lint

* split notebook

* Lint

examples/parse/batch_parse_folder.py

@@ -0,0 +1,183 @@
+"""
+Example: Batch Processing a Folder of PDFs with LlamaParse
+
+This script demonstrates how to process multiple PDFs from a folder
+using LlamaParse with controlled concurrency using asyncio and semaphores.
+
+Usage:
+    python batch_parse_folder.py --input-dir ./pdfs --max-concurrent 5
+"""
+
+import asyncio
+import argparse
+from pathlib import Path
+from typing import List, Dict, Any
+from datetime import datetime
+from dotenv import load_dotenv
+import os
+
+from llama_cloud_services import LlamaParse
+
+# Load environment variables from .env file
+load_dotenv()
+
+
+async def parse_single_file(
+    parser: LlamaParse,
+    file_path: Path,
+    semaphore: asyncio.Semaphore,
+) -> Dict[str, Any]:
+    """
+    Parse a single PDF file with concurrency control.
+
+    Args:
+        parser: LlamaParse instance
+        file_path: Path to the PDF file
+        semaphore: Semaphore to control concurrent requests
+
+    Returns:
+        Dictionary with file info and parse result
+    """
+    async with semaphore:
+        try:
+            print(f"Starting parse: {file_path.name}")
+
+            result = await parser.aparse(str(file_path))
+
+            print(f"✓ Completed: {file_path.name} ({len(result.pages)} pages)")
+
+            return {
+                "file": file_path.name,
+                "status": "success",
+                "result": result,
+                "pages": len(result.pages) if result.pages else 0,
+            }
+        except Exception as e:
+            print(f"✗ Error parsing {file_path.name}: {str(e)}")
+            return {
+                "file": file_path.name,
+                "status": "error",
+                "error": str(e),
+            }
+
+
+async def parse_folder(
+    input_dir: Path,
+    max_concurrent: int = 5,
+    api_key: str = None,
+) -> List[Dict[str, any]]:
+    """
+    Parse all PDFs in a folder with controlled concurrency.
+
+    Args:
+        input_dir: Directory containing PDF files
+        max_concurrent: Maximum number of concurrent parse operations
+        api_key: LlamaCloud API key (loaded from .env file)
+
+    Returns:
+        List of parse results for each file
+    """
+    # Find all PDF files
+    pdf_files = list(input_dir.glob("*.pdf"))
+
+    if not pdf_files:
+        print(f"No PDF files found in {input_dir}")
+        return []
+
+    print(f"Found {len(pdf_files)} PDF files to parse")
+
+    # Initialize parser
+    parser = LlamaParse(
+        api_key=api_key,
+        num_workers=1,  # We control concurrency with semaphore
+        show_progress=False,  # We'll show our own progress
+    )
+
+    # Create semaphore to limit concurrent requests
+    semaphore = asyncio.Semaphore(max_concurrent)
+
+    # Create tasks for all files
+    tasks = [parse_single_file(parser, pdf_file, semaphore) for pdf_file in pdf_files]
+
+    # Run all tasks concurrently (but limited by semaphore)
+    print(
+        f"Processing {len(tasks)} files with max {max_concurrent} concurrent operations..."
+    )
+    start_time = datetime.now()
+
+    results = await asyncio.gather(*tasks)
+
+    end_time = datetime.now()
+    duration = (end_time - start_time).total_seconds()
+
+    # Process results
+    successful = [
+        r for r in results if isinstance(r, dict) and r.get("status") == "success"
+    ]
+    failed = [r for r in results if isinstance(r, dict) and r.get("status") == "error"]
+
+    # Print summary
+    print("PARSE SUMMARY \n")
+    print(f"Total files: {len(pdf_files)}")
+    print(f"Successful: {len(successful)}")
+    print(f"Failed: {len(failed)}")
+    print(f"Total time: {duration:.2f} seconds")
+    print(f"Average time per file: {duration / len(pdf_files):.2f} seconds")
+
+    if failed:
+        print("\nFailed files:")
+        for result in failed:
+            print(f"  - {result['file']}: {result.get('error', 'Unknown error')}")
+
+    return results
+
+
+def main():
+    """Main entry point for the script."""
+    parser = argparse.ArgumentParser(
+        description="Batch process PDFs in a folder with LlamaParse"
+    )
+    parser.add_argument(
+        "--input-dir",
+        type=str,
+        required=True,
+        help="Directory containing PDF files to parse",
+    )
+    parser.add_argument(
+        "--max-concurrent",
+        type=int,
+        default=5,
+        help="Maximum number of concurrent parse operations (default: 5)",
+    )
+
+    args = parser.parse_args()
+
+    input_dir = Path(args.input_dir)
+
+    # Validate input directory
+    if not input_dir.exists():
+        print(f"Error: Input directory does not exist: {input_dir}")
+        return
+
+    if not input_dir.is_dir():
+        print(f"Error: Input path is not a directory: {input_dir}")
+        return
+
+    # Get API key from environment (loaded from .env file)
+    api_key = os.getenv("LLAMA_CLOUD_API_KEY")
+    if not api_key:
+        print("Error: LLAMA_CLOUD_API_KEY not found. Please set it in your .env file")
+        return
+
+    # Run async function
+    asyncio.run(
+        parse_folder(
+            input_dir=input_dir,
+            max_concurrent=args.max_concurrent,
+            api_key=api_key,
+        )
+    )
+
+
+if __name__ == "__main__":
+    main()

