update CLAUDE.md

.pre-commit-config.yaml

@@ -34,7 +34,7 @@ repos:
     rev: v1.0.1
     hooks:
       - id: mypy
-        exclude: ^py/tests|^py/unit_tests
+        exclude: ^py/tests|^py/unit_tests|^examples
         additional_dependencies:
           [
             "types-requests",

examples/sheets/coding-agent-analysis/CLAUDE.md

@@ -0,0 +1,73 @@
+This project uses LlamaSheets to extract data from spreadsheets for analysis.
+
+## Current Project Structure
+
+- `data/` - Contains extracted parquet files from LlamaSheets
+  - `{name}_region_{N}.parquet` - Table data files
+  - `{name}_metadata_{N}.parquet` - Cell metadata files
+  - `{name}_job_metadata.json` - Extraction job information
+- `scripts/` - Analysis and helper scripts
+- `reports/` - Your generated reports and outputs
+
+## Working with LlamaSheets Data
+
+### Understanding the Files
+
+When a spreadsheet is extracted, you'll find:
+
+1. **Table parquet files** (`region_*.parquet`): The actual table data
+   - Columns correspond to spreadsheet columns
+   - Data types are preserved (dates, numbers, strings, booleans)
+
+2. **Metadata parquet files** (`metadata_*.parquet`): Rich cell-level metadata
+   - Formatting: `font_bold`, `font_italic`, `font_size`, `background_color_rgb`
+   - Position: `row_number`, `column_number`, `coordinate` (e.g., "A1")
+   - Type detection: `data_type`, `is_date_like`, `is_percentage`, `is_currency`
+   - Layout: `is_in_first_row`, `is_merged_cell`, `horizontal_alignment`
+   - Content: `cell_value`, `raw_cell_value`
+
+3. **Job metadata JSON** (`job_metadata.json`): Overall extraction results
+   - `regions[]`: List of extracted regions with IDs, locations, and titles/descriptions
+   - `worksheet_metadata[]`: Generated titles and descriptions
+   - `status`: Success/failure status
+
+### Key Principles
+
+1. **Use metadata to understand structure**: Bold cells often indicate headers, colors indicate groupings
+2. **Validate before analysis**: Check data types, look for missing values
+3. **Preserve formatting context**: The metadata tells you what the spreadsheet author emphasized
+4. **Save intermediate results**: Store cleaned data as new parquet files
+
+### Common Patterns
+
+**Loading data:**
+```python
+import pandas as pd
+
+df = pd.read_parquet("data/region_1_Sheet1.parquet")
+meta_df = pd.read_parquet("data/metadata_1_Sheet1.parquet")
+```
+
+**Finding headers:**
+```python
+headers = meta_df[meta_df["font_bold"] == True]["cell_value"].tolist()
+```
+
+**Finding date columns:**
+```python
+date_cols = meta_df[meta_df["is_date_like"] == True]["column_number"].unique()
+```
+
+## Tools Available
+
+- **Python 3.11+**: For data analysis
+- **pandas**: DataFrame manipulation
+- **pyarrow**: Parquet file reading
+- **matplotlib**: Visualization (optional)
+
+## Guidelines
+
+- Always read the job_metadata.json first to understand what was extracted
+- Check both table data and metadata before making assumptions
+- Write reusable functions for common operations
+- Document any data quality issues discovered

examples/sheets/coding-agent-analysis/generate_sample_data.py

@@ -0,0 +1,278 @@
+"""
+Generate sample spreadsheets for LlamaSheets + Claude workflows.
+
+This script creates example Excel files that demonstrate different use cases:
+1. Simple data table (for Workflow 1)
+2. Regional sales data (for Workflow 2)
+3. Complex budget with formatting (for Workflow 3)
+4. Weekly sales report (for Workflow 4)
+
+Usage:
+    python generate_sample_data.py
+"""
+
+import random
+from datetime import datetime, timedelta
+from pathlib import Path
+
+import pandas as pd
+from openpyxl import Workbook
+from openpyxl.styles import Font, PatternFill, Alignment
+
+
+def generate_workflow_1_data(output_dir: Path) -> None:
+    """Generate simple financial report for Workflow 1."""
+    print("📊 Generating Workflow 1: financial_report_q1.xlsx")
+
+    # Create sample quarterly data
+    months = ["January", "February", "March"]
+    categories = ["Revenue", "Cost of Goods Sold", "Operating Expenses", "Net Income"]
+
+    data = []
+    for category in categories:
+        row: dict[str, str | int] = {"Category": category}
+        for month in months:
+            if category == "Revenue":
+                value = random.randint(80000, 120000)
+            elif category == "Cost of Goods Sold":
+                value = random.randint(30000, 50000)
+            elif category == "Operating Expenses":
+                value = random.randint(20000, 35000)
+            else:  # Net Income
+                value = int(
+                    int(row.get("January", 0))
+                    + int(row.get("February", 0))
+                    + int(row.get("March", 0))
+                )
+                value = random.randint(15000, 40000)
+            row[month] = value
+        data.append(row)
+
+    df = pd.DataFrame(data)
+
+    # Write to Excel
+    output_file = output_dir / "financial_report_q1.xlsx"
+    with pd.ExcelWriter(output_file, engine="openpyxl") as writer:
+        df.to_excel(writer, sheet_name="Q1 Summary", index=False)
+
+        # Format it nicely
+        worksheet = writer.sheets["Q1 Summary"]
+        for cell in worksheet[1]:  # Header row
+            cell.font = Font(bold=True)
+            cell.fill = PatternFill(
+                start_color="4F81BD", end_color="4F81BD", fill_type="solid"
+            )
+            cell.font = Font(color="FFFFFF", bold=True)
+
+    print(f"  ✅ Created {output_file}")
+
+
+def generate_workflow_2_data(output_dir: Path) -> None:
+    """Generate regional sales data for Workflow 2."""
+    print("\n📊 Generating Workflow 2: Regional sales data")
+
+    regions = ["northeast", "southeast", "west"]
+    products = ["Widget A", "Widget B", "Widget C", "Gadget X", "Gadget Y"]
+
+    for region in regions:
+        data = []
+        start_date = datetime(2024, 1, 1)
+
+        # Generate 90 days of sales data
+        for day in range(90):
+            date = start_date + timedelta(days=day)
+            # Random number of sales per day (3-8)
+            for _ in range(random.randint(3, 8)):
+                product = random.choice(products)
+                units_sold = random.randint(1, 20)
+                price_per_unit = random.randint(50, 200)
+                revenue = units_sold * price_per_unit
+
+                data.append(
+                    {
+                        "Date": date.strftime("%Y-%m-%d"),
+                        "Product": product,
+                        "Units_Sold": units_sold,
+                        "Revenue": revenue,
+                    }
+                )
+
+        df = pd.DataFrame(data)
+
+        # Write to Excel
+        output_file = output_dir / f"sales_{region}.xlsx"
+        df.to_excel(output_file, sheet_name="Sales", index=False)
+        print(f"  ✅ Created {output_file} ({len(df)} rows)")
+
+
+def generate_workflow_3_data(output_dir: Path) -> None:
+    """Generate complex budget spreadsheet with formatting for Workflow 3."""
+    print("\n📊 Generating Workflow 3: company_budget_2024.xlsx")
+
+    wb = Workbook()
+    ws = wb.active
+    ws.title = "Budget"
+
+    # Define departments with colors
+    departments = {
+        "Engineering": "C6E0B4",
+        "Marketing": "FFD966",
+        "Sales": "F4B084",
+        "Operations": "B4C7E7",
+    }
+
+    # Define categories
+    categories = {
+        "Personnel": ["Salaries", "Benefits", "Training"],
+        "Infrastructure": ["Office Rent", "Equipment", "Software Licenses"],
+        "Operations": ["Travel", "Supplies", "Miscellaneous"],
+    }
+
+    # Styles
+    header_font = Font(bold=True, size=12)
+    category_font = Font(bold=True, size=11)
+
+    row = 1
+
+    # Title
+    ws.merge_cells(f"A{row}:E{row}")
+    ws[f"A{row}"] = "2024 Annual Budget"
+    ws[f"A{row}"].font = Font(bold=True, size=14)
+    ws[f"A{row}"].alignment = Alignment(horizontal="center")
+    row += 2
+
+    # Headers
+    ws[f"A{row}"] = "Category"
+    ws[f"B{row}"] = "Item"
+    for i, dept in enumerate(departments.keys()):
+        ws.cell(row, 3 + i, dept)
+        ws.cell(row, 3 + i).font = header_font
+
+    for cell in ws[row]:
+        cell.font = header_font
+    row += 1
+
+    # Data
+    for category, items in categories.items():
+        # Category header (bold)
+        ws[f"A{row}"] = category
+        ws[f"A{row}"].font = category_font
+        row += 1
+
+        # Items with department budgets
+        for item in items:
+            ws[f"A{row}"] = ""
+            ws[f"B{row}"] = item
+
+            # Add budget amounts for each department (with color)
+            for i, (dept, color) in enumerate(departments.items()):
+                amount = random.randint(5000, 50000)
+                cell = ws.cell(row, 3 + i, amount)
+                cell.fill = PatternFill(
+                    start_color=color, end_color=color, fill_type="solid"
+                )
+                cell.number_format = "$#,##0"
+
+            row += 1
+
+        row += 1  # Blank row between categories
+
+    # Adjust column widths
+    ws.column_dimensions["A"].width = 20
+    ws.column_dimensions["B"].width = 25
+    for i in range(len(departments)):
+        ws.column_dimensions[chr(67 + i)].width = 15  # C, D, E, F
+
+    output_file = output_dir / "company_budget_2024.xlsx"
+    wb.save(output_file)
+    print(f"  ✅ Created {output_file}")
+    print("     • Bold categories, colored departments, merged title cell")
+
+
+def generate_workflow_4_data(output_dir: Path) -> None:
+    """Generate weekly sales report for Workflow 4."""
+    print("\n📊 Generating Workflow 4: sales_weekly.xlsx")
+
+    products = [
+        "Product A",
+        "Product B",
+        "Product C",
+        "Product D",
+        "Product E",
+        "Product F",
+        "Product G",
+        "Product H",
+    ]
+
+    # Generate one week of data
+    data = []
+    start_date = datetime(2024, 11, 4)  # Monday
+
+    for day in range(7):
+        date = start_date + timedelta(days=day)
+        # Each product has 3-10 transactions per day
+        for product in products:
+            for _ in range(random.randint(3, 10)):
+                units = random.randint(1, 15)
+                price = random.randint(20, 150)
+                revenue = units * price
+
+                data.append(
+                    {
+                        "Date": date.strftime("%Y-%m-%d"),
+                        "Product": product,
+                        "Units": units,
+                        "Revenue": revenue,
+                    }
+                )
+
+    df = pd.DataFrame(data)
+
+    # Write to Excel with some formatting
+    output_file = output_dir / "sales_weekly.xlsx"
+    with pd.ExcelWriter(output_file, engine="openpyxl") as writer:
+        df.to_excel(writer, sheet_name="Weekly Sales", index=False)
+
+        # Format header
+        worksheet = writer.sheets["Weekly Sales"]
+        for cell in worksheet[1]:
+            cell.font = Font(bold=True)
+
+    print(f"  ✅ Created {output_file} ({len(df)} rows)")
+
+
+def main() -> None:
+    """Generate all sample data files."""
+    print("=" * 60)
+    print("Generating Sample Data for LlamaSheets + Coding Agent Workflows")
+    print("=" * 60)
+
+    # Create output directory
+    output_dir = Path("input_data")
+    output_dir.mkdir(exist_ok=True)
+
+    # Generate data for each workflow
+    generate_workflow_1_data(output_dir)
+    generate_workflow_2_data(output_dir)
+    generate_workflow_3_data(output_dir)
+    generate_workflow_4_data(output_dir)
+
+    print("\n" + "=" * 60)
+    print("✅ All sample data generated!")
+    print("=" * 60)
+    print(f"\nFiles created in {output_dir.absolute()}:")
+    print("\nWorkflow 1 (Understanding a New Spreadsheet):")
+    print("  • financial_report_q1.xlsx")
+    print("\nWorkflow 2 (Generating Analysis Scripts):")
+    print("  • sales_northeast.xlsx")
+    print("  • sales_southeast.xlsx")
+    print("  • sales_west.xlsx")
+    print("\nWorkflow 3 (Using Cell Metadata):")
+    print("  • company_budget_2024.xlsx")
+    print("\nWorkflow 4 (Complete Automation):")
+    print("  • sales_weekly.xlsx")
+    print("\nYou can now use these files with the workflows in the documentation!")
+
+
+if __name__ == "__main__":
+    main()

