malama 0.1.6

Universal document assistant using popular AI models (PDF, Word, Excel)

📁 Malama FileParse

Malama FileParse is a flexible and extensible framework for querying content from PDF, DOCX, and Excel files using any large language model (LLM) such as OpenAI, Gemini, Claude, Mistral, DeepSeek, and more.

🔍 Upload a file → Load context (optionally with page/range control) → Ask intelligent questions → Get accurate LLM-powered responses.

---

✨ Features

• ✅ Extracts content from PDF, DOCX, and XLSX files
• 📄 Supports full-document or partial (page-based) extraction for applicable formats
• 🤖 Works with any LLM by passing the desired model name
• 🔌 Easily extendable to new AI providers or formats
• 🧠 Unified prompt structure and clean response interface
• ⚙️ Minimal dependencies and easy integration

---

📂 Supported File Types

File Type	Description
PDF	Supports full and paginated text extraction
DOCX	Extracts paragraphs and tables in document order
XLSX	Reads sheet names and cell contents in readable format

---

🤖 Supported LLMs

Malama FileParse accepts any valid model name for the provider you configure.
Here are some examples you can use out of the box:

Provider	Example Models (not restricted)
OpenAI	gpt-3.5-turbo, gpt-4
Gemini	gemini-1.5-pro, gemini-1.0
Anthropic	claude-3-sonnet, claude-3-opus
Groq	llama3-70b-8192, mixtral-8x7b
Mistral	mistral-medium, mistral-tiny
Cohere	command-r, command-r-plus
DeepSeek	deepseek-coder, deepseek-r1
Together.ai	Qwen2-72B, Falcon-180B (via API key)
Amazon Titan	(Coming Soon via Boto3 integration)

You can pass any model name recognized by your chosen provider—no hardcoded restrictions. Prefer your own model name to work around different models of LLM provider.

---

📦 Installation

pip install malama