examples/split/document_splitting/document_splitting.ipynb

@@ -0,0 +1,540 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Document Splitting with LlamaCloud\n",
+    "\n",
+    "This notebook demonstrates how to use the LlamaCloud **Split** API to automatically segment a concatenated PDF into logical document sections based on content categories.\n",
+    "\n",
+    "## Use Case\n",
+    "\n",
+    "When dealing with large PDFs that contain multiple distinct documents or sections (e.g., a bundle of research papers, a collection of reports), you often need to split them into individual segments. The Split API uses AI to:\n",
+    "\n",
+    "1. Analyze each page's content\n",
+    "2. Classify pages into user-defined categories\n",
+    "3. Group consecutive pages of the same category into segments\n",
+    "\n",
+    "## Example Document\n",
+    "\n",
+    "We'll use a PDF containing three concatenated documents:\n",
+    "- **Alan Turing's essay** \"Intelligent Machinery, A Heretical Theory\" (an essay)\n",
+    "- **ImageNet paper** (a research paper)\n",
+    "- **\"Attention is All You Need\"** paper (a research paper)\n",
+    "\n",
+    "We'll split this into segments categorized as either `essay` or `research_paper`.\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Setup\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Requirement already satisfied: llama-cloud in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (0.1.44)\n",
+      "Requirement already satisfied: python-dotenv in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (1.2.1)\n",
+      "Requirement already satisfied: requests in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (2.32.5)\n",
+      "Requirement already satisfied: certifi>=2024.7.4 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from llama-cloud) (2025.11.12)\n",
+      "Requirement already satisfied: httpx>=0.20.0 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from llama-cloud) (0.28.1)\n",
+      "Requirement already satisfied: pydantic>=1.10 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from llama-cloud) (2.12.5)\n",
+      "Requirement already satisfied: charset_normalizer<4,>=2 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from requests) (3.4.4)\n",
+      "Requirement already satisfied: idna<4,>=2.5 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from requests) (3.11)\n",
+      "Requirement already satisfied: urllib3<3,>=1.21.1 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from requests) (2.5.0)\n",
+      "Requirement already satisfied: anyio in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from httpx>=0.20.0->llama-cloud) (4.11.0)\n",
+      "Requirement already satisfied: httpcore==1.* in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from httpx>=0.20.0->llama-cloud) (1.0.9)\n",
+      "Requirement already satisfied: h11>=0.16 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from httpcore==1.*->httpx>=0.20.0->llama-cloud) (0.16.0)\n",
+      "Requirement already satisfied: annotated-types>=0.6.0 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from pydantic>=1.10->llama-cloud) (0.7.0)\n",
+      "Requirement already satisfied: pydantic-core==2.41.5 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from pydantic>=1.10->llama-cloud) (2.41.5)\n",
+      "Requirement already satisfied: typing-extensions>=4.14.1 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from pydantic>=1.10->llama-cloud) (4.15.0)\n",
+      "Requirement already satisfied: typing-inspection>=0.4.2 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from pydantic>=1.10->llama-cloud) (0.4.2)\n",
+      "Requirement already satisfied: sniffio>=1.1 in /Users/javier/llama_cloud_services/.venv/lib/python3.11/site-packages (from anyio->httpx>=0.20.0->llama-cloud) (1.3.1)\n",
+      "\n",
+      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m25.0.1\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m25.3\u001b[0m\n",
+      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpip install --upgrade pip\u001b[0m\n",
+      "Note: you may need to restart the kernel to use updated packages.\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Install required packages\n",
+    "%pip install llama-cloud python-dotenv requests"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "✅ API configured with base URL: https://api.cloud.llamaindex.ai\n",
