+digit_842,6
+digit_843,6
+digit_844,4
+digit_845,9
+digit_846,1
+digit_847,5
+digit_848,0
+digit_849,9
+digit_850,5
+digit_851,2
+digit_852,8
+digit_853,2
+digit_854,0
+digit_855,0
+digit_856,1
+digit_857,7
+digit_858,6
+digit_859,3
+digit_860,2
+digit_861,1
+digit_862,7
+digit_863,4
+digit_864,6
+digit_865,3
+digit_866,1
+digit_867,3
+digit_868,9
+digit_869,1
+digit_870,7
+digit_871,6
+digit_872,8
+digit_873,4
+digit_874,3
+digit_875,1
+digit_876,4
+digit_877,0
+digit_878,5
+digit_879,3
+digit_880,6
+digit_881,9
+digit_882,6
+digit_883,1
+digit_884,7
+digit_885,5
+digit_886,4
+digit_887,4
+digit_888,7
+digit_889,2
+digit_890,8
+digit_891,2
+digit_892,2
+digit_893,5
+digit_894,7
+digit_895,9
+digit_896,5
+digit_897,4
+digit_898,8
+digit_899,8
+digit_900,4
+digit_901,9
+digit_902,0
+digit_903,8
+digit_904,9
+digit_905,8
+digit_906,1
+digit_907,2
+digit_908,3
+digit_909,4
+digit_910,5
+digit_911,6
+digit_912,7
+digit_913,8
+digit_914,9
+digit_915,0
+digit_916,1
+digit_917,2
+digit_918,3
+digit_919,4
+digit_920,5
+digit_921,6
+digit_922,7
+digit_923,8
+digit_924,9
+digit_925,0
+digit_926,1
+digit_927,2
+digit_928,3
+digit_929,4
+digit_930,5
+digit_931,6
+digit_932,7
+digit_933,8
+digit_934,9
+digit_935,0
+digit_936,9
+digit_937,5
+digit_938,5
+digit_939,6
+digit_940,5
+digit_941,0
+digit_942,9
+digit_943,8
+digit_944,9
+digit_945,8
+digit_946,4
+digit_947,1
+digit_948,7
+digit_949,7
+digit_950,3
+digit_951,5
+digit_952,1
+digit_953,2
+digit_954,7
+digit_955,8
+digit_956,2
+digit_957,0
+digit_958,1
+digit_959,2
+digit_960,6
+digit_961,3
+digit_962,3
+digit_963,7
+digit_964,3
+digit_965,3
+digit_966,4
+digit_967,6
+digit_968,6
+digit_969,6
+digit_970,4
+digit_971,9
+digit_972,1
+digit_973,5
+digit_974,0
+digit_975,9
+digit_976,5
+digit_977,2
+digit_978,8
+digit_979,2
+digit_980,0
+digit_981,0
+digit_982,1
+digit_983,7
+digit_984,6
+digit_985,3
+digit_986,2
+digit_987,1
+digit_988,4
+digit_989,6
+digit_990,3
+digit_991,1
+digit_992,3
+digit_993,9
+digit_994,1
+digit_995,7
+digit_996,6
+digit_997,8
+digit_998,4
+digit_999,3
+digit_1000,1
+digit_1001,4
+digit_1002,0
+digit_1003,5
+digit_1004,3
+digit_1005,6
+digit_1006,9
+digit_1007,6
+digit_1008,1
+digit_1009,7
+digit_1010,5
+digit_1011,4
+digit_1012,4
+digit_1013,7
+digit_1014,2
+digit_1015,8
+digit_1016,2
+digit_1017,2
+digit_1018,5
+digit_1019,7
+digit_1020,9
+digit_1021,5
+digit_1022,4
+digit_1023,4
+digit_1024,9
+digit_1025,0
+digit_1026,8
+digit_1027,9
+digit_1028,8
+digit_1029,0
+digit_1030,1
+digit_1031,2
+digit_1032,3
+digit_1033,4
+digit_1034,5
+digit_1035,6
+digit_1036,7
