Skip to content

Python SDK for AI agent monitoring, LLM cost tracking, benchmarking, and more. Integrates with most LLMs and agent frameworks including CrewAI, Agno, OpenAI Agents SDK, Langchain, Autogen, AG2, and CamelAI

8.3 KB
Raw Blame
import os
import re
from pathlib import Path


def clean_html_content(text):
    """Remove HTML tags and clean content for llms.txt compatibility."""
    text = re.sub(r"<[^>]+>", "", text)

    lines = text.split("\n")
    cleaned_lines = []
    in_table = False

    for line in lines:
        stripped = line.strip()

        if "|" in stripped and (stripped.startswith("|") or stripped.count("|") >= 2):
            in_table = True
            continue
        elif in_table and (stripped.startswith("-") or not stripped):
            continue
        else:
            in_table = False

        cleaned_line = re.sub(r"[^\x00-\x7F]+", "", line)

        if cleaned_line.strip() or (cleaned_lines and cleaned_lines[-1].strip()):
            cleaned_lines.append(cleaned_line)

    return "\n".join(cleaned_lines)


def convert_relative_urls(text, base_url="https://github.com/AgentOps-AI/agentops/blob/main"):
    """Convert relative URLs to absolute URLs for llms.txt compliance."""

    def replace_relative_link(match):
        link_text = match.group(1)
        url = match.group(2)

        if url.startswith(("http://", "https://", "mailto:")):
            return match.group(0)

        if url.startswith("#"):
            absolute_url = f"{base_url}/README.md{url}"
            return f"[{link_text}]({absolute_url})"

        if url.startswith("./"):
            url = url[2:]
        elif url.startswith("../"):
            url = url[3:]

        url = re.sub(r"/+", "/", url)
        url = url.strip("/")

        if not url:
            return match.group(0)

        absolute_url = f"{base_url}/{url}"
        return f"[{link_text}]({absolute_url})"

    text = re.sub(r"\[([^\]]+)\]\(([^)]+)\)", replace_relative_link, text)

    return text


def compile_llms_txt():
    """Compile a comprehensive llms.txt file with actual repository content."""

    content = "# AgentOps\n\n"

    content += "> AgentOps is the developer favorite platform for testing, debugging, and deploying AI agents and LLM apps. Monitor, analyze, and optimize your agent workflows with comprehensive observability and analytics.\n\n"

    try:
        with open("../README.md", "r", encoding="utf-8") as f:
            readme_content = f.read()
        cleaned_readme = clean_html_content(readme_content)
        cleaned_readme = convert_relative_urls(cleaned_readme)
        content += "## Repository Overview\n\n"
        content += cleaned_readme + "\n\n"
    except Exception as e:
        print(f"Warning: Could not read README.md: {e}")

    try:
        with open("../CONTRIBUTING.md", "r", encoding="utf-8") as f:
            contributing_content = f.read()
        cleaned_contributing = clean_html_content(contributing_content)
        cleaned_contributing = convert_relative_urls(cleaned_contributing)
        content += "## Contributing Guide\n\n"
        content += cleaned_contributing + "\n\n"
    except Exception as e:
        print(f"Warning: Could not read CONTRIBUTING.md: {e}")

    content += "## Core SDK Implementation\n\n"

    sdk_files = ["../agentops/__init__.py", "../agentops/client/client.py", "../agentops/sdk/decorators/__init__.py"]

    for file_path in sdk_files:
        if os.path.exists(file_path):
            try:
                with open(file_path, "r", encoding="utf-8") as f:
                    file_content = f.read()
                relative_path = os.path.relpath(file_path, "..")
                content += f"### {relative_path}\n\n```python\n{file_content}\n```\n\n"
            except Exception as e:
                print(f"Warning: Could not read {file_path}: {e}")

    content += "## Documentation\n\n"

    doc_files = ["v2/introduction.mdx", "v2/quickstart.mdx", "v2/concepts/core-concepts.mdx", "v1/quickstart.mdx"]

