ostruct-cli 0.6.1

CLI for OpenAI Structured Output

ostruct-cli

Command-line interface for working with OpenAI models and structured output, powered by the openai-structured library.

Features

• Generate structured JSON output from natural language using OpenAI models and a JSON schema
• Rich template system for defining prompts (Jinja2-based)
• Automatic token counting and context window management
• Streaming support for real-time output
• Secure handling of sensitive data

Installation

For Users

To install the latest stable version from PyPI:

pip install ostruct-cli

For Developers

If you plan to contribute to the project, see the Development Setup section below for instructions on setting up the development environment with Poetry.

Shell Completion

ostruct-cli supports shell completion for Bash, Zsh, and Fish shells. To enable it:

Bash

Add this to your ~/.bashrc:

eval "$(_OSTRUCT_COMPLETE=bash_source ostruct)"

Zsh

Add this to your ~/.zshrc:

eval "$(_OSTRUCT_COMPLETE=zsh_source ostruct)"

Fish

Add this to your ~/.config/fish/completions/ostruct.fish:

eval (env _OSTRUCT_COMPLETE=fish_source ostruct)

After adding the appropriate line, restart your shell or source the configuration file. Shell completion will help you with:

• Command options and their arguments
• File paths for template and schema files
• Directory paths for -d and --base-dir options
• And more!

Quick Start

1. Set your OpenAI API key:

export OPENAI_API_KEY=your-api-key

Example 1: Using stdin (Simplest)

1. Create a template file extract_person.j2:

Extract information about the person from this text: {{ stdin }}

2. Create a schema file schema.json:

{
  "type": "object",
  "properties": {
    "person": {
      "type": "object",
      "properties": {
        "name": {
          "type": "string",
          "description": "The person's full name"
        },
        "age": {
          "type": "integer",
          "description": "The person's age"
        },
        "occupation": {
          "type": "string",
          "description": "The person's job or profession"
        }
      },
      "required": ["name", "age", "occupation"],
      "additionalProperties": false
    }
  },
  "required": ["person"],
  "additionalProperties": false
}

3. Run the CLI:

# Basic usage
echo "John Smith is a 35-year-old software engineer" | ostruct run extract_person.j2 schema.json

# For longer text using heredoc
cat << EOF | ostruct run extract_person.j2 schema.json
John Smith is a 35-year-old software engineer
working at Tech Corp. He has been programming
for over 10 years.
EOF

# With advanced options
echo "John Smith is a 35-year-old software engineer" | \
  ostruct run extract_person.j2 schema.json \
  --model gpt-4o \
  --sys-prompt "Extract precise information about the person" \
  --temperature 0.7

The command will output:

{
  "person": {
    "name": "John Smith",
    "age": 35,
    "occupation": "software engineer"
  }
}

Example 2: Processing a Single File

1. Create a template file extract_from_file.j2:

Extract information about the person from this text: {{ text.content }}

2. Use the same schema file schema.json as above.

3. Run the CLI:

# Basic usage
ostruct run extract_from_file.j2 schema.json -f text input.txt

# With advanced options
ostruct run extract_from_file.j2 schema.json \
  -f text input.txt \
  --model gpt-4o \
  --max-output-tokens 1000 \
  --temperature 0.7

The command will output:

{
  "person": {
    "name": "John Smith",
    "age": 35,
    "occupation": "software engineer"
  }
}

Example 3: Processing Multiple Files

1. Create a template file extract_from_profiles.j2:

Extract information about the people from this data:

{% for profile in profiles %}
== {{ profile.name }}

{{ profile.content }}

{% endfor %}

2. Use the same schema file schema.json as above, but updated for multiple people:

{
  "type": "object",
  "properties": {
    "people": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "name": {
            "type": "string",
            "description": "The person's full name"
          },
          "age": {
            "type": "integer",
            "description": "The person's age"
          },
          "occupation": {
            "type": "string",
            "description": "The person's job or profession"
          }
        },
        "required": ["name", "age", "occupation"],
        "additionalProperties": false
      }
    }
  },
  "required": ["people"],
  "additionalProperties": false
}