+      "✅ Project ID: using default project\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "import time\n",
+    "import requests\n",
+    "from dotenv import load_dotenv\n",
+    "\n",
+    "# Load environment variables\n",
+    "load_dotenv()\n",
+    "\n",
+    "# Configuration\n",
+    "LLAMA_CLOUD_API_KEY = os.environ.get(\"LLAMA_CLOUD_API_KEY\", \"llx-...\")\n",
+    "BASE_URL = os.environ.get(\"LLAMA_CLOUD_BASE_URL\", \"https://api.cloud.llamaindex.ai\")\n",
+    "PROJECT_ID = os.environ.get(\"LLAMA_CLOUD_PROJECT_ID\", None)\n",
+    "\n",
+    "# Headers for API requests\n",
+    "headers = {\n",
+    "    \"Authorization\": f\"Bearer {LLAMA_CLOUD_API_KEY}\",\n",
+    "    \"Content-Type\": \"application/json\",\n",
+    "}\n",
+    "\n",
+    "print(f\"✅ API configured with base URL: {BASE_URL}\")\n",
+    "print(f\"✅ Project ID: {PROJECT_ID or 'using default project'}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Step 1: Upload the PDF File\n",
+    "\n",
+    "First, we'll upload our concatenated PDF to LlamaCloud using the Files API. This can be done using the `llama-cloud` SDK.\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "📤 Uploading ./data/turing+imagenet+attention.pdf...\n",
+      "✅ File uploaded successfully!\n",
+      "   File name: turing+imagenet+attention.pdf\n"
+     ]
+    }
+   ],
+   "source": [
+    "from llama_cloud.client import LlamaCloud\n",
+    "\n",
+    "# Initialize the client\n",
+    "client = LlamaCloud(token=LLAMA_CLOUD_API_KEY, base_url=BASE_URL)\n",
+    "\n",
+    "# Path to the PDF file\n",
+    "pdf_path = \"./data/turing+imagenet+attention.pdf\"\n",
+    "\n",
+    "# Upload the file\n",
+    "print(f\"📤 Uploading {pdf_path}...\")\n",
+    "\n",
+    "with open(pdf_path, \"rb\") as f:\n",
+    "    uploaded_file = client.files.upload_file(upload_file=f, project_id=PROJECT_ID)\n",
+    "\n",
+    "file_id = uploaded_file.id\n",
+    "print(f\"✅ File uploaded successfully!\")\n",
+    "print(f\"   File name: {uploaded_file.name}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Step 2: Create a Split Job\n",
+    "\n",
+    "Now we'll create a split job using the Split API. Since the Split API is in beta and not yet available in the SDK, we'll use raw HTTP requests.\n",
+    "\n",
+    "We define two categories:\n",
+    "- **essay**: For philosophical or reflective writing\n",
+    "- **research_paper**: For formal academic documents with methodology and citations\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "🔄 Creating split job...\n",
+      "✅ Split job created!\n",
+      "   Job ID: spl-zsssb632a742aikliu96pqkb56t5\n",
+      "   Status: pending\n",
+      "   Categories: ['essay', 'research_paper']\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Define the split job request\n",
+    "split_request = {\n",
+    "    \"document_input\": {\n",
+    "        \"type\": \"file_id\",  # only file_id is supported for now\n",
+    "        \"value\": file_id,\n",
+    "    },\n",
+    "    \"categories\": [\n",
+    "        {\n",
+    "            \"name\": \"essay\",\n",
+    "            \"description\": \"A philosophical or reflective piece of writing that presents personal viewpoints, arguments, or thoughts on a topic without strict formal structure\",\n",
+    "        },\n",
+    "        {\n",
+    "            \"name\": \"research_paper\",\n",
+    "            \"description\": \"A formal academic document presenting original research, methodology, experiments, results, and conclusions with citations and references\",\n",
+    "        },\n",
+    "    ],\n",
+    "}\n",
+    "\n",
+    "# Create the split job\n",
+    "print(\"🔄 Creating split job...\")\n",
+    "response = requests.post(\n",
+    "    f\"{BASE_URL}/api/v1/beta/split/jobs\",\n",
+    "    params={\"project_id\": PROJECT_ID},\n",
+    "    headers=headers,\n",
+    "    json=split_request,\n",
+    ")\n",
+    "response.raise_for_status()\n",
+    "\n",
+    "split_job = response.json()\n",
+    "job_id = split_job[\"id\"]\n",
+    "\n",
+    "print(f\"✅ Split job created!\")\n",
+    "print(f\"   Job ID: {job_id}\")\n",
+    "print(f\"   Status: {split_job['status']}\")\n",
+    "print(f\"   Categories: {[c['name'] for c in split_job['categories']]}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Step 3: Poll for Job Completion\n",
+    "\n",