+digit_1037,8
+digit_1038,9
+digit_1039,0
+digit_1040,1
+digit_1041,2
+digit_1042,3
+digit_1043,4
+digit_1044,5
+digit_1045,6
+digit_1046,7
+digit_1047,8
+digit_1048,9
+digit_1049,0
+digit_1050,1
+digit_1051,2
+digit_1052,3
+digit_1053,4
+digit_1054,5
+digit_1055,6
+digit_1056,7
+digit_1057,8
+digit_1058,9
+digit_1059,0
+digit_1060,9
+digit_1061,5
+digit_1062,5
+digit_1063,6
+digit_1064,5
+digit_1065,0
+digit_1066,9
+digit_1067,8
+digit_1068,9
+digit_1069,8
+digit_1070,4
+digit_1071,1
+digit_1072,7
+digit_1073,7
+digit_1074,3
+digit_1075,5
+digit_1076,1
+digit_1077,0
+digit_1078,0
+digit_1079,7
+digit_1080,8
+digit_1081,2
+digit_1082,0
+digit_1083,1
+digit_1084,2
+digit_1085,6
+digit_1086,3
+digit_1087,3
+digit_1088,7
+digit_1089,3
+digit_1090,3
+digit_1091,4
+digit_1092,6
+digit_1093,6
+digit_1094,6
+digit_1095,4
+digit_1096,9
+digit_1097,1
+digit_1098,5
+digit_1099,0
+digit_1100,9
+digit_1101,5
+digit_1102,2
+digit_1103,8
+digit_1104,2
+digit_1105,0
+digit_1106,0
+digit_1107,1
+digit_1108,7
+digit_1109,6
+digit_1110,3
+digit_1111,2
+digit_1112,1
+digit_1113,7
+digit_1114,4
+digit_1115,6
+digit_1116,3
+digit_1117,1
+digit_1118,3
+digit_1119,9
+digit_1120,1
+digit_1121,7
+digit_1122,6
+digit_1123,8
+digit_1124,4
+digit_1125,3
+digit_1126,1
+digit_1127,4
+digit_1128,0
+digit_1129,5
+digit_1130,3
+digit_1131,6
+digit_1132,9
+digit_1133,6
+digit_1134,1
+digit_1135,7
+digit_1136,5
+digit_1137,4
+digit_1138,4
+digit_1139,7
+digit_1140,2
+digit_1141,8
+digit_1142,2
+digit_1143,2
+digit_1144,5
+digit_1145,7
+digit_1146,9
+digit_1147,5
+digit_1148,4
+digit_1149,8
+digit_1150,8
+digit_1151,4
+digit_1152,9
+digit_1153,0
+digit_1154,8
+digit_1155,9
+digit_1156,8
+digit_1157,0
+digit_1158,1
+digit_1159,2
+digit_1160,3
+digit_1161,4
+digit_1162,5
+digit_1163,6
+digit_1164,7
+digit_1165,8
+digit_1166,9
+digit_1167,0
+digit_1168,1
+digit_1169,2
+digit_1170,3
+digit_1171,4
+digit_1172,5
+digit_1173,6
+digit_1174,7
+digit_1175,8
+digit_1176,9
+digit_1177,0
+digit_1178,1
+digit_1179,2
+digit_1180,3
+digit_1181,4
+digit_1182,5
+digit_1183,6
+digit_1184,7
+digit_1185,8
+digit_1186,9
+digit_1187,0
+digit_1188,9
+digit_1189,5
+digit_1190,5
+digit_1191,6
+digit_1192,5
+digit_1193,0
+digit_1194,9
+digit_1195,8
+digit_1196,9
+digit_1197,8
+digit_1198,4
+digit_1199,1
+digit_1200,7
+digit_1201,7
+digit_1202,3
+digit_1203,5
+digit_1204,1
+digit_1205,0
+digit_1206,0
+digit_1207,2
+digit_1208,2
+digit_1209,7
+digit_1210,8
+digit_1211,2
+digit_1212,0
+digit_1213,1
+digit_1214,2
+digit_1215,6
+digit_1216,3
+digit_1217,3
+digit_1218,7
+digit_1219,3