examples/sheets/coding-agent-analysis/requirements.txt

@@ -0,0 +1,5 @@
+llama-cloud-services  # LlamaSheets SDK
+pandas>=2.0.0
+pyarrow>=12.0.0
+openpyxl>=3.0.0       # For Excel file support
+matplotlib>=3.7.0     # For visualizations (optional)

examples/sheets/coding-agent-analysis/scripts/extract.py

@@ -0,0 +1,100 @@
+"""Helper script to extract spreadsheets using LlamaSheets."""
+
+import asyncio
+import json
+import os
+import dotenv
+from pathlib import Path
+
+from llama_cloud_services.beta.sheets import LlamaSheets
+from llama_cloud_services.beta.sheets.types import (
+    SpreadsheetParsingConfig,
+    SpreadsheetResultType,
+)
+
+dotenv.load_dotenv()
+
+
+async def extract_spreadsheet(
+    file_path: str, output_dir: str = "data", generate_metadata: bool = True
+) -> dict:
+    """Extract a spreadsheet using LlamaSheets."""
+
+    client = LlamaSheets(
+        base_url="https://api.cloud.llamaindex.ai",
+        api_key=os.getenv("LLAMA_CLOUD_API_KEY"),
+    )
+
+    print(f"Extracting {file_path}...")
+
+    # Extract regions
+    config = SpreadsheetParsingConfig(
+        sheet_names=None,  # Extract all sheets
+        generate_additional_metadata=generate_metadata,
+    )
+
+    job_result = await client.aextract_regions(file_path, config=config)
+
+    print(f"Extracted {len(job_result.regions)} region(s)")
+
+    # Create output directory
+    output_path = Path(output_dir)
+    output_path.mkdir(parents=True, exist_ok=True)
+
+    # Get base name for files
+    base_name = Path(file_path).stem
+
+    # Save job metadata
+    job_metadata_path = output_path / f"{base_name}_job_metadata.json"
+    with open(job_metadata_path, "w") as f:
+        json.dump(job_result.model_dump(mode="json"), f, indent=2)
+    print(f"Saved job metadata to {job_metadata_path}")
+
+    # Download each region
+    for idx, region in enumerate(job_result.regions, 1):
+        sheet_name = region.sheet_name.replace(" ", "_")
+
+        # Download region data
+        region_bytes = await client.adownload_region_result(
+            job_id=job_result.id,
+            region_id=region.region_id,
+            result_type=region.region_type,
+        )
+
+        region_path = output_path / f"{base_name}_region_{idx}_{sheet_name}.parquet"
+        with open(region_path, "wb") as f:
+            f.write(region_bytes)
+        print(f"  Table {idx}: {region_path}")
+
+        # Download metadata
+        metadata_bytes = await client.adownload_region_result(
+            job_id=job_result.id,
+            region_id=region.region_id,
+            result_type=SpreadsheetResultType.CELL_METADATA,
+        )
+
+        metadata_path = output_path / f"{base_name}_metadata_{idx}_{sheet_name}.parquet"
+        with open(metadata_path, "wb") as f:
+            f.write(metadata_bytes)
+        print(f"  Metadata {idx}: {metadata_path}")
+
+    print(f"\nAll files saved to {output_path}/")
+
+    return job_result.model_dump(mode="json")
+
+
+if __name__ == "__main__":
+    import sys
+
+    if len(sys.argv) < 2:
+        print("Usage: python scripts/extract.py <spreadsheet_file>")
+        sys.exit(1)
+
+    file_path = sys.argv[1]
+
+    if not Path(file_path).exists():
+        print(f"❌ File not found: {file_path}")
+        sys.exit(1)
+
+    result = asyncio.run(extract_spreadsheet(file_path))
+    print(f"\n✅ Extraction complete! Job ID: {result['id']}")