+    "The split job runs asynchronously. We'll poll the job status until it completes.\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "⏳ Waiting for split job to complete...\n",
+      "   Status: processing (elapsed: 0s)\n",
+      "   Status: processing (elapsed: 5s)\n",
+      "   Status: processing (elapsed: 11s)\n",
+      "   Status: completed (elapsed: 16s)\n",
+      "\n",
+      "✅ Split job completed successfully!\n"
+     ]
+    }
+   ],
+   "source": [
+    "def poll_split_job(job_id: str, max_wait_seconds: int = 180, poll_interval: int = 5):\n",
+    "    \"\"\"\n",
+    "    Poll a split job until it reaches a terminal state.\n",
+    "\n",
+    "    Args:\n",
+    "        job_id: The split job ID\n",
+    "        max_wait_seconds: Maximum time to wait for completion\n",
+    "        poll_interval: Seconds between poll attempts\n",
+    "\n",
+    "    Returns:\n",
+    "        The completed job response\n",
+    "    \"\"\"\n",
+    "    start_time = time.time()\n",
+    "\n",
+    "    while (time.time() - start_time) < max_wait_seconds:\n",
+    "        response = requests.get(\n",
+    "            f\"{BASE_URL}/api/v1/beta/split/jobs/{job_id}\",\n",
+    "            params={\"project_id\": PROJECT_ID},\n",
+    "            headers=headers,\n",
+    "        )\n",
+    "        response.raise_for_status()\n",
+    "        job = response.json()\n",
+    "\n",
+    "        status = job[\"status\"]\n",
+    "        elapsed = int(time.time() - start_time)\n",
+    "        print(f\"   Status: {status} (elapsed: {elapsed}s)\")\n",
+    "\n",
+    "        if status in [\"completed\", \"failed\"]:\n",
+    "            return job\n",
+    "\n",
+    "        time.sleep(poll_interval)\n",
+    "\n",
+    "    raise TimeoutError(f\"Job did not complete within {max_wait_seconds} seconds\")\n",
+    "\n",
+    "\n",
+    "print(\"⏳ Waiting for split job to complete...\")\n",
+    "completed_job = poll_split_job(job_id)\n",
+    "\n",
+    "if completed_job[\"status\"] == \"completed\":\n",
+    "    print(\"\\n✅ Split job completed successfully!\")\n",
+    "else:\n",
+    "    print(\n",
+    "        f\"\\n❌ Split job failed: {completed_job.get('error_message', 'Unknown error')}\"\n",
+    "    )"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Step 4: Analyze the Results\n",
+    "\n",
+    "Let's examine the split results to see how the document was segmented.\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "📊 Split Results Summary\n",
+      "==================================================\n",
+      "Total segments found: 3\n",
+      "\n",
+      "Segments by category:\n",
+      "   • essay: 1 segment(s)\n",
+      "   • research_paper: 2 segment(s)\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Get the segments from the result\n",
+    "segments = completed_job.get(\"result\", {}).get(\"segments\", [])\n",
+    "\n",
+    "print(f\"📊 Split Results Summary\")\n",
+    "print(f\"=\" * 50)\n",
+    "print(f\"Total segments found: {len(segments)}\")\n",
+    "print()\n",
+    "\n",
+    "# Count by category\n",
+    "category_counts = {}\n",
+    "for segment in segments:\n",
+    "    cat = segment[\"category\"]\n",
+    "    category_counts[cat] = category_counts.get(cat, 0) + 1\n",
+    "\n",
+    "print(\"Segments by category:\")\n",
+    "for cat, count in category_counts.items():\n",
+    "    print(f\"   • {cat}: {count} segment(s)\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "📄 Segment Details\n",
+      "==================================================\n",
+      "\n",
+      "Segment 1:\n",
+      "   Category: essay\n",
+      "   Pages 1-4 (4 pages)\n",
+      "   Confidence: high\n",
+      "\n",
+      "Segment 2:\n",
+      "   Category: research_paper\n",
+      "   Pages 5-13 (9 pages)\n",
+      "   Confidence: high\n",
+      "\n",
+      "Segment 3:\n",
+      "   Category: research_paper\n",
+      "   Pages 14-24 (11 pages)\n",
+      "   Confidence: high\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Display detailed segment information\n",
+    "print(f\"\\n📄 Segment Details\")\n",
+    "print(f\"=\" * 50)\n",
+    "\n",
+    "for i, segment in enumerate(segments, 1):\n",
+    "    category = segment[\"category\"]\n",
+    "    pages = segment[\"pages\"]\n",
+    "    confidence = segment[\"confidence_category\"]\n",
+    "\n",
+    "    # Format page range\n",
+    "    if len(pages) == 1:\n",
+    "        page_range = f\"Page {pages[0]}\"\n",
+    "    else:\n",