+digit_1220,3
+digit_1221,4
+digit_1222,6
+digit_1223,6
+digit_1224,6
+digit_1225,4
+digit_1226,9
+digit_1227,1
+digit_1228,5
+digit_1229,0
+digit_1230,9
+digit_1231,5
+digit_1232,2
+digit_1233,8
+digit_1234,2
+digit_1235,0
+digit_1236,0
+digit_1237,1
+digit_1238,7
+digit_1239,6
+digit_1240,3
+digit_1241,2
+digit_1242,1
+digit_1243,7
+digit_1244,4
+digit_1245,6
+digit_1246,3
+digit_1247,1
+digit_1248,3
+digit_1249,9
+digit_1250,1
+digit_1251,7
+digit_1252,6
+digit_1253,8
+digit_1254,4
+digit_1255,3
+digit_1256,1
+digit_1257,4
+digit_1258,0
+digit_1259,5
+digit_1260,3
+digit_1261,6
+digit_1262,9
+digit_1263,6
+digit_1264,1
+digit_1265,7
+digit_1266,5
+digit_1267,4
+digit_1268,4
+digit_1269,7
+digit_1270,2
+digit_1271,8
+digit_1272,2
+digit_1273,2
+digit_1274,5
+digit_1275,7
+digit_1276,9
+digit_1277,5
+digit_1278,4
+digit_1279,8
+digit_1280,8
+digit_1281,4
+digit_1282,9
+digit_1283,0
+digit_1284,8
+digit_1285,9
+digit_1286,8
+digit_1287,0
+digit_1288,1
+digit_1289,2
+digit_1290,3
+digit_1291,4
+digit_1292,5
+digit_1293,6
+digit_1294,7
+digit_1295,8
+digit_1296,9
+digit_1297,0
+digit_1298,1
+digit_1299,2
+digit_1300,3
+digit_1301,4
+digit_1302,5
+digit_1303,6
+digit_1304,7
+digit_1305,8
+digit_1306,9
+digit_1307,0
+digit_1308,1
+digit_1309,2
+digit_1310,3
+digit_1311,4
+digit_1312,5
+digit_1313,6
+digit_1314,7
+digit_1315,8
+digit_1316,9
+digit_1317,0
+digit_1318,9
+digit_1319,5
+digit_1320,5
+digit_1321,6
+digit_1322,5
+digit_1323,0
+digit_1324,9
+digit_1325,8
+digit_1326,9
+digit_1327,8
+digit_1328,4
+digit_1329,1
+digit_1330,7
+digit_1331,7
+digit_1332,3
+digit_1333,5
+digit_1334,1
+digit_1335,0
+digit_1336,0
+digit_1337,2
+digit_1338,2
+digit_1339,7
+digit_1340,8
+digit_1341,2
+digit_1342,0
+digit_1343,1
+digit_1344,2
+digit_1345,6
+digit_1346,3
+digit_1347,3
+digit_1348,7
+digit_1349,3
+digit_1350,3
+digit_1351,4
+digit_1352,6
+digit_1353,6
+digit_1354,6
+digit_1355,4
+digit_1356,9
+digit_1357,1
+digit_1358,5
+digit_1359,0
+digit_1360,9
+digit_1361,5
+digit_1362,2
+digit_1363,8
+digit_1364,2
+digit_1365,0
+digit_1366,0
+digit_1367,1
+digit_1368,7
+digit_1369,6
+digit_1370,3
+digit_1371,2
+digit_1372,1
+digit_1373,7
+digit_1374,4
+digit_1375,6
+digit_1376,3
+digit_1377,1
+digit_1378,3
+digit_1379,9
+digit_1380,1
+digit_1381,7
+digit_1382,6
+digit_1383,8
+digit_1384,4
+digit_1385,3
+digit_1386,1
+digit_1387,4
+digit_1388,0
+digit_1389,5
+digit_1390,3
+digit_1391,6
+digit_1392,9
+digit_1393,6
+digit_1394,1
+digit_1395,7
+digit_1396,5
+digit_1397,4
+digit_1398,4
+digit_1399,7
+digit_1400,2
+digit_1401,8
+digit_1402,2
+digit_1403,2