examples/sheets/llama-index-agent/generate_sample_data.py

@@ -0,0 +1,278 @@
+"""
+Generate sample spreadsheets for LlamaSheets + LlamaIndex Agent workflows.
+
+This script creates example Excel files that demonstrate different use cases:
+1. Simple data table (for Workflow 1)
+2. Regional sales data (for Workflow 2)
+3. Complex budget with formatting (for Workflow 3)
+4. Weekly sales report (for Workflow 4)
+
+Usage:
+    python generate_sample_data.py
+"""
+
+import random
+from datetime import datetime, timedelta
+from pathlib import Path
+
+import pandas as pd
+from openpyxl import Workbook
+from openpyxl.styles import Font, PatternFill, Alignment
+
+
+def generate_workflow_1_data(output_dir: Path) -> None:
+    """Generate simple financial report for Workflow 1."""
+    print("📊 Generating Workflow 1: financial_report_q1.xlsx")
+
+    # Create sample quarterly data
+    months = ["January", "February", "March"]
+    categories = ["Revenue", "Cost of Goods Sold", "Operating Expenses", "Net Income"]
+
+    data = []
+    for category in categories:
+        row: dict[str, str | int] = {"Category": category}
+        for month in months:
+            if category == "Revenue":
+                value = random.randint(80000, 120000)
+            elif category == "Cost of Goods Sold":
+                value = random.randint(30000, 50000)
+            elif category == "Operating Expenses":
+                value = random.randint(20000, 35000)
+            else:  # Net Income
+                value = int(
+                    int(row.get("January", 0))
+                    + int(row.get("February", 0))
+                    + int(row.get("March", 0))
+                )
+                value = random.randint(15000, 40000)
+            row[month] = value
+        data.append(row)
+
+    df = pd.DataFrame(data)
+
+    # Write to Excel
+    output_file = output_dir / "financial_report_q1.xlsx"
+    with pd.ExcelWriter(output_file, engine="openpyxl") as writer:
+        df.to_excel(writer, sheet_name="Q1 Summary", index=False)
+
+        # Format it nicely
+        worksheet = writer.sheets["Q1 Summary"]
+        for cell in worksheet[1]:  # Header row
+            cell.font = Font(bold=True)
+            cell.fill = PatternFill(
+                start_color="4F81BD", end_color="4F81BD", fill_type="solid"
+            )
+            cell.font = Font(color="FFFFFF", bold=True)
+
+    print(f"  ✅ Created {output_file}")
+
+
+def generate_workflow_2_data(output_dir: Path) -> None:
+    """Generate regional sales data for Workflow 2."""
+    print("\n📊 Generating Workflow 2: Regional sales data")
+
+    regions = ["northeast", "southeast", "west"]
+    products = ["Widget A", "Widget B", "Widget C", "Gadget X", "Gadget Y"]
+
+    for region in regions:
+        data = []
+        start_date = datetime(2024, 1, 1)
+
+        # Generate 90 days of sales data
+        for day in range(90):
+            date = start_date + timedelta(days=day)
+            # Random number of sales per day (3-8)
+            for _ in range(random.randint(3, 8)):
+                product = random.choice(products)
+                units_sold = random.randint(1, 20)
+                price_per_unit = random.randint(50, 200)
+                revenue = units_sold * price_per_unit
+
+                data.append(
+                    {
+                        "Date": date.strftime("%Y-%m-%d"),
+                        "Product": product,
+                        "Units_Sold": units_sold,
+                        "Revenue": revenue,
+                    }
+                )
+
+        df = pd.DataFrame(data)
+
+        # Write to Excel
+        output_file = output_dir / f"sales_{region}.xlsx"
+        df.to_excel(output_file, sheet_name="Sales", index=False)
+        print(f"  ✅ Created {output_file} ({len(df)} rows)")
+
+
+def generate_workflow_3_data(output_dir: Path) -> None:
+    """Generate complex budget spreadsheet with formatting for Workflow 3."""
+    print("\n📊 Generating Workflow 3: company_budget_2024.xlsx")
+
+    wb = Workbook()
+    ws = wb.active
+    ws.title = "Budget"
+
+    # Define departments with colors
+    departments = {
+        "Engineering": "C6E0B4",
+        "Marketing": "FFD966",
+        "Sales": "F4B084",
+        "Operations": "B4C7E7",
+    }
+
+    # Define categories
+    categories = {
+        "Personnel": ["Salaries", "Benefits", "Training"],
+        "Infrastructure": ["Office Rent", "Equipment", "Software Licenses"],
+        "Operations": ["Travel", "Supplies", "Miscellaneous"],
+    }
+
+    # Styles
+    header_font = Font(bold=True, size=12)
+    category_font = Font(bold=True, size=11)
+
+    row = 1
+
+    # Title
+    ws.merge_cells(f"A{row}:E{row}")
+    ws[f"A{row}"] = "2024 Annual Budget"
+    ws[f"A{row}"].font = Font(bold=True, size=14)
+    ws[f"A{row}"].alignment = Alignment(horizontal="center")
+    row += 2
+
+    # Headers
+    ws[f"A{row}"] = "Category"
+    ws[f"B{row}"] = "Item"
+    for i, dept in enumerate(departments.keys()):
+        ws.cell(row, 3 + i, dept)
+        ws.cell(row, 3 + i).font = header_font
+
+    for cell in ws[row]:
+        cell.font = header_font
+    row += 1
+
+    # Data
+    for category, items in categories.items():
+        # Category header (bold)
+        ws[f"A{row}"] = category
+        ws[f"A{row}"].font = category_font
+        row += 1
+
+        # Items with department budgets
+        for item in items:
+            ws[f"A{row}"] = ""
+            ws[f"B{row}"] = item
+
+            # Add budget amounts for each department (with color)
+            for i, (dept, color) in enumerate(departments.items()):
+                amount = random.randint(5000, 50000)
+                cell = ws.cell(row, 3 + i, amount)
+                cell.fill = PatternFill(
+                    start_color=color, end_color=color, fill_type="solid"
+                )
+                cell.number_format = "$#,##0"
+
+            row += 1
+
+        row += 1  # Blank row between categories
+
+    # Adjust column widths
+    ws.column_dimensions["A"].width = 20
+    ws.column_dimensions["B"].width = 25
+    for i in range(len(departments)):
+        ws.column_dimensions[chr(67 + i)].width = 15  # C, D, E, F
+
+    output_file = output_dir / "company_budget_2024.xlsx"
+    wb.save(output_file)
+    print(f"  ✅ Created {output_file}")
+    print("     • Bold categories, colored departments, merged title cell")
+
+
+def generate_workflow_4_data(output_dir: Path) -> None:
+    """Generate weekly sales report for Workflow 4."""
+    print("\n📊 Generating Workflow 4: sales_weekly.xlsx")
+
+    products = [
+        "Product A",
+        "Product B",
+        "Product C",
+        "Product D",
+        "Product E",
+        "Product F",
+        "Product G",
+        "Product H",
+    ]
+
+    # Generate one week of data
+    data = []
+    start_date = datetime(2024, 11, 4)  # Monday
+
+    for day in range(7):
+        date = start_date + timedelta(days=day)
+        # Each product has 3-10 transactions per day
+        for product in products:
+            for _ in range(random.randint(3, 10)):
+                units = random.randint(1, 15)
+                price = random.randint(20, 150)
+                revenue = units * price
+
+                data.append(
+                    {
+                        "Date": date.strftime("%Y-%m-%d"),
+                        "Product": product,
+                        "Units": units,
+                        "Revenue": revenue,
+                    }
+                )
+
+    df = pd.DataFrame(data)
+
+    # Write to Excel with some formatting
+    output_file = output_dir / "sales_weekly.xlsx"
+    with pd.ExcelWriter(output_file, engine="openpyxl") as writer:
+        df.to_excel(writer, sheet_name="Weekly Sales", index=False)
+
+        # Format header
+        worksheet = writer.sheets["Weekly Sales"]
+        for cell in worksheet[1]:
+            cell.font = Font(bold=True)
+
+    print(f"  ✅ Created {output_file} ({len(df)} rows)")
+
+
+def main() -> None:
+    """Generate all sample data files."""
+    print("=" * 60)
+    print("Generating Sample Data for LlamaSheets + Coding Agent Workflows")
+    print("=" * 60)
+
+    # Create output directory
+    output_dir = Path("input_data")
+    output_dir.mkdir(exist_ok=True)
+
+    # Generate data for each workflow
+    generate_workflow_1_data(output_dir)
+    generate_workflow_2_data(output_dir)
+    generate_workflow_3_data(output_dir)
+    generate_workflow_4_data(output_dir)
+
+    print("\n" + "=" * 60)
+    print("✅ All sample data generated!")
+    print("=" * 60)
+    print(f"\nFiles created in {output_dir.absolute()}:")
+    print("\nWorkflow 1 (Understanding a New Spreadsheet):")
+    print("  • financial_report_q1.xlsx")
+    print("\nWorkflow 2 (Generating Analysis Scripts):")
+    print("  • sales_northeast.xlsx")
+    print("  • sales_southeast.xlsx")
+    print("  • sales_west.xlsx")
+    print("\nWorkflow 3 (Using Cell Metadata):")
+    print("  • company_budget_2024.xlsx")
+    print("\nWorkflow 4 (Complete Automation):")
+    print("  • sales_weekly.xlsx")
+    print("\nYou can now use these files with the workflows in the documentation!")
+
+
+if __name__ == "__main__":
+    main()