+    "        page_range = f\"Pages {min(pages)}-{max(pages)}\"\n",
+    "\n",
+    "    print(f\"\\nSegment {i}:\")\n",
+    "    print(f\"   Category: {category}\")\n",
+    "    print(f\"   {page_range} ({len(pages)} page{'s' if len(pages) > 1 else ''})\")\n",
+    "    print(f\"   Confidence: {confidence}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Expected Results\n",
+    "\n",
+    "Based on our test document, we expect:\n",
+    "- **1 essay segment**: Alan Turing's \"Intelligent Machinery, A Heretical Theory\"\n",
+    "- **2 research paper segments**: ImageNet paper and \"Attention is All You Need\" paper\n",
+    "\n",
+    "The pages should be grouped consecutively, with no overlap between segments.\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "✅ Validation\n",
+      "==================================================\n",
+      "Total pages assigned: 24\n",
+      "Unique pages: 24\n",
+      "✅ No page overlap detected - each page belongs to exactly one segment\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Verify no page overlap\n",
+    "all_pages = []\n",
+    "for segment in segments:\n",
+    "    all_pages.extend(segment[\"pages\"])\n",
+    "\n",
+    "unique_pages = set(all_pages)\n",
+    "\n",
+    "print(f\"\\n✅ Validation\")\n",
+    "print(f\"=\" * 50)\n",
+    "print(f\"Total pages assigned: {len(all_pages)}\")\n",
+    "print(f\"Unique pages: {len(unique_pages)}\")\n",
+    "\n",
+    "if len(all_pages) == len(unique_pages):\n",
+    "    print(f\"✅ No page overlap detected - each page belongs to exactly one segment\")\n",
+    "else:\n",
+    "    print(\n",
+    "        f\"⚠️  Page overlap detected - {len(all_pages) - len(unique_pages)} duplicate assignments\"\n",
+    "    )"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Using `allow_uncategorized` Strategy\n",
+    "\n",
+    "You can also use the `allow_uncategorized` splitting strategy. This is useful when you want to capture pages that don't match any defined category.\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "📝 With allow_uncategorized=True and only 'essay' category defined,\n",
+      "   pages that don't match 'essay' will be grouped as 'uncategorized'.\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Example with allow_uncategorized strategy\n",
+    "split_request_uncategorized = {\n",
+    "    \"document_input\": {\"type\": \"file_id\", \"value\": file_id},\n",
+    "    \"categories\": [\n",
+    "        {\n",
+    "            \"name\": \"essay\",\n",
+    "            \"description\": \"A philosophical or reflective piece of writing that presents personal viewpoints, arguments, or thoughts on a topic\",\n",
+    "        }\n",
+    "        # Note: We only define 'essay' category\n",
+    "        # Research papers will be classified as 'uncategorized'\n",
+    "    ],\n",
+    "    \"splitting_strategy\": {\"allow_uncategorized\": True},\n",
+    "}\n",
+    "\n",
+    "print(\"📝 With allow_uncategorized=True and only 'essay' category defined,\")\n",
+    "print(\"   pages that don't match 'essay' will be grouped as 'uncategorized'.\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Conclusion\n",
+    "\n",
+    "The LlamaCloud Split API provides a powerful way to automatically segment concatenated documents based on content categories. This is useful for:\n",
+    "\n",
+    "- **Document processing pipelines**: Automatically separate bundled documents before further processing\n",
+    "- **Content organization**: Categorize and organize mixed document collections\n",
+    "- **Information extraction**: Identify different document types within a single file\n",
+    "\n",
+    "### Key Features\n",
+    "\n",
+    "- **AI-powered classification**: Uses LLMs to understand page content and assign categories\n",
+    "- **Flexible categories**: Define any categories relevant to your use case\n",
+    "- **Confidence scoring**: Each segment includes a confidence level\n",
+    "- **Page-level granularity**: Results include exact page numbers for each segment\n",
+    "\n",
+    "### API Reference\n",
+    "\n",
+    "- **Create Split Job**: `POST /api/v1/beta/split/jobs`\n",
+    "- **Get Split Job**: `GET /api/v1/beta/split/jobs/{job_id}`\n",
+    "- **List Split Jobs**: `GET /api/v1/beta/split/jobs`\n"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": ".venv",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