+digit_1404,5
+digit_1405,7
+digit_1406,9
+digit_1407,5
+digit_1408,4
+digit_1409,8
+digit_1410,8
+digit_1411,4
+digit_1412,9
+digit_1413,0
+digit_1414,8
+digit_1415,0
+digit_1416,1
+digit_1417,2
+digit_1418,3
+digit_1419,4
+digit_1420,5
+digit_1421,6
+digit_1422,7
+digit_1423,8
+digit_1424,9
+digit_1425,0
+digit_1426,1
+digit_1427,2
+digit_1428,3
+digit_1429,4
+digit_1430,5
+digit_1431,6
+digit_1432,7
+digit_1433,8
+digit_1434,9
+digit_1435,0
+digit_1436,1
+digit_1437,2
+digit_1438,3
+digit_1439,4
+digit_1440,5
+digit_1441,6
+digit_1442,7
+digit_1443,8
+digit_1444,9
+digit_1445,0
+digit_1446,9
+digit_1447,5
+digit_1448,5
+digit_1449,6
+digit_1450,5
+digit_1451,0
+digit_1452,9
+digit_1453,8
+digit_1454,9
+digit_1455,8
+digit_1456,4
+digit_1457,1
+digit_1458,7
+digit_1459,7
+digit_1460,3
+digit_1461,5
+digit_1462,1
+digit_1463,0
+digit_1464,0
+digit_1465,2
+digit_1466,2
+digit_1467,7
+digit_1468,8
+digit_1469,2
+digit_1470,0
+digit_1471,1
+digit_1472,2
+digit_1473,6
+digit_1474,3
+digit_1475,3
+digit_1476,7
+digit_1477,3
+digit_1478,3
+digit_1479,4
+digit_1480,6
+digit_1481,6
+digit_1482,6
+digit_1483,4
+digit_1484,9
+digit_1485,1
+digit_1486,5
+digit_1487,0
+digit_1488,9
+digit_1489,5
+digit_1490,2
+digit_1491,8
+digit_1492,2
+digit_1493,0
+digit_1494,0
+digit_1495,1
+digit_1496,7
+digit_1497,6
+digit_1498,3
+digit_1499,2
+digit_1500,1
+digit_1501,7
+digit_1502,4
+digit_1503,6
+digit_1504,3
+digit_1505,1
+digit_1506,3
+digit_1507,9
+digit_1508,1
+digit_1509,7
+digit_1510,6
+digit_1511,8
+digit_1512,4
+digit_1513,3
+digit_1514,1
+digit_1515,4
+digit_1516,0
+digit_1517,5
+digit_1518,3
+digit_1519,6
+digit_1520,9
+digit_1521,6
+digit_1522,1
+digit_1523,7
+digit_1524,5
+digit_1525,4
+digit_1526,4
+digit_1527,7
+digit_1528,2
+digit_1529,8
+digit_1530,2
+digit_1531,2
+digit_1532,5
+digit_1533,7
+digit_1534,9
+digit_1535,5
+digit_1536,4
+digit_1537,8
+digit_1538,8
+digit_1539,4
+digit_1540,9
+digit_1541,0
+digit_1542,8
+digit_1543,9
+digit_1544,8
+digit_1545,0
+digit_1546,1
+digit_1547,2
+digit_1548,3
+digit_1549,4
+digit_1550,5
+digit_1551,6
+digit_1552,7
+digit_1553,8
+digit_1554,9
+digit_1555,0
+digit_1556,1
+digit_1557,2
+digit_1558,3
+digit_1559,4
+digit_1560,5
+digit_1561,6
+digit_1562,9
+digit_1563,0
+digit_1564,1
+digit_1565,2
+digit_1566,3
+digit_1567,4
+digit_1568,5
+digit_1569,6
+digit_1570,7
+digit_1571,8
+digit_1572,9
+digit_1573,0
+digit_1574,9
+digit_1575,5
+digit_1576,5
+digit_1577,6
+digit_1578,5
+digit_1579,0
+digit_1580,9
+digit_1581,8
+digit_1582,9
+digit_1583,8
+digit_1584,4
+digit_1585,1
+digit_1586,7
+digit_1587,7
