v0.2.0

v0.2.0 — Adapter Core Cleanup, Unified Errors, and Better DX

Highlights

• Unified chat API for OpenAI / Anthropic / Google (consistent params & message schema).
• Standardized error classes across providers → easier debugging.
• Cleaner request handling (temperature, max_tokens, etc.) with a single interface.

Added

• Basic error taxonomy (consistent codes/types for all supported providers).
• Minimal request/response logging for easier debugging.
• Helper utilities for normalizing generation parameters (temperature, max_tokens, top_p, etc.).
• README usage examples (minimal “chat” scenario).

Changed

• Unified message format (Prompt / User / AI → normalized internal format).
• Consistent parameter names in adapter.generate_chat_answer(...) across providers.
• Reduced dependencies — removed provider SDK wrappers in favor of direct HTTP calls.

Fixed

• Correct passing of the system prompt for all supported LLMs.
• Stable handling of network timeouts and invalid API keys (mapped to unified errors).

Breaking changes

• Updated internal message format — if you were creating messages directly from the old classes, check compatibility.
• Unified parameter names — make sure you now use temperature, max_tokens, etc. consistently.

Migration

Minimal example:

from llm_api_adapter.models.messages.chat_message import (
    AIMessage, Prompt, UserMessage
)                                               
from llm_api_adapter.universal_adapter import UniversalLLMAPIAdapter

messages = [
    Prompt(
        "You are a friendly assistant who explains complex concepts "
        "in simple terms."
    ),
    UserMessage("Hi! Can you explain how artificial intelligence works?"),
    AIMessage(
        "Sure! Artificial intelligence (AI) is a system that can perform "
        "tasks requiring human-like intelligence, such as recognizing images "
        "or understanding language. It learns by analyzing large amounts of "
        "data, finding patterns, and making predictions."
    ),
    UserMessage("How does AI learn?"),
]

adapter = UniversalLLMAPIAdapter(
    organization="openai",
    model="gpt-3.5-turbo",
    api_key=openai_api_key
)

response = adapter.generate_chat_answer(
    messages=messages,
    max_tokens=256,
    temperature=1.0,
    top_p=1.0
)
print(response.content)

Roadmap (next)

Async & streaming API (async_chat, streaming deltas).

Thanks

Thanks to everyone who provided feedback and tested early versions