py/CHANGELOG.md

@@ -1,5 +1,11 @@
 # llama-cloud-services-py
 
+## 0.6.88
+
+### Patch Changes
+
+- 71db318: Add tier and version
+
 ## 0.6.87
 
 ### Patch Changes

py/llama_cloud_services/parse/base.py

@@ -568,6 +568,13 @@ class LlamaParse(BasePydanticReader):
         default=False,
         description="If set to true, the parser will include line-level bounding boxes in the result.",
     )
+    tier: Optional[str] = Field(
+        default=None, description="The tier to use for the parsing job."
+    )
+    version: Optional[str] = Field(
+        default=None,
+        description="The version of the parser to use at the specified tier.",
+    )
 
     # Deprecated
     bounding_box: Optional[str] = Field(
@@ -1125,6 +1132,12 @@ class LlamaParse(BasePydanticReader):
         if self.line_level_bounding_box is not None:
             data["line_level_bounding_box"] = self.line_level_bounding_box
 
+        if self.tier is not None:
+            data["tier"] = self.tier
+
+        if self.version is not None:
+            data["version"] = self.version
+
         # Deprecated
         if self.bounding_box is not None:
             data["bounding_box"] = self.bounding_box

py/llama_parse/CHANGELOG.md

@@ -1,5 +1,12 @@
 # llama_parse
 
+## 0.6.88
+
+### Patch Changes
+
+- Updated dependencies [71db318]
+  - llama-cloud-services-py@0.6.88
+
 ## 0.6.87
 
 ### Patch Changes

py/llama_parse/package.json

@@ -1,6 +1,6 @@
 {
   "name": "llama_parse",
-  "version": "0.6.87",
+  "version": "0.6.88",
   "description": "",
   "main": "index.js",
   "private": false,

py/llama_parse/pyproject.toml

@@ -11,13 +11,13 @@ dev = [
 
 [project]
 name = "llama-parse"
-version = "0.6.87"
+version = "0.6.88"
 description = "Parse files into RAG-Optimized formats."
 authors = [{name = "Logan Markewich", email = "logan@llamaindex.ai"}]
 requires-python = ">=3.9,<4.0"
 readme = "README.md"
 license = "MIT"
-dependencies = ["llama-cloud-services>=0.6.87"]
+dependencies = ["llama-cloud-services>=0.6.88"]
 