3. Run the CLI:

# Basic usage
ostruct run extract_from_profiles.j2 schema.json -p profiles "profiles/*.txt"

# With advanced options
ostruct run extract_from_profiles.j2 schema.json \
  -p profiles "profiles/*.txt" \
  --model gpt-4o \
  --sys-prompt "Extract precise information about the person" \
  --temperature 0.5

The command will output:

{
  "people": [
    {
      "name": "John Smith",
      "age": 35,
      "occupation": "software engineer"
    },
    {
      "name": "Jane Doe",
      "age": 28,
      "occupation": "data scientist"
    }
  ]
}

About Template Files

Template files use the .j2 extension to indicate they contain Jinja2 template syntax. This convention:

• Enables proper syntax highlighting in most editors
• Makes it clear the file contains template logic
• Follows industry standards for Jinja2 templates

CLI Options

The CLI revolves around a single subcommand called run. Basic usage:

ostruct run <TASK_TEMPLATE> <SCHEMA_FILE> [OPTIONS]

Common options include:

• File & Directory Inputs:

  • -f <NAME> <PATH>: Map a single file to a variable name
  • -d <NAME> <DIR>: Map a directory to a variable name
  • -p <NAME> <PATTERN>: Map files matching a glob pattern to a variable name
  • -R, --recursive: Enable recursive directory/pattern scanning

• Variables:

  • -V name=value: Define a simple string variable
  • -J name='{"key":"value"}': Define a JSON variable

• Model Parameters:

  • -m, --model MODEL: Select the OpenAI model (supported: gpt-4o, o1, o3-mini)
  • --temperature FLOAT: Set sampling temperature (0.0-2.0)
  • --max-output-tokens INT: Set maximum output tokens
  • --top-p FLOAT: Set top-p sampling parameter (0.0-1.0)
  • --frequency-penalty FLOAT: Adjust frequency penalty (-2.0-2.0)
  • --presence-penalty FLOAT: Adjust presence penalty (-2.0-2.0)
  • --reasoning-effort [low|medium|high]: Control model reasoning effort

• System Prompt:

  • --sys-prompt TEXT: Provide system prompt directly
  • --sys-file FILE: Load system prompt from file
  • --ignore-task-sysprompt: Ignore system prompt in template frontmatter

• API Configuration:

  • --api-key KEY: OpenAI API key (defaults to OPENAI_API_KEY env var)
  • --timeout FLOAT: API timeout in seconds (default: 60.0)

Debug Options

• --debug-validation: Show detailed schema validation debugging
• --debug-openai-stream: Enable low-level debug output for OpenAI streaming
• --progress-level {none,basic,detailed}: Set progress reporting level
  • none: No progress indicators
  • basic: Show key operation steps (default)
  • detailed: Show all steps with additional info
• --show-model-schema: Display the generated Pydantic model schema
• --verbose: Enable verbose logging
• --dry-run: Validate and render template without making API calls
• --no-progress: Disable all progress indicators

All debug and error logs are written to:

• ~/.ostruct/logs/ostruct.log: General application logs
• ~/.ostruct/logs/openai_stream.log: OpenAI streaming operations logs

For more detailed documentation and examples, visit our documentation.

Development

To contribute or report issues, please visit our GitHub repository.

Development Setup

1. Clone the repository:

git clone https://github.com/yanivgolan/ostruct.git
cd ostruct

2. Install Poetry if you haven't already:

curl -sSL https://install.python-poetry.org | python3 -

3. Install dependencies:

poetry install

4. Install openai-structured in editable mode:

poetry add --editable ../openai-structured  # Adjust path as needed

5. Run tests:

poetry run pytest

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.