examples/sheets/llama-index-agent/llama_index_example.py

@@ -0,0 +1,308 @@
+"""
+LlamaSheets Agent with LlamaIndex
+
+This example shows how to build an agent that can work with spreadsheet data
+extracted by LlamaSheets using Python code execution.
+
+The agent has minimal tools but maximum flexibility - it can execute arbitrary
+pandas code against the extracted data, similar to a coding agent.
+
+NOTE: Code execution should be handled safely in a sandboxed environment for security.
+"""
+
+import io
+import json
+import sys
+from pathlib import Path
+from typing import Any, Dict, Optional
+
+import dotenv
+import pandas as pd
+from llama_index.core.agent import FunctionAgent, ToolCall, ToolCallResult, AgentStream
+from llama_index.llms.openai import OpenAI
+from workflows import Context
+
+dotenv.load_dotenv()
+
+# Global context for loaded dataframes
+_dataframe_context: Dict[str, Any] = {}
+
+
+# Helper function for initial agent context
+def list_extracted_data(data_dir: str = "data") -> str:
+    """
+    List all regions and metadata files extracted by LlamaSheets.
+
+    This helps discover what data is available to work with.
+
+    Args:
+        data_dir: Directory containing extracted parquet files (default: "data")
+
+    Returns:
+        JSON string with information about available files
+    """
+    data_path = Path(data_dir)
+
+    if not data_path.exists():
+        return json.dumps({"error": f"Data directory '{data_dir}' not found"})
+
+    # Find all parquet and metadata files
+    region_files = list(data_path.glob("*_region_*.parquet"))
+    job_metadata_files = list(data_path.glob("*_job_metadata.json"))
+
+    regions = []
+    for region_file in region_files:
+        # Quick peek at dimensions
+        df = pd.read_parquet(region_file)
+
+        # Find corresponding metadata file
+        base_name = region_file.stem.replace("_region_", "_metadata_")
+        metadata_path = region_file.parent / f"{base_name}.parquet"
+
+        regions.append(
+            {
+                "region_file": str(region_file),
+                "metadata_file": str(metadata_path) if metadata_path.exists() else None,
+                "shape": {"rows": len(df), "columns": len(df.columns)},
+                "columns": list(df.columns),
+            }
+        )
+
+    result = {
+        "data_directory": str(data_path.absolute()),
+        "num_regions": len(regions),
+        "regions": regions,
+        "job_metadata_files": [str(f) for f in job_metadata_files],
+    }
+
+    return json.dumps(result, indent=2)
+
+
+# Agent tool for code execution against dataframes
+def execute_dataframe_code(
+    code: str, load_files: Optional[Dict[str, str]] = None
+) -> str:
+    """
+        Execute Python pandas code against LlamaSheets extracted data.
+
+        This tool allows flexible data analysis by executing arbitrary pandas code.
+        You can load parquet files, manipulate dataframes, and return results.
+
+        The code executes in a context where:
+        - pandas is available as 'pd'
+        - json is available for formatting output
+        - Previously loaded dataframes are accessible by their variable names
+
+        Args:
+            code: Python code to execute. Any print() statements or stdout/stderr
+                  will be captured and returned. Optionally set a 'result' variable
+                  for structured output.
+            load_files: Optional dict mapping variable names to file paths to load
+                       Example: {"df": "data/sales_region_1.parquet",
+                                "meta": "data/sales_metadata_1.parquet"}
+
+        Returns:
+            String containing:
+            - Any stdout/stderr output from the code execution
+            - The 'result' variable if it was set (formatted appropriately)
+            - Error message if execution failed
+
+        Example usage:
+            code = '''
+    # Load and inspect data
+    df = pd.read_parquet("data/sales_region_1.parquet")
+    print(f"Loaded {len(df)} rows")
+
+    result = {
+        "shape": df.shape,
+        "columns": list(df.columns),
+        "sample": df.head(3).to_dict(orient="records")
+    }
+            '''
+    """
+    global _dataframe_context
+
+    # Capture stdout and stderr
+    stdout_capture = io.StringIO()
+    stderr_capture = io.StringIO()
+    old_stdout = sys.stdout
+    old_stderr = sys.stderr
+
+    try:
+        # Redirect stdout/stderr
+        sys.stdout = stdout_capture
+        sys.stderr = stderr_capture
+
+        # Create execution context with pandas, json, and previously loaded dfs
+        exec_context = {
+            "pd": pd,
+            "json": json,
+            "Path": Path,
+            **_dataframe_context,  # Include previously loaded dataframes
+        }
+
+        # Load any requested files into context
+        if load_files:
+            for var_name, file_path in load_files.items():
+                if file_path.endswith(".parquet"):
+                    exec_context[var_name] = pd.read_parquet(file_path)
+                    # Also save to global context for future calls
+                    _dataframe_context[var_name] = exec_context[var_name]
+                elif file_path.endswith(".json"):
+                    with open(file_path, "r") as f:
+                        exec_context[var_name] = json.load(f)
+                        _dataframe_context[var_name] = exec_context[var_name]
+
+        # Execute the code
+        exec(code, exec_context)
+
+        # Restore stdout/stderr
+        sys.stdout = old_stdout
+        sys.stderr = old_stderr
+
+        # Collect output
+        stdout_output = stdout_capture.getvalue()
+        stderr_output = stderr_capture.getvalue()
+
+        output_parts = []
+
+        # Add stdout if any
+        if stdout_output:
+            output_parts.append(f"<stdout>{stdout_output}</stdout>")
+
+        # Add stderr if any
+        if stderr_output:
+            output_parts.append(f"<stderr>{stderr_output}</stderr>")
+
+        # Try to get a result (if code set a 'result' variable)
+        if "result" in exec_context:
+            result = exec_context["result"]
+            result_str = None
+
+            if isinstance(result, pd.DataFrame):
+                # Convert DataFrame to readable format
+                result_str = result.to_string()
+            elif isinstance(result, (dict, list)):
+                result_str = json.dumps(result, indent=2, default=str)
+            else:
+                result_str = str(result)
+
+            if result_str:
+                output_parts.append(f"<result_var>{result_str}</result_var>")
+
+        # Return combined output or success message
+        if output_parts:
+            return "\n\n".join(output_parts)
+        else:
+            return "Code executed successfully (no output or result)"
+
+    except Exception as e:
+        # Restore stdout/stderr in case of error
+        sys.stdout = old_stdout
+        sys.stderr = old_stderr
+
+        # Get any partial output
+        stdout_output = stdout_capture.getvalue()
+        stderr_output = stderr_capture.getvalue()
+
+        error_parts = []
+        if stdout_output:
+            error_parts.append(f"=== STDOUT (before error) ===\n{stdout_output}")
+        if stderr_output:
+            error_parts.append(f"=== STDERR (before error) ===\n{stderr_output}")
+
+        error_parts.append(f"=== ERROR ===\n{str(e)}")
+        error_parts.append(f"\n=== CODE ===\n{code}")
+
+        return "\n\n".join(error_parts)
+
+
+def create_llamasheets_agent(
+    llm_model: str = "gpt-4.1", api_key: Optional[str] = None
+) -> FunctionAgent:
+    # Initialize LLM
+    llm = OpenAI(model=llm_model, api_key=api_key)
+
+    # Create tools - just 4 simple but powerful tools
+    tools = [execute_dataframe_code]
+
+    # System prompt to guide the agent
+    available_regions = list_extracted_data()
+    system_prompt = f"""You are an AI assistant that helps analyze spreadsheet data extracted by LlamaSheets.
+
+LlamaSheets extracts messy spreadsheets into clean parquet files with two types of outputs:
+1. Region files (*_region_*.parquet) - The actual data with columns and rows
+2. Metadata files (*_metadata_*.parquet) - Rich cell-level metadata including:
+   - Formatting: font_bold, font_italic, font_size, background_color_rgb
+   - Position: row_number, column_number, coordinate
+   - Type detection: data_type, is_date_like, is_percentage, is_currency
+   - Layout: is_in_first_row, is_merged_cell, horizontal_alignment
+
+Your approach:
+1. Use list_extracted_data() to discover available files
+2. Use execute_dataframe_code() to load and analyze data with pandas
+3. Use metadata to understand structure (bold = headers, colors = groups)
+4. Use save_dataframe() to export results
+
+Key tips:
+- Bold cells in metadata often indicate headers
+- Background colors often indicate groupings or departments
+- Load both region and metadata files for complete analysis
+- Write clear pandas code - you have full pandas functionality available
+- Store results in variables for reuse across multiple code executions
+
+Existing Processed Regions:
+{available_regions}
+"""
+
+    # Configure agent
+    return FunctionAgent(tools=tools, llm=llm, system_prompt=system_prompt)
+
+
+async def main():
+    """Example of using the LlamaSheets agent."""
+
+    # Create the agent
+    agent = create_llamasheets_agent()
+    ctx = Context(agent)
+
+    # Example queries the agent can handle:
+    queries = [
+        # Discovery
+        "What spreadsheet data is available?",
+        # Simple analysis
+        "Load the sales data and show me the first few rows with column info",
+        # Using metadata
+        "Find all bold cells in the metadata - these are likely headers",
+    ]
+
+    # Example: Run a query
+    for query in queries:
+        print(f"\n=== Query: {query} ===")
+        handler = agent.run(query, ctx=ctx)
+        async for ev in handler.stream_events():
+            if isinstance(ev, ToolCall):
+                tool_kwargs_str = (
+                    str(ev.tool_kwargs)[:500] + " ..."
+                    if len(str(ev.tool_kwargs)) > 500
+                    else str(ev.tool_kwargs)
+                )
+                print(f"\n[Tool Call] {ev.tool_name} with args:\n{tool_kwargs_str}\n\n")
+            elif isinstance(ev, ToolCallResult):
+                result_str = (
+                    str(ev.tool_output)[:500] + " ..."
+                    if len(str(ev.tool_output)) > 500
+                    else str(ev.tool_output)
+                )
+                print(f"\n[Tool Result] {ev.tool_name}:\n{result_str}\n\n")
+            elif isinstance(ev, AgentStream):
+                print(ev.delta, end="", flush=True)
+
+        _ = await handler
+        print("=== End Query ===\n")
+
+
+if __name__ == "__main__":
+    import asyncio
+
+    asyncio.run(main())