 [project.scripts]
 llama-parse = "llama_parse.cli.main:parse"

py/package.json

@@ -1,6 +1,6 @@
 {
   "name": "llama-cloud-services-py",
-  "version": "0.6.87",
+  "version": "0.6.88",
   "private": false,
   "license": "MIT",
   "scripts": {},

py/pyproject.toml

@@ -23,7 +23,7 @@ dev = [
 
 [project]
 name = "llama-cloud-services"
-version = "0.6.87"
+version = "0.6.88"
 description = "Tailored SDK clients for LlamaCloud services."
 authors = [{name = "Logan Markewich", email = "logan@runllama.ai"}]
 requires-python = ">=3.9,<4.0"

ts/llama_cloud_services/CHANGELOG.md

@@ -1,5 +1,11 @@
 # llama-cloud-services
 
+## 0.4.3
+
+### Patch Changes
+
+- 71db318: Add tier and version
+
 ## 0.4.2
 
 ### Patch Changes

ts/llama_cloud_services/package.json

@@ -1,6 +1,6 @@
 {
   "name": "llama-cloud-services",
-  "version": "0.4.2",
+  "version": "0.4.3",
   "type": "module",
   "license": "MIT",
   "scripts": {

ts/llama_cloud_services/src/LLamaCloudFileService.ts

@@ -1,7 +1,7 @@
 import {
   addFilesToPipelineApiApiV1PipelinesPipelineIdFilesPut,
   getPipelineFileStatusApiV1PipelinesPipelineIdFilesFileIdStatusGet,
-  listPipelineFilesApiV1PipelinesPipelineIdFilesGet,
+  listPipelineFiles2ApiV1PipelinesPipelineIdFiles2Get,
   listProjectsApiV1ProjectsGet,
   readFileContentApiV1FilesIdContentGet,
   searchPipelinesApiV1PipelinesGet,
@@ -97,21 +97,20 @@ export class LLamaCloudFileService {
    */
   public static async getFileUrl(pipelineId: string, filename: string) {
     initService();
-    const { data: allPipelineFiles } =
-      await listPipelineFilesApiV1PipelinesPipelineIdFilesGet({
-        path: {
-          pipeline_id: pipelineId,
-        },
-        throwOnError: true,
-      });
-    const file = allPipelineFiles.find((file) => file.name === filename);
+    const response = await listPipelineFiles2ApiV1PipelinesPipelineIdFiles2Get({
+      path: {
+        pipeline_id: pipelineId,
+      },
+      throwOnError: true,
+    });
+    const file = response.data.files.find((file) => file.name === filename);
     if (!file?.file_id) return null;
     const { data: fileContent } = await readFileContentApiV1FilesIdContentGet({
       path: {
         id: file.file_id,
       },
       query: {
-        project_id: file.project_id,
+        project_id: file.project_id || null,
       },
       throwOnError: true,
     });

ts/llama_cloud_services/src/reader.ts

@@ -186,6 +186,8 @@ export class LlamaParseReader extends FileReader {
   page_footer_suffix?: string | undefined;
   merge_tables_across_pages_in_markdown?: boolean | undefined;
   extract_printed_page_number?: boolean | undefined;
+  tier?: string | undefined;
+  version?: string | undefined;
 
   constructor(
     params: Partial<Omit<LlamaParseReader, "language" | "apiKey">> & {
@@ -383,6 +385,8 @@ export class LlamaParseReader extends FileReader {
       merge_tables_across_pages_in_markdown:
         this.merge_tables_across_pages_in_markdown,
       extract_printed_page_number: this.extract_printed_page_number,
+      tier: this.tier,
+      version: this.version,
     } satisfies {
       [Key in keyof BodyUploadFileApiParsingUploadPost]-?:
         | BodyUploadFileApiParsingUploadPost[Key]