structquery

structquery is a lightweight Python package that converts unstructured text queries into highly structured, search‑optimized outputs. By leveraging the llmatch‑messages utility and a default LLM7 model, the package extracts key elements from user input (e.g., location, difficulty, length) and returns them in a format that is ready for downstream search engines or information‑retrieval pipelines.

---

✨ Features

• One‑line transformation from free‑form text to a list of structured fields.
• Built‑in LLM7 support (ChatLLM7 from langchain_llm7) with automatic API‑key handling.
• Pluggable LLMs – you can swap in any LangChain‑compatible chat model (OpenAI, Anthropic, Google, etc.).
• Regex‑driven output validation via llmatch to guarantee format correctness.
• Zero‑config defaults – works out‑of‑the‑box for common use‑cases.

---

📦 Installation

pip install structquery

---

🚀 Quick Start

fromstructqueryimportstructquery# Simple call – uses the default ChatLLM7 modelresponse=structquery(
user_input="best hiking trails in Colorado"
)
print(response) # → List of extracted structured strings

Parameters

Name	Type	Description
user_input	str	The raw user query or description you want to structure.
llm	Optional[BaseChatModel]	Your own LangChain chat model. If omitted, the package creates a ChatLLM7 instance automatically.
api_key	Optional[str]	LLM7 API key. If omitted, the package reads LLM7_API_KEY from the environment; otherwise it falls back to the placeholder "None" (which triggers an error from the LLM service).

---

🔧 Using a Custom LLM

You can provide any LangChain‑compatible chat model that inherits from BaseChatModel. Below are examples for the most popular providers.

OpenAI

fromlangchain_openaiimportChatOpenAIfromstructqueryimportstructqueryllm=ChatOpenAI(model="gpt-4o-mini")
response=structquery(
user_input="latest smartphones under $500",
llm=llm
)

Anthropic

fromlangchain_anthropicimportChatAnthropicfromstructqueryimportstructqueryllm=ChatAnthropic(model_name="claude-3-haiku-20240307")
response=structquery(
user_input="affordable vegan meal plan",
llm=llm
)

Google Generative AI

fromlangchain_google_genaiimportChatGoogleGenerativeAIfromstructqueryimportstructqueryllm=ChatGoogleGenerativeAI(model="gemini-1.5-flash")
response=structquery(
user_input="top-rated bike touring routes in Europe",
llm=llm
)

---

🔑 LLM7 API Key

• The free tier of LLM7 supplies generous rate limits that satisfy most experimentation and production needs.
• To obtain a free API key, register at https://token.llm7.io/.
• You can provide the key in three ways:
  1. Environment variable: export LLM7_API_KEY=your_key
  2. Direct argument: structquery(user_input, api_key="your_key")
  3. Pass a pre‑configured ChatLLM7 instance via the llm argument.

---

🐞 Issues & Contributions

• Bug reports / feature requests: https://github.com/chigwell/structquery/issues
• Contributions are welcome! Fork the repository, make your changes, and submit a pull request.

---

📝 License

Distributed under the MIT License. See the LICENSE file for details.

---

📧 Contact

Eugene Evstafev – hi@euegne.plus
GitHub: https://github.com/chigwell