+digit_1588,3
+digit_1589,5
+digit_1590,1
+digit_1591,0
+digit_1592,0
+digit_1593,2
+digit_1594,2
+digit_1595,7
+digit_1596,8
+digit_1597,2
+digit_1598,0
+digit_1599,1
+digit_1600,2
+digit_1601,6
+digit_1602,3
+digit_1603,3
+digit_1604,7
+digit_1605,3
+digit_1606,3
+digit_1607,4
+digit_1608,6
+digit_1609,6
+digit_1610,6
+digit_1611,4
+digit_1612,9
+digit_1613,1
+digit_1614,5
+digit_1615,0
+digit_1616,9
+digit_1617,5
+digit_1618,2
+digit_1619,8
+digit_1620,0
+digit_1621,1
+digit_1622,7
+digit_1623,6
+digit_1624,3
+digit_1625,2
+digit_1626,1
+digit_1627,7
+digit_1628,4
+digit_1629,6
+digit_1630,3
+digit_1631,1
+digit_1632,3
+digit_1633,9
+digit_1634,1
+digit_1635,7
+digit_1636,6
+digit_1637,8
+digit_1638,4
+digit_1639,3
+digit_1640,1
+digit_1641,4
+digit_1642,0
+digit_1643,5
+digit_1644,3
+digit_1645,6
+digit_1646,9
+digit_1647,6
+digit_1648,1
+digit_1649,7
+digit_1650,5
+digit_1651,4
+digit_1652,4
+digit_1653,7
+digit_1654,2
+digit_1655,2
+digit_1656,5
+digit_1657,7
+digit_1658,9
+digit_1659,5
+digit_1660,4
+digit_1661,4
+digit_1662,9
+digit_1663,0
+digit_1664,8
+digit_1665,9
+digit_1666,8
+digit_1667,0
+digit_1668,1
+digit_1669,2
+digit_1670,3
+digit_1671,4
+digit_1672,5
+digit_1673,6
+digit_1674,7
+digit_1675,8
+digit_1676,9
+digit_1677,0
+digit_1678,1
+digit_1679,2
+digit_1680,3
+digit_1681,4
+digit_1682,5
+digit_1683,6
+digit_1684,7
+digit_1685,8
+digit_1686,9
+digit_1687,0
+digit_1688,1
+digit_1689,2
+digit_1690,3
+digit_1691,4
+digit_1692,5
+digit_1693,6
+digit_1694,7
+digit_1695,8
+digit_1696,9
+digit_1697,0
+digit_1698,9
+digit_1699,5
+digit_1700,5
+digit_1701,6
+digit_1702,5
+digit_1703,0
+digit_1704,9
+digit_1705,8
+digit_1706,9
+digit_1707,8
+digit_1708,4
+digit_1709,1
+digit_1710,7
+digit_1711,7
+digit_1712,3
+digit_1713,5
+digit_1714,1
+digit_1715,0
+digit_1716,0
+digit_1717,2
+digit_1718,2
+digit_1719,7
+digit_1720,8
+digit_1721,2
+digit_1722,0
+digit_1723,1
+digit_1724,2
+digit_1725,6
+digit_1726,3
+digit_1727,3
+digit_1728,7
+digit_1729,3
+digit_1730,3
+digit_1731,4
+digit_1732,6
+digit_1733,6
+digit_1734,6
+digit_1735,4
+digit_1736,9
+digit_1737,1
+digit_1738,5
+digit_1739,0
+digit_1740,9
+digit_1741,5
+digit_1742,2
+digit_1743,8
+digit_1744,2
+digit_1745,0
+digit_1746,0
+digit_1747,1
+digit_1748,7
+digit_1749,6
+digit_1750,3
+digit_1751,2
+digit_1752,1
+digit_1753,7
+digit_1754,4
+digit_1755,6
+digit_1756,3
+digit_1757,1
+digit_1758,3
+digit_1759,9
+digit_1760,1
+digit_1761,7
+digit_1762,6
+digit_1763,8
+digit_1764,4
+digit_1765,3
+digit_1766,1
+digit_1767,4
+digit_1768,0
+digit_1769,5
+digit_1770,3