    for doc_file in doc_files:
        if os.path.exists(doc_file):
            try:
                with open(doc_file, "r", encoding="utf-8") as f:
                    file_content = f.read()
                cleaned_content = clean_html_content(file_content)
                cleaned_content = convert_relative_urls(cleaned_content)
                content += f"### {doc_file}\n\n{cleaned_content}\n\n"
            except Exception as e:
                print(f"Warning: Could not read {doc_file}: {e}")

    content += "## Instrumentation Architecture\n\n"

    instrumentation_files = [
        "../agentops/instrumentation/__init__.py",
        "../agentops/instrumentation/README.md",
        "../agentops/instrumentation/providers/openai/instrumentor.py",
    ]

    for file_path in instrumentation_files:
        if os.path.exists(file_path):
            try:
                with open(file_path, "r", encoding="utf-8") as f:
                    file_content = f.read()
                relative_path = os.path.relpath(file_path, "..")
                if file_path.endswith(".py"):
                    content += f"### {relative_path}\n\n```python\n{file_content}\n```\n\n"
                else:
                    cleaned_content = clean_html_content(file_content)
                    cleaned_content = convert_relative_urls(cleaned_content)
                    content += f"### {relative_path}\n\n{cleaned_content}\n\n"
            except Exception as e:
                print(f"Warning: Could not read {file_path}: {e}")

    content += "## Examples\n\n"

    example_files = [
        "../examples/openai/openai_example_sync.py",
        "../examples/crewai/job_posting.py",
        "../examples/langchain/langchain_examples.py",
        "../examples/README.md",
    ]

    for file_path in example_files:
        if os.path.exists(file_path):
            try:
                with open(file_path, "r", encoding="utf-8") as f:
                    file_content = f.read()
                relative_path = os.path.relpath(file_path, "..")
                if file_path.endswith(".py"):
                    content += f"### {relative_path}\n\n```python\n{file_content}\n```\n\n"
                else:
                    cleaned_content = clean_html_content(file_content)
                    cleaned_content = convert_relative_urls(cleaned_content)
                    content += f"### {relative_path}\n\n{cleaned_content}\n\n"
            except Exception as e:
                print(f"Warning: Could not read {file_path}: {e}")

    output_path = Path("../llms.txt")
    output_path.write_text(content, encoding="utf-8")
    print(f"Successfully compiled comprehensive llms.txt to {output_path.absolute()}")
    print(f"Total content length: {len(content)} characters")

    try:
        import llms_txt

        print("SUCCESS: llms-txt package available for validation")

        import re

        link_pattern = r"\[([^\]]+)\]\(([^)]+)\)"
        links = re.findall(link_pattern, content)

        has_h1 = content.startswith("# ")
        has_blockquote = "> " in content[:500]  # Check first 500 chars for summary
        h2_count = content.count("\n## ")

        title_match = re.match(r"^# (.+)$", content.split("\n")[0])
        title = title_match.group(1) if title_match else "Unknown"

        summary_match = re.search(r"> (.+)", content)
        summary = summary_match.group(1) if summary_match else "No summary"

        print("SUCCESS: Manual validation results:")
        print(f"   - Title: {title}")
        print(f"   - Summary: {summary[:100]}{'...' if len(summary) > 100 else ''}")
        print(f"   - H2 sections: {h2_count}")
        print(f"   - Links found: {len(links)}")
        print(f"   - Content size: {len(content)} characters")

        print("SUCCESS: Structure validation:")
        print(f"   - H1 header: {'PASS' if has_h1 else 'FAIL'}")
        print(f"   - Blockquote summary: {'PASS' if has_blockquote else 'FAIL'}")
        print(f"   - Multiple sections: {'PASS' if h2_count > 0 else 'FAIL'}")

        try:
            simple_test = "# Test\n\n> Test summary\n\n## Section\n\nContent here."
            llms_txt.parse_llms_file(simple_test)
            print("SUCCESS: llms-txt library functional (tested with simple content)")
        except Exception as simple_error:
            print(f"WARNING: llms-txt library has parsing issues: {simple_error}")

        print("INFO: For comprehensive content validation, use: https://llmstxtvalidator.dev")

    except ImportError:
        print("WARNING: llms-txt package not available, skipping library validation")
        print("INFO: Install with: pip install llms-txt")


if __name__ == "__main__":
    compile_llms_txt()