examples/sheets/llama-index-agent/requirements.txt

@@ -0,0 +1,7 @@
+llama-cloud-services  # LlamaSheets SDK
+llama-index-core
+llama-index-llms-openai
+pandas>=2.0.0
+pyarrow>=12.0.0
+openpyxl>=3.0.0       # For Excel file support
+matplotlib>=3.7.0     # For visualizations (optional)

examples/sheets/llama-index-agent/scripts/extract.py

@@ -0,0 +1,100 @@
+"""Helper script to extract spreadsheets using LlamaSheets."""
+
+import asyncio
+import json
+import os
+import dotenv
+from pathlib import Path
+
+from llama_cloud_services.beta.sheets import LlamaSheets
+from llama_cloud_services.beta.sheets.types import (
+    SpreadsheetParsingConfig,
+    SpreadsheetResultType,
+)
+
+dotenv.load_dotenv()
+
+
+async def extract_spreadsheet(
+    file_path: str, output_dir: str = "data", generate_metadata: bool = True
+) -> dict:
+    """Extract a spreadsheet using LlamaSheets."""
+
+    client = LlamaSheets(
+        base_url="https://api.cloud.llamaindex.ai",
+        api_key=os.getenv("LLAMA_CLOUD_API_KEY"),
+    )
+
+    print(f"Extracting {file_path}...")
+
+    # Extract regions
+    config = SpreadsheetParsingConfig(
+        sheet_names=None,  # Extract all sheets
+        generate_additional_metadata=generate_metadata,
+    )
+
+    job_result = await client.aextract_regions(file_path, config=config)
+
+    print(f"Extracted {len(job_result.regions)} region(s)")
+
+    # Create output directory
+    output_path = Path(output_dir)
+    output_path.mkdir(parents=True, exist_ok=True)
+
+    # Get base name for files
+    base_name = Path(file_path).stem
+
+    # Save job metadata
+    job_metadata_path = output_path / f"{base_name}_job_metadata.json"
+    with open(job_metadata_path, "w") as f:
+        json.dump(job_result.model_dump(mode="json"), f, indent=2)
+    print(f"Saved job metadata to {job_metadata_path}")
+
+    # Download each region
+    for idx, region in enumerate(job_result.regions, 1):
+        sheet_name = region.sheet_name.replace(" ", "_")
+
+        # Download region data
+        region_bytes = await client.adownload_region_result(
+            job_id=job_result.id,
+            region_id=region.region_id,
+            result_type=region.region_type,
+        )
+
+        region_path = output_path / f"{base_name}_region_{idx}_{sheet_name}.parquet"
+        with open(region_path, "wb") as f:
+            f.write(region_bytes)
+        print(f"  Table {idx}: {region_path}")
+
+        # Download metadata
+        metadata_bytes = await client.adownload_region_result(
+            job_id=job_result.id,
+            region_id=region.region_id,
+            result_type=SpreadsheetResultType.CELL_METADATA,
+        )
+
+        metadata_path = output_path / f"{base_name}_metadata_{idx}_{sheet_name}.parquet"
+        with open(metadata_path, "wb") as f:
+            f.write(metadata_bytes)
+        print(f"  Metadata {idx}: {metadata_path}")
+
+    print(f"\nAll files saved to {output_path}/")
+
+    return job_result.model_dump(mode="json")
+
+
+if __name__ == "__main__":
+    import sys
+
+    if len(sys.argv) < 2:
+        print("Usage: python scripts/extract.py <spreadsheet_file>")
+        sys.exit(1)
+
+    file_path = sys.argv[1]
+
+    if not Path(file_path).exists():
+        print(f"❌ File not found: {file_path}")
+        sys.exit(1)
+
+    result = asyncio.run(extract_spreadsheet(file_path))
+    print(f"\n✅ Extraction complete! Job ID: {result['id']}")

py/llama_cloud_services/beta/sheets/__init__.py

@@ -0,0 +1,43 @@
+"""LlamaCloud Spreadsheet API SDK
+
+This module provides a Python SDK for the LlamaCloud Spreadsheet API.
+"""
+
+from llama_cloud_services.beta.sheets.client import (
+    LlamaSheets,
+    SpreadsheetAPIError,
+    SpreadsheetJobError,
+    SpreadsheetTimeoutError,
+)
+from llama_cloud_services.beta.sheets.types import (
+    ExtractedRegionSummary,
+    FileUploadResponse,
+    JobStatus,
+    PresignedUrlResponse,
+    SpreadsheetJob,
+    SpreadsheetJobResult,
+    SpreadsheetParseResult,
+    SpreadsheetParsingConfig,
+    SpreadsheetResultType,
+    WorksheetMetadata,
+)
+
+__all__ = [
+    # Client
+    "LlamaSheets",
+    # Exceptions
+    "SpreadsheetAPIError",
+    "SpreadsheetJobError",
+    "SpreadsheetTimeoutError",
+    # Types
+    "ExtractedRegionSummary",
+    "FileUploadResponse",
+    "JobStatus",
+    "PresignedUrlResponse",
+    "SpreadsheetJob",
+    "SpreadsheetJobResult",
+    "SpreadsheetParseResult",
+    "SpreadsheetParsingConfig",
+    "SpreadsheetResultType",
+    "WorksheetMetadata",
+]