+digit_1771,6
+digit_1772,9
+digit_1773,6
+digit_1774,1
+digit_1775,7
+digit_1776,5
+digit_1777,4
+digit_1778,4
+digit_1779,7
+digit_1780,2
+digit_1781,8
+digit_1782,2
+digit_1783,2
+digit_1784,5
+digit_1785,7
+digit_1786,9
+digit_1787,5
+digit_1788,4
+digit_1789,8
+digit_1790,8
+digit_1791,4
+digit_1792,9
+digit_1793,0
+digit_1794,8
+digit_1795,9
+digit_1796,8

tmp/inputs/iris_data.csv

@@ -0,0 +1,151 @@
+,sepal length (cm),sepal width (cm),petal length (cm),petal width (cm)
+sample_0,5.1,3.5,1.4,0.2
+sample_1,4.9,3.0,1.4,0.2
+sample_2,4.7,3.2,1.3,0.2
+sample_3,4.6,3.1,1.5,0.2
+sample_4,5.0,3.6,1.4,0.2
+sample_5,5.4,3.9,1.7,0.4
+sample_6,4.6,3.4,1.4,0.3
+sample_7,5.0,3.4,1.5,0.2
+sample_8,4.4,2.9,1.4,0.2
+sample_9,4.9,3.1,1.5,0.1
+sample_10,5.4,3.7,1.5,0.2
+sample_11,4.8,3.4,1.6,0.2
+sample_12,4.8,3.0,1.4,0.1
+sample_13,4.3,3.0,1.1,0.1
+sample_14,5.8,4.0,1.2,0.2
+sample_15,5.7,4.4,1.5,0.4
+sample_16,5.4,3.9,1.3,0.4
+sample_17,5.1,3.5,1.4,0.3
+sample_18,5.7,3.8,1.7,0.3
+sample_19,5.1,3.8,1.5,0.3
+sample_20,5.4,3.4,1.7,0.2
+sample_21,5.1,3.7,1.5,0.4
+sample_22,4.6,3.6,1.0,0.2
+sample_23,5.1,3.3,1.7,0.5
+sample_24,4.8,3.4,1.9,0.2
+sample_25,5.0,3.0,1.6,0.2
+sample_26,5.0,3.4,1.6,0.4
+sample_27,5.2,3.5,1.5,0.2
+sample_28,5.2,3.4,1.4,0.2
+sample_29,4.7,3.2,1.6,0.2
+sample_30,4.8,3.1,1.6,0.2
+sample_31,5.4,3.4,1.5,0.4
+sample_32,5.2,4.1,1.5,0.1
+sample_33,5.5,4.2,1.4,0.2
+sample_34,4.9,3.1,1.5,0.2
+sample_35,5.0,3.2,1.2,0.2
+sample_36,5.5,3.5,1.3,0.2
+sample_37,4.9,3.6,1.4,0.1
+sample_38,4.4,3.0,1.3,0.2
+sample_39,5.1,3.4,1.5,0.2
+sample_40,5.0,3.5,1.3,0.3
+sample_41,4.5,2.3,1.3,0.3
+sample_42,4.4,3.2,1.3,0.2
+sample_43,5.0,3.5,1.6,0.6
+sample_44,5.1,3.8,1.9,0.4
+sample_45,4.8,3.0,1.4,0.3
+sample_46,5.1,3.8,1.6,0.2
+sample_47,4.6,3.2,1.4,0.2
+sample_48,5.3,3.7,1.5,0.2
+sample_49,5.0,3.3,1.4,0.2
+sample_50,7.0,3.2,4.7,1.4
+sample_51,6.4,3.2,4.5,1.5
+sample_52,6.9,3.1,4.9,1.5
+sample_53,5.5,2.3,4.0,1.3
+sample_54,6.5,2.8,4.6,1.5
+sample_55,5.7,2.8,4.5,1.3
+sample_56,6.3,3.3,4.7,1.6
+sample_57,4.9,2.4,3.3,1.0
+sample_58,6.6,2.9,4.6,1.3
+sample_59,5.2,2.7,3.9,1.4
+sample_60,5.0,2.0,3.5,1.0
+sample_61,5.9,3.0,4.2,1.5
+sample_62,6.0,2.2,4.0,1.0
+sample_63,6.1,2.9,4.7,1.4
+sample_64,5.6,2.9,3.6,1.3
+sample_65,6.7,3.1,4.4,1.4
+sample_66,5.6,3.0,4.5,1.5
+sample_67,5.8,2.7,4.1,1.0
+sample_68,6.2,2.2,4.5,1.5
+sample_69,5.6,2.5,3.9,1.1
+sample_70,5.9,3.2,4.8,1.8
+sample_71,6.1,2.8,4.0,1.3
+sample_72,6.3,2.5,4.9,1.5
+sample_73,6.1,2.8,4.7,1.2
+sample_74,6.4,2.9,4.3,1.3
+sample_75,6.6,3.0,4.4,1.4
+sample_76,6.8,2.8,4.8,1.4
+sample_77,6.7,3.0,5.0,1.7