py/llama_cloud_services/beta/sheets/client.py

@@ -0,0 +1,518 @@
+import asyncio
+import io
+import os
+import time
+from typing import TYPE_CHECKING
+
+import httpx
+from llama_cloud.client import AsyncLlamaCloud
+from tenacity import (
+    AsyncRetrying,
+    retry_if_exception,
+    stop_after_attempt,
+    wait_exponential,
+)
+
+from llama_cloud_services.beta.sheets.types import (
+    FileUploadResponse,
+    JobStatus,
+    PresignedUrlResponse,
+    SpreadsheetJob,
+    SpreadsheetJobResult,
+    SpreadsheetParsingConfig,
+    SpreadsheetResultType,
+)
+from llama_cloud_services.constants import BASE_URL
+from llama_cloud_services.files.client import FileClient
+from llama_cloud_services.utils import (
+    augment_async_errors,
+    FileInput,
+)
+
+if TYPE_CHECKING:
+    import pandas as pd
+
+
+def _should_retry_exception(exception: BaseException) -> bool:
+    """Determine if an exception should be retried."""
+    if isinstance(exception, httpx.HTTPStatusError):
+        return exception.response.status_code in (429, 500, 502, 503, 504)
+    return False
+
+
+class SpreadsheetAPIError(Exception):
+    """Base exception for spreadsheet API errors"""
+
+    pass
+
+
+class SpreadsheetJobError(SpreadsheetAPIError):
+    """Exception raised when a spreadsheet job fails"""
+
+    pass
+
+
+class SpreadsheetTimeoutError(SpreadsheetAPIError):
+    """Exception raised when a job times out"""
+
+    pass
+
+
+class LlamaSheets:
+    """Client for the LlamaCloud Spreadsheet API"""
+
+    def __init__(
+        self,
+        api_key: str | None = None,
+        base_url: str | None = None,
+        max_timeout: int = 300,
+        poll_interval: int = 5,
+        max_retries: int = 3,
+        async_httpx_client: httpx.AsyncClient | None = None,
+    ) -> None:
+        """Initialize the LlamaSheets client.
+
+        Args:
+            api_key: API key for authentication. If not provided, will use LLAMA_CLOUD_API_KEY env var
+            base_url: Base URL for the API
+            max_timeout: Maximum time to wait for job completion in seconds
+            poll_interval: Interval between status checks in seconds
+            max_retries: Maximum number of retries for failed requests
+            async_httpx_client: Optional custom async httpx client
+        """
+        self.api_key = api_key or os.environ.get("LLAMA_CLOUD_API_KEY")
+        if not self.api_key:
+            raise ValueError(
+                "An API key must be provided either as an argument or via the LLAMA_CLOUD_API_KEY environment variable."
+            )
+
+        base_url = base_url or os.environ.get("LLAMA_CLOUD_BASE_URL", BASE_URL)
+        self.base_url = str(base_url).rstrip("/")
+
+        self.max_timeout = max_timeout
+        self.poll_interval = poll_interval
+        self.max_retries = max_retries
+
+        self._async_client: httpx.AsyncClient | None = async_httpx_client
+        self._files_client = FileClient(
+            AsyncLlamaCloud(
+                token=self.api_key,
+                base_url=self.base_url,
+                httpx_client=async_httpx_client,
+            )
+        )
+
+    def _get_async_client(self) -> httpx.AsyncClient:
+        """Get or create the async httpx client"""
+        if self._async_client is None:
+            self._async_client = httpx.AsyncClient(
+                timeout=httpx.Timeout(60.0),
+                follow_redirects=True,
+            )
+        return self._async_client
+
+    def _get_headers(self) -> dict[str, str]:
+        """Get common headers for API requests"""
+        return {
+            "Authorization": f"Bearer {self.api_key}",
+            "Content-Type": "application/json",
+        }
+
+    # Sync methods
+
+    def upload_file(
+        self, file_obj: FileInput, file_name: str | None = None
+    ) -> FileUploadResponse:
+        """Upload a file to the Files API.
+
+        Args:
+            file_obj: File to upload (path, bytes, or file-like object)
+            file_name: Optional name for the uploaded filename
+
+        Returns:
+            FileUploadResponse with the uploaded file ID
+        """
+        with augment_async_errors():
+            return asyncio.run(self.aupload_file(file_obj))
+
+    def create_job(
+        self,
+        file_id: str,
+        config: dict | SpreadsheetParsingConfig | None = None,
+    ) -> SpreadsheetJob:
+        """Create a new spreadsheet parsing job.
+
+        Args:
+            file_id: ID of the uploaded file
+            config: Parsing configuration
+
+        Returns:
+            SpreadsheetJob with job details
+        """
+        with augment_async_errors():
+            return asyncio.run(self.acreate_job(file_id, config))
+
+    def get_job(
+        self, job_id: str, include_results_metadata: bool = True
+    ) -> SpreadsheetJobResult:
+        """Get the status of a spreadsheet parsing job.
+
+        Args:
+            job_id: ID of the job
+            include_results_metadata: Whether to include results metadata in the response
+
+        Returns:
+            SpreadsheetJobResult with job status and optionally results
+        """
+        with augment_async_errors():
+            return asyncio.run(self.aget_job(job_id, include_results_metadata))
+
+    def wait_for_completion(self, job_id: str) -> SpreadsheetJobResult:
+        """Wait for a job to complete by polling.
+
+        Args:
+            job_id: ID of the job to wait for
+
+        Returns:
+            SpreadsheetJobResult when job is complete
+
+        Raises:
+            SpreadsheetTimeoutError: If job doesn't complete within max_timeout
+            SpreadsheetJobError: If job fails
+        """
+        with augment_async_errors():
+            return asyncio.run(self.await_for_completion(job_id))
+
+    def download_region_result(
+        self,
+        job_id: str,
+        region_id: str,
+        result_type: SpreadsheetResultType = SpreadsheetResultType.TABLE,
+    ) -> bytes:
+        """Download a region result (either region data or cell metadata).
+
+        Args:
+            job_id: ID of the job
+            region_id: ID of the region
+            result_type: Type of result to download (region or cell_metadata)
+
+        Returns:
+            Raw bytes of the parquet file
+        """
+        with augment_async_errors():
+            return asyncio.run(
+                self.adownload_region_result(job_id, region_id, result_type)
+            )
+
+    def download_region_as_dataframe(
+        self,
+        job_id: str,
+        region_id: str,
+        result_type: SpreadsheetResultType = SpreadsheetResultType.TABLE,
+    ) -> "pd.DataFrame":
+        """Download a region result as a pandas DataFrame.
+
+        Args:
+            job_id: ID of the job
+            region_id: ID of the region
+            result_type: Type of result to download (region or cell_metadata)
+
+        Returns:
+            pandas DataFrame
+        """
+        with augment_async_errors():
+            return asyncio.run(
+                self.adownload_region_as_dataframe(job_id, region_id, result_type)
+            )
+
+    def extract_regions(
+        self,
+        file_obj: FileInput,
+        config: dict | SpreadsheetParsingConfig | None = None,
+    ) -> SpreadsheetJobResult:
+        """High-level method to parse a spreadsheet file.
+
+        This method handles the entire workflow:
+        1. Upload the file
+        2. Create a parsing job
+        3. Wait for completion
+        4. Return results
+
+        Args:
+            file_obj: File to parse (path, bytes, or file-like object)
+            config: Parsing configuration
+
+        Returns:
+            SpreadsheetJobResult with parsing results
+        """
+        with augment_async_errors():
+            return asyncio.run(self.aextract_regions(file_obj, config))
+
+    # Async methods
+
+    async def aupload_file(
+        self, file_obj: FileInput, file_name: str | None = None
+    ) -> FileUploadResponse:
+        """Upload a file to the Files API.
+
+        Args:
+            file_obj: File to upload (path, bytes, or file-like object)
+            file_name: Optional name for the uploaded filename
+
+        Returns:
+            FileUploadResponse with the uploaded file ID
+        """
+        try:
+            async for attempt in AsyncRetrying(
+                stop=stop_after_attempt(self.max_retries),
+                wait=wait_exponential(multiplier=1, min=1, max=32),
+                retry=retry_if_exception(_should_retry_exception),
+                reraise=True,
+            ):
+                with attempt:
+                    return await self._files_client.upload_content(
+                        file_obj, external_file_id=file_name
+                    )
+        except Exception as e:
+            raise SpreadsheetAPIError(f"Failed to upload file: {e}") from e
+        raise RuntimeError("Tenacity did not execute")
+
+    async def acreate_job(
+        self,
+        file_id: str,
+        config: dict | SpreadsheetParsingConfig | None = None,
+    ) -> SpreadsheetJob:
+        """Create a new spreadsheet parsing job.
+
+        Args:
+            file_id: ID of the uploaded file
+            config: Parsing configuration
+
+        Returns:
+            SpreadsheetJob with job details
+        """
+        if config is None:
+            config = SpreadsheetParsingConfig()
+        elif isinstance(config, dict):
+            config = SpreadsheetParsingConfig.model_validate(config)
+
+        if not isinstance(config, SpreadsheetParsingConfig):
+            raise ValueError(
+                "config must be a dict or SpreadsheetParsingConfig instance"
+            )
+
+        payload = {
+            "file_id": file_id,
+            "config": config.model_dump(mode="json", exclude_none=True),
+        }
+
+        try:
+            async for attempt in AsyncRetrying(
+                stop=stop_after_attempt(self.max_retries),
+                wait=wait_exponential(multiplier=1, min=1, max=32),
+                retry=retry_if_exception(_should_retry_exception),
+                reraise=True,
+            ):
+                with attempt:
+                    client = self._get_async_client()
+                    response = await client.post(
+                        f"{self.base_url}/api/v1/beta/sheets/jobs",
+                        headers=self._get_headers(),
+                        json=payload,
+                    )
+                    response.raise_for_status()
+                    return SpreadsheetJob.model_validate(response.json())
+        except Exception as e:
+            raise SpreadsheetAPIError(f"Failed to create job: {e}") from e
+        raise RuntimeError("Tenacity did not execute")
+
+    async def aget_job(
+        self, job_id: str, include_results_metadata: bool = True
+    ) -> SpreadsheetJobResult:
+        """Get the status of a spreadsheet parsing job.
+
+        Args:
+            job_id: ID of the job
+            include_results_metadata: Whether to include results in the response
+
+        Returns:
+            SpreadsheetJobResult with job status and optionally results
+        """
+        try:
+            async for attempt in AsyncRetrying(
+                stop=stop_after_attempt(self.max_retries),
+                wait=wait_exponential(multiplier=1, min=1, max=32),
+                retry=retry_if_exception(_should_retry_exception),
+                reraise=True,
+            ):
+                with attempt:
+                    client = self._get_async_client()
+                    response = await client.get(
+                        f"{self.base_url}/api/v1/beta/sheets/jobs/{job_id}",
+                        headers=self._get_headers(),
+                        params={"include_results": include_results_metadata},
+                    )
+                    response.raise_for_status()
+                    return SpreadsheetJobResult.model_validate(response.json())
+        except Exception as e:
+            raise SpreadsheetAPIError(f"Failed to get job status: {e}") from e
+        raise RuntimeError("Tenacity did not execute")
+
+    async def await_for_completion(self, job_id: str) -> SpreadsheetJobResult:
+        """Wait for a job to complete by polling.
+
+        Args:
+            job_id: ID of the job to wait for
+
+        Returns:
+            SpreadsheetJobResult when job is complete
+
+        Raises:
+            SpreadsheetTimeoutError: If job doesn't complete within max_timeout
+            SpreadsheetJobError: If job fails
+        """
+        start_time = time.time()
+
+        while (time.time() - start_time) < self.max_timeout:
+            job_result = await self.aget_job(job_id, include_results_metadata=True)
+
+            if job_result.status in (
+                JobStatus.SUCCESS,
+                JobStatus.PARTIAL_SUCCESS,
+                JobStatus.ERROR,
+                JobStatus.FAILURE,
+            ):
+                if job_result.status in (JobStatus.SUCCESS, JobStatus.PARTIAL_SUCCESS):
+                    return job_result
+                else:
+                    error_msg = f"Job failed with status: {job_result.status}"
+                    if job_result.errors:
+                        error_msg += f"\nErrors: {', '.join(job_result.errors)}"
+                    raise SpreadsheetJobError(error_msg)
+
+            await asyncio.sleep(self.poll_interval)
+
+        raise SpreadsheetTimeoutError(
+            f"Job did not complete within {self.max_timeout} seconds"
+        )
+
+    async def adownload_region_result(
+        self,
+        job_id: str,
+        region_id: str,
+        result_type: SpreadsheetResultType = SpreadsheetResultType.TABLE,
+    ) -> bytes:
+        """Download a region result (either region data or cell metadata).
+
+        Args:
+            job_id: ID of the job
+            region_id: ID of the region
+            result_type: Type of result to download (region or cell_metadata)
+
+        Returns:
+            Raw bytes of the parquet file
+        """
+        # Get presigned URL
+        presigned_response = None
+        result_type_str = str(result_type)
+        try:
+            async for attempt in AsyncRetrying(
+                stop=stop_after_attempt(self.max_retries),
+                wait=wait_exponential(multiplier=1, min=1, max=32),
+                retry=retry_if_exception(_should_retry_exception),
+                reraise=True,
+            ):
+                with attempt:
+                    client = self._get_async_client()
+                    response = await client.get(
+                        f"{self.base_url}/api/v1/beta/sheets/jobs/{job_id}/regions/{region_id}/result/{result_type_str}",
+                        headers=self._get_headers(),
+                    )
+                    response.raise_for_status()
+                    presigned_response = PresignedUrlResponse.model_validate(
+                        response.json()
+                    )
+        except Exception as e:
+            raise SpreadsheetAPIError(f"Failed to get presigned URL: {e}") from e
+
+        # Download using presigned URL
+        if presigned_response is None:
+            raise SpreadsheetAPIError("Failed to obtain presigned URL.")
+
+        try:
+            async for attempt in AsyncRetrying(
+                stop=stop_after_attempt(self.max_retries),
+                wait=wait_exponential(multiplier=1, min=1, max=32),
+                retry=retry_if_exception(_should_retry_exception),
+                reraise=True,
+            ):
+                with attempt:
+                    download_response = await client.get(presigned_response.url)
+                    download_response.raise_for_status()
+                    return download_response.content
+        except Exception as e:
+            raise SpreadsheetAPIError(f"Failed to download result: {e}") from e
+        raise RuntimeError("Tenacity did not execute")
+
+    async def adownload_region_as_dataframe(
+        self,
+        job_id: str,
+        region_id: str,
+        result_type: SpreadsheetResultType = SpreadsheetResultType.TABLE,
+    ) -> "pd.DataFrame":
+        """Download a region result as a pandas DataFrame.
+
+        Args:
+            job_id: ID of the job
+            region_id: ID of the region
+            result_type: Type of result to download (region or cell_metadata)
+
+        Returns:
+            pandas DataFrame
+        """
+        import pandas as pd
+
+        parquet_bytes = await self.adownload_region_result(
+            job_id, region_id, result_type
+        )
+        return pd.read_parquet(io.BytesIO(parquet_bytes))
+
+    async def aextract_regions(
+        self,
+        file_obj: FileInput,
+        config: dict | SpreadsheetParsingConfig | None = None,
+    ) -> SpreadsheetJobResult:
+        """High-level method to parse a spreadsheet file.
+
+        This method handles the entire workflow:
+        1. Upload the file
+        2. Create a parsing job
+        3. Wait for completion
+        4. Return results
+
+        Args:
+            file_obj: File to parse (path, bytes, or file-like object)
+            config: Parsing configuration
+
+        Returns:
+            SpreadsheetJobResult with parsing results
+        """
+        # Upload file
+        file_response = await self.aupload_file(file_obj)
+
+        # Create job
+        job = await self.acreate_job(file_response.id, config)
+
+        # Wait for completion
+        return await self.await_for_completion(job.id)
+
+    async def aclose(self) -> None:
+        """Close all HTTP clients (async)"""
+        if self._async_client:
+            await self._async_client.aclose()
+
+    async def __aenter__(self) -> "LlamaSheets":
+        return self
+
+    async def __aexit__(self, _exc_type, _exc_val, _exc_tb) -> None:  # type: ignore
+        await self.aclose()

py/llama_cloud_services/beta/sheets/types.py

@@ -0,0 +1,156 @@
+from datetime import datetime
+from enum import Enum
+
+from pydantic import BaseModel, ConfigDict, Field, field_validator
+
+
+class SpreadsheetResultType(str, Enum):
+    TABLE = "table"
+    EXTRA = "extra"
+    CELL_METADATA = "cell_metadata"
+
+    def __str__(self) -> str:
+        return self.value
+
+
+class ExtractedRegionSummary(BaseModel):
+    """A summary of a single extracted region from a spreadsheet"""
+
+    region_id: str = Field(
+        ...,
+        description="Unique identifier for this region within the file",
+    )
+    sheet_name: str = Field(..., description="Worksheet name where region was found")
+    location: str = Field(..., description="Location of the region in the spreadsheet")
+    title: str | None = Field(None, description="Generated title for the region")
+    description: str | None = Field(
+        None, description="Generated description of the region"
+    )
+    region_type: SpreadsheetResultType = Field(
+        ..., description="Type of the extracted region"
+    )
+
+
+class WorksheetMetadata(BaseModel):
+    """Metadata about a worksheet in a spreadsheet"""
+
+    sheet_name: str = Field(..., description="Name of the worksheet")
+    title: str | None = Field(None, description="Generated title for the worksheet")
+    description: str | None = Field(
+        None, description="Generated description of the worksheet"
+    )
+
+
+class SpreadsheetParseResult(BaseModel):
+    """Result of parsing a single spreadsheet file"""
+
+    success: bool = Field(..., description="Whether parsing was successful")
+    file_name: str = Field(..., description="Original filename")
+
+    regions: list[ExtractedRegionSummary] = Field(
+        default_factory=list, description="All successfully extracted regions"
+    )
+    worksheet_metadata: list[WorksheetMetadata] = Field(
+        default_factory=list, description="Metadata for each processed worksheet"
+    )
+
+    # Error information
+    errors: list[str] = Field(
+        default_factory=list, description="Any errors encountered during parsing"
+    )
+
+
+class SpreadsheetParsingConfig(BaseModel):
+    """Configuration for spreadsheet parsing and region extraction"""
+
+    model_config = ConfigDict(extra="forbid")
+
+    sheet_names: list[str] | None = Field(
+        default=None,
+        description="The names of the sheets to extract regions from. If empty, the default sheet is extracted.",
+    )
+    include_hidden_cells: bool = Field(
+        default=True,
+        description="Whether to include hidden cells when extracting regions from the spreadsheet.",
+    )
+    extraction_range: str | None = Field(
+        default=None,
+        description="A1 notation of the range to extract a single region from. If None, the entire sheet is used.",
+    )
+    generate_additional_metadata: bool = Field(
+        default=True,
+        description="Whether to generate additional metadata (title, description) for each extracted region.",
+    )
+    use_experimental_processing: bool = Field(
+        default=False,
+        description="Enables experimental processing. Accuracy may be impacted.",
+    )
+
+
+class SpreadsheetJob(BaseModel):
+    """A spreadsheet parsing job"""
+
+    id: str = Field(..., description="The ID of the job")
+    user_id: str = Field(..., description="The ID of the user")
+    project_id: str = Field(..., description="The ID of the project")
+    file: dict = Field(..., description="The file object being parsed")
+    config: SpreadsheetParsingConfig = Field(
+        ..., description="Configuration for the parsing job"
+    )
+    status: str = Field(..., description="The status of the parsing job")
+    created_at: str = Field(..., description="When the job was created")
+    updated_at: str = Field(..., description="When the job was last updated")
+
+    @field_validator("created_at", "updated_at", mode="before")
+    def validate_dates(cls, v: str) -> str:
+        """Validate that the dates are in the correct format"""
+        if isinstance(v, datetime):
+            return v.isoformat()
+        else:
+            return v
+
+
+class SpreadsheetJobResult(SpreadsheetJob):
+    """A spreadsheet parsing job result."""
+
+    # Results are included when the job is complete
+    success: bool | None = Field(
+        None, description="Whether the job completed successfully"
+    )
+    regions: list[ExtractedRegionSummary] = Field(
+        default_factory=list,
+        description="All extracted regions (populated when job is complete)",
+    )
+    worksheet_metadata: list[WorksheetMetadata] = Field(
+        default_factory=list,
+        description="Metadata for each processed worksheet (populated when job is complete)",
+    )
+    errors: list[str] = Field(
+        default_factory=list, description="Any errors encountered"
+    )
+
+
+class JobStatus(str, Enum):
+    """Status of a spreadsheet parsing job"""
+
+    PENDING = "PENDING"
+    IN_PROGRESS = "IN_PROGRESS"
+    SUCCESS = "SUCCESS"
+    PARTIAL_SUCCESS = "PARTIAL_SUCCESS"
+    ERROR = "ERROR"
+    FAILURE = "FAILURE"
+
+
+class PresignedUrlResponse(BaseModel):
+    """Response containing a presigned URL for downloading results"""
+
+    url: str = Field(..., description="The presigned URL for downloading")
+
+
+class FileUploadResponse(BaseModel):
+    """Response from uploading a file"""
+
+    id: str = Field(..., description="The ID of the uploaded file")
+    name: str = Field(..., description="The name of the file")
+    project_id: str = Field(..., description="The project ID")
+    user_id: str = Field(..., description="The user ID")