+sample_78,6.0,2.9,4.5,1.5
+sample_79,5.7,2.6,3.5,1.0
+sample_80,5.5,2.4,3.8,1.1
+sample_81,5.5,2.4,3.7,1.0
+sample_82,5.8,2.7,3.9,1.2
+sample_83,6.0,2.7,5.1,1.6
+sample_84,5.4,3.0,4.5,1.5
+sample_85,6.0,3.4,4.5,1.6
+sample_86,6.7,3.1,4.7,1.5
+sample_87,6.3,2.3,4.4,1.3
+sample_88,5.6,3.0,4.1,1.3
+sample_89,5.5,2.5,4.0,1.3
+sample_90,5.5,2.6,4.4,1.2
+sample_91,6.1,3.0,4.6,1.4
+sample_92,5.8,2.6,4.0,1.2
+sample_93,5.0,2.3,3.3,1.0
+sample_94,5.6,2.7,4.2,1.3
+sample_95,5.7,3.0,4.2,1.2
+sample_96,5.7,2.9,4.2,1.3
+sample_97,6.2,2.9,4.3,1.3
+sample_98,5.1,2.5,3.0,1.1
+sample_99,5.7,2.8,4.1,1.3
+sample_100,6.3,3.3,6.0,2.5
+sample_101,5.8,2.7,5.1,1.9
+sample_102,7.1,3.0,5.9,2.1
+sample_103,6.3,2.9,5.6,1.8
+sample_104,6.5,3.0,5.8,2.2
+sample_105,7.6,3.0,6.6,2.1
+sample_106,4.9,2.5,4.5,1.7
+sample_107,7.3,2.9,6.3,1.8
+sample_108,6.7,2.5,5.8,1.8
+sample_109,7.2,3.6,6.1,2.5
+sample_110,6.5,3.2,5.1,2.0
+sample_111,6.4,2.7,5.3,1.9
+sample_112,6.8,3.0,5.5,2.1
+sample_113,5.7,2.5,5.0,2.0
+sample_114,5.8,2.8,5.1,2.4
+sample_115,6.4,3.2,5.3,2.3
+sample_116,6.5,3.0,5.5,1.8
+sample_117,7.7,3.8,6.7,2.2
+sample_118,7.7,2.6,6.9,2.3
+sample_119,6.0,2.2,5.0,1.5
+sample_120,6.9,3.2,5.7,2.3
+sample_121,5.6,2.8,4.9,2.0
+sample_122,7.7,2.8,6.7,2.0
+sample_123,6.3,2.7,4.9,1.8
+sample_124,6.7,3.3,5.7,2.1
+sample_125,7.2,3.2,6.0,1.8
+sample_126,6.2,2.8,4.8,1.8
+sample_127,6.1,3.0,4.9,1.8
+sample_128,6.4,2.8,5.6,2.1
+sample_129,7.2,3.0,5.8,1.6
+sample_130,7.4,2.8,6.1,1.9
+sample_131,7.9,3.8,6.4,2.0
+sample_132,6.4,2.8,5.6,2.2
+sample_133,6.3,2.8,5.1,1.5
+sample_134,6.1,2.6,5.6,1.4
+sample_135,7.7,3.0,6.1,2.3
+sample_136,6.3,3.4,5.6,2.4
+sample_137,6.4,3.1,5.5,1.8
+sample_138,6.0,3.0,4.8,1.8
+sample_139,6.9,3.1,5.4,2.1
+sample_140,6.7,3.1,5.6,2.4
+sample_141,6.9,3.1,5.1,2.3
+sample_142,5.8,2.7,5.1,1.9
+sample_143,6.8,3.2,5.9,2.3
+sample_144,6.7,3.3,5.7,2.5
+sample_145,6.7,3.0,5.2,2.3
+sample_146,6.3,2.5,5.0,1.9
+sample_147,6.5,3.0,5.2,2.0
+sample_148,6.2,3.4,5.4,2.3
+sample_149,5.9,3.0,5.1,1.8

tmp/inputs/iris_labels.csv

@@ -0,0 +1,151 @@
+,label
+sample_0,setosa
+sample_1,setosa
+sample_2,setosa
+sample_3,setosa
+sample_4,setosa
+sample_5,setosa
+sample_6,setosa
+sample_7,setosa
+sample_8,setosa
+sample_9,setosa
+sample_10,setosa
+sample_11,setosa
+sample_12,setosa
+sample_13,setosa
+sample_14,setosa
+sample_15,setosa
+sample_16,setosa
+sample_17,setosa
+sample_18,setosa
+sample_19,setosa
+sample_20,setosa
+sample_21,setosa
+sample_22,setosa
+sample_23,setosa
+sample_24,setosa
+sample_25,setosa
+sample_26,setosa
+sample_27,setosa
+sample_28,setosa