py/llama_cloud_services/constants.py

@@ -1,2 +1,3 @@
+BASE_URL = "https://api.cloud.llamaindex.ai"
 EU_BASE_URL = "https://api.cloud.eu.llamaindex.ai"
 POLLING_TIMEOUT_SECONDS = 300.0

py/pyproject.toml

@@ -14,7 +14,10 @@ dev = [
   "ipython>=8.12.3,<9",
   "jupyter>=1.1.1,<2",
   "mypy>=1.14.1,<2",
-  "pydantic-settings>=2.10.1"
+  "pydantic-settings>=2.10.1",
+  "pandas",
+  "openpyxl",
+  "pyarrow"
 ]
 
 [project]

py/tests/beta/sheets/test_sheets.py

@@ -0,0 +1,163 @@
+import os
+import tempfile
+import pytest
+import pandas as pd
+
+from llama_cloud_services.beta.sheets import LlamaSheets
+from llama_cloud_services.beta.sheets.types import SpreadsheetParsingConfig
+
+
+@pytest.fixture
+def sheets_client():
+    """Create a LlamaSheets client for testing."""
+    api_key = os.getenv("LLAMA_CLOUD_API_KEY")
+    base_url = os.getenv("LLAMA_CLOUD_BASE_URL", "https://api.cloud.llamaindex.ai")
+
+    client = LlamaSheets(
+        api_key=api_key,
+        base_url=base_url,
+        max_timeout=300,
+        poll_interval=2,
+    )
+    return client
+
+
+@pytest.fixture
+def sample_excel_file():
+    """Create a temporary Excel file with sample data."""
+    # Create a simple dataframe with various data types
+    data = {
+        "Name": ["Alice", "Bob", "Charlie", "David", "Eve"],
+        "Age": [25, 30, 35, 40, 45],
+        "City": ["New York", "Los Angeles", "Chicago", "Houston", "Phoenix"],
+        "Salary": [50000.50, 75000.75, 100000.00, 125000.25, 150000.50],
+    }
+    df = pd.DataFrame(data)
+
+    # Create a temporary file
+    with tempfile.NamedTemporaryFile(suffix=".xlsx", delete=False) as tmp:
+        tmp_path = tmp.name
+        df.to_excel(tmp_path, index=False, sheet_name="TestSheet")
+
+    yield tmp_path
+
+    # Cleanup
+    try:
+        os.unlink(tmp_path)
+    except Exception:
+        pass
+
+
+@pytest.mark.skipif(
+    os.environ.get("LLAMA_CLOUD_API_KEY", "") == "",
+    reason="LLAMA_CLOUD_API_KEY not set",
+)
+@pytest.mark.asyncio
+async def test_spreadsheet_extraction_e2e(
+    sheets_client: LlamaSheets, sample_excel_file: str
+):
+    """End-to-end test for spreadsheet extraction.
+
+    This test:
+    1. Creates a temporary Excel file with sample data
+    2. Uploads and extracts tables from the file
+    3. Downloads the extracted table as a DataFrame
+    4. Verifies the extracted data matches the original data
+    """
+    # Extract tables from the spreadsheet
+    result = await sheets_client.aextract_tables(sample_excel_file)
+
+    # Verify job completed successfully
+    assert result.status in ("SUCCESS", "PARTIAL_SUCCESS")
+    assert result.success is True
+
+    # Verify we extracted at least one table
+    assert len(result.tables) > 0, "Expected at least one table to be extracted"
+
+    # Get the first table
+    first_table = result.tables[0]
+    assert first_table.sheet_name == "TestSheet"
+
+    # Download the table as a DataFrame
+    extracted_df = await sheets_client.adownload_table_as_dataframe(
+        job_id=result.id,
+        table_id=first_table.table_id,
+    )
+
+    # Load the original dataframe for comparison
+    original_df = pd.read_excel(sample_excel_file)
+
+    # Verify the extracted DataFrame has the expected shape
+    breakpoint()
+    assert extracted_df.shape[0] == original_df.shape[0], (
+        f"Row count mismatch: extracted {extracted_df.shape[0]}, "
+        f"original {original_df.shape[0]}"
+    )
+    assert extracted_df.shape[1] == original_df.shape[1], (
+        f"Column count mismatch: extracted {extracted_df.shape[1]}, "
+        f"original {original_df.shape[1]}"
+    )
+
+    # Verify column names match
+    assert list(extracted_df.columns) == list(original_df.columns), (
+        f"Column names mismatch: extracted {list(extracted_df.columns)}, "
+        f"original {list(original_df.columns)}"
+    )
+
+    # Verify data types are preserved (at least numerically)
+    for col in original_df.columns:
+        if original_df[col].dtype in ["int64", "float64"]:
+            assert extracted_df[col].dtype in ["int64", "float64"], (
+                f"Column {col} type mismatch: extracted {extracted_df[col].dtype}, "
+                f"original {original_df[col].dtype}"
+            )
+
+    # Verify the data values match (allowing for minor type conversions)
+    for col in original_df.columns:
+        original_values = original_df[col].tolist()
+        extracted_values = extracted_df[col].tolist()
+
+        # Convert both to strings for comparison to handle type differences
+        original_str = [str(v) for v in original_values]
+        extracted_str = [str(v) for v in extracted_values]
+
+        assert original_str == extracted_str, (
+            f"Column {col} values mismatch:\n"
+            f"Original: {original_str}\n"
+            f"Extracted: {extracted_str}"
+        )
+
+
+@pytest.mark.skipif(
+    os.environ.get("LLAMA_CLOUD_API_KEY", "") == "",
+    reason="LLAMA_CLOUD_API_KEY not set",
+)
+@pytest.mark.asyncio
+async def test_spreadsheet_extraction_with_config(
+    sheets_client: LlamaSheets, sample_excel_file: str
+):
+    """Test spreadsheet extraction with custom configuration."""
+    # Create a config with specific settings
+    config = SpreadsheetParsingConfig(
+        sheet_names=["TestSheet"],
+        include_hidden_cells=True,
+        generate_additional_metadata=True,
+    )
+
+    # Extract tables with the config
+    result = await sheets_client.aextract_tables(sample_excel_file, config=config)
+
+    # Verify job completed successfully
+    assert result.status in ("SUCCESS", "PARTIAL_SUCCESS")
+    assert result.success is True
+
+    # Verify that additional metadata was generated
+    assert len(result.worksheet_metadata) > 0
+    assert result.worksheet_metadata[0].title is not None
+    assert result.worksheet_metadata[0].description is not None
+
+    # Verify we extracted at least one table
+    assert len(result.tables) > 0
+
+    # Verify the sheet name matches
+    assert result.tables[0].sheet_name == "TestSheet"