+sample_29,setosa
+sample_30,setosa
+sample_31,setosa
+sample_32,setosa
+sample_33,setosa
+sample_34,setosa
+sample_35,setosa
+sample_36,setosa
+sample_37,setosa
+sample_38,setosa
+sample_39,setosa
+sample_40,setosa
+sample_41,setosa
+sample_42,setosa
+sample_43,setosa
+sample_44,setosa
+sample_45,setosa
+sample_46,setosa
+sample_47,setosa
+sample_48,setosa
+sample_49,setosa
+sample_50,versicolor
+sample_51,versicolor
+sample_52,versicolor
+sample_53,versicolor
+sample_54,versicolor
+sample_55,versicolor
+sample_56,versicolor
+sample_57,versicolor
+sample_58,versicolor
+sample_59,versicolor
+sample_60,versicolor
+sample_61,versicolor
+sample_62,versicolor
+sample_63,versicolor
+sample_64,versicolor
+sample_65,versicolor
+sample_66,versicolor
+sample_67,versicolor
+sample_68,versicolor
+sample_69,versicolor
+sample_70,versicolor
+sample_71,versicolor
+sample_72,versicolor
+sample_73,versicolor
+sample_74,versicolor
+sample_75,versicolor
+sample_76,versicolor
+sample_77,versicolor
+sample_78,versicolor
+sample_79,versicolor
+sample_80,versicolor
+sample_81,versicolor
+sample_82,versicolor
+sample_83,versicolor
+sample_84,versicolor
+sample_85,versicolor
+sample_86,versicolor
+sample_87,versicolor
+sample_88,versicolor
+sample_89,versicolor
+sample_90,versicolor
+sample_91,versicolor
+sample_92,versicolor
+sample_93,versicolor
+sample_94,versicolor
+sample_95,versicolor
+sample_96,versicolor
+sample_97,versicolor
+sample_98,versicolor
+sample_99,versicolor
+sample_100,virginica
+sample_101,virginica
+sample_102,virginica
+sample_103,virginica
+sample_104,virginica
+sample_105,virginica
+sample_106,virginica
+sample_107,virginica
+sample_108,virginica
+sample_109,virginica
+sample_110,virginica
+sample_111,virginica
+sample_112,virginica
+sample_113,virginica
+sample_114,virginica
+sample_115,virginica
+sample_116,virginica
+sample_117,virginica
+sample_118,virginica
+sample_119,virginica
+sample_120,virginica
+sample_121,virginica
+sample_122,virginica
+sample_123,virginica
+sample_124,virginica
+sample_125,virginica
+sample_126,virginica
+sample_127,virginica
+sample_128,virginica
+sample_129,virginica
+sample_130,virginica
+sample_131,virginica
+sample_132,virginica
+sample_133,virginica
+sample_134,virginica
+sample_135,virginica
+sample_136,virginica
+sample_137,virginica
+sample_138,virginica
+sample_139,virginica
+sample_140,virginica
+sample_141,virginica
+sample_142,virginica
+sample_143,virginica
+sample_144,virginica
+sample_145,virginica
+sample_146,virginica
+sample_147,virginica
+sample_148,virginica
+sample_149,virginica