aurora-research 0.1.0__tar.gz

aurora_research-0.1.0/PKG-INFO

@@ -0,0 +1,165 @@
+Metadata-Version: 2.3
+Name: aurora-research
+Version: 0.1.0
+Summary: Aurora 🌌 — Autonomous AI research scientist. Multi-model, ARIS-powered.
+Author: dududu
+Author-email: dududu <1816619485@qq.com>
+Requires-Dist: httpx>=0.28.1
+Requires-Dist: openai>=2.43.0
+Requires-Dist: pydantic>=2.13.4
+Requires-Dist: python-dotenv>=1.2.2
+Requires-Dist: rich>=15.0.0
+Requires-Dist: typer>=0.26.7
+Requires-Python: >=3.12
+Project-URL: Homepage, https://github.com/dududu/aurora-research
+Project-URL: Repository, https://github.com/dududu/aurora-research
+Description-Content-Type: text/markdown
+
+# Aurora Research Agent
+
+**Autonomous AI research scientist** -- multi-model, ARIS-powered scientific research agent.
+
+Aurora handles the full research workflow: literature survey, idea discovery, experiment implementation, cross-model review, and paper writing -- all from the command line.
+
+## Features
+
+- **Multi-model support** -- DeepSeek (default), OpenAI, Anthropic, and any OpenAI-compatible API
+- **Autonomous agent loop** -- Tool-using AI that reads/writes files, runs code, searches the web
+- **ARIS research workflows** -- Idea discovery, auto-review loops, paper writing, full pipeline
+- **White-label** -- Branded as "Aurora", never exposes underlying model/provider identity
+- **Cross-model review** -- Executor and reviewer use different models for adversarial quality control
+- **Resumable runs** -- State persistence enables recovery from crashes or context compaction
+
+## Quick Start
+
+### Prerequisites
+
+- Python 3.12+
+- [uv](https://docs.astral.sh/uv/) (package manager)
+- A DeepSeek API key ([get one here](https://platform.deepseek.com/api_keys))
+
+### Install
+
+```bash
+# Install from PyPI
+uv tool install aurora-research
+# Creates the `aurora-ai` command
+
+# Or try without installing
+uvx aurora-research run "Search arXiv for recent papers on LLM reasoning"
+```
+
+### Configure
+
+Set your API key:
+
+```bash
+# Option 1: Environment variable
+export DEEPSEEK_API_KEY="sk-your-key-here"
+
+# Option 2: .env file
+echo 'DEEPSEEK_API_KEY=sk-your-key-here' > .env
+
+# Option 3: Interactive config
+aurora-ai config
+```
+
+### Basic Usage
+
+```bash
+# One-shot task
+aurora-ai run "Explain the Transformer architecture with code examples"
+
+# Interactive chat
+aurora-ai chat
+
+# Plan-first execution (read-only research, then execute)
+aurora-ai plan "Build a PyTorch training pipeline for CIFAR-10"
+
+# List available models
+aurora-ai models
+```
+
+## Research Workflows
+
+### Idea Discovery
+
+```bash
+aurora-ai idea "efficient attention mechanisms for long-context LLMs" --num-ideas 8
+```
+
+Pipeline: literature survey -> idea generation (8-12 ideas) -> novelty check -> critical review -> ranked report
+
+### Auto Review Loop
+
+```bash
+aurora-ai review --target ./experiment_results/ --rounds 4
+```
+
+Cross-model adversarial review: reviewer evaluates work -> executor fixes issues -> re-review -> repeat until score >= 7/10
+
+### Paper Writing
+
+```bash
+aurora-ai paper "Efficient Attention via Hierarchical Token Pruning" --assure draft
+```
+
+Pipeline: paper plan -> figures -> LaTeX writing -> compilation -> improvement loop
+
+### Full Research Pipeline
+
+```bash
+aurora-ai pipeline "continual learning without catastrophic forgetting" --write
+```
+
+End-to-end: idea discovery -> experiments -> review -> paper
+
+## Model Configuration
+
+Aurora uses **model aliases** -- user-friendly names that hide the underlying provider:
+
+| Alias | Description |
+|-------|-------------|
+| `fast` | Fast & affordable (default) |
+| `pro` | Balanced performance |
+| `deep` | Deep reasoning |
+
+```bash
+# Use a specific model
+aurora-ai run --model pro "Complex reasoning task"
+
+# Add a custom model
+aurora-ai config add turbo --provider openai --model gpt-4o --key sk-xxx
+```
+
+## Architecture
+
+```
+CLI (aurora-ai) -> Agent Loop (tool-use engine) -> LLM Backend (multi-model)
+                         |
+              Tools: read, write, edit, bash, glob, grep, web_search, web_fetch
+                         |
+              Skills: idea-discovery, auto-review, paper-writing, pipeline
+```
+
+## Permission Modes
+
+| Mode | Behavior |
+|------|----------|
+| `default` | Ask before destructive operations |
+| `acceptEdits` | Auto-approve file edits (default for `run`) |
+| `bypassPermissions` | Auto-approve everything (use `-y` flag) |
+| `plan` | Read-only, research only |
+
+## Development
+
+```bash
+git clone https://github.com/user/aurora-research.git
+cd aurora-research
+uv sync
+uv run pytest
+```
+
+## License
+
+MIT

aurora_research-0.1.0/README.md

@@ -0,0 +1,148 @@
+# Aurora Research Agent
+
+**Autonomous AI research scientist** -- multi-model, ARIS-powered scientific research agent.
+
+Aurora handles the full research workflow: literature survey, idea discovery, experiment implementation, cross-model review, and paper writing -- all from the command line.
+
+## Features
+
+- **Multi-model support** -- DeepSeek (default), OpenAI, Anthropic, and any OpenAI-compatible API
+- **Autonomous agent loop** -- Tool-using AI that reads/writes files, runs code, searches the web
+- **ARIS research workflows** -- Idea discovery, auto-review loops, paper writing, full pipeline
+- **White-label** -- Branded as "Aurora", never exposes underlying model/provider identity
+- **Cross-model review** -- Executor and reviewer use different models for adversarial quality control
+- **Resumable runs** -- State persistence enables recovery from crashes or context compaction
+
+## Quick Start
+
+### Prerequisites
+
+- Python 3.12+
+- [uv](https://docs.astral.sh/uv/) (package manager)
+- A DeepSeek API key ([get one here](https://platform.deepseek.com/api_keys))
+
+### Install
+
+```bash
+# Install from PyPI
+uv tool install aurora-research
+# Creates the `aurora-ai` command
+
+# Or try without installing
+uvx aurora-research run "Search arXiv for recent papers on LLM reasoning"
+```
+
+### Configure
+
+Set your API key:
+
+```bash
+# Option 1: Environment variable
+export DEEPSEEK_API_KEY="sk-your-key-here"
+
+# Option 2: .env file
+echo 'DEEPSEEK_API_KEY=sk-your-key-here' > .env
+
+# Option 3: Interactive config
+aurora-ai config
+```
+
+### Basic Usage
+
+```bash
+# One-shot task
+aurora-ai run "Explain the Transformer architecture with code examples"
+
+# Interactive chat
+aurora-ai chat
+
+# Plan-first execution (read-only research, then execute)
+aurora-ai plan "Build a PyTorch training pipeline for CIFAR-10"
+
+# List available models
+aurora-ai models
+```
+
+## Research Workflows
+
+### Idea Discovery
+
+```bash
+aurora-ai idea "efficient attention mechanisms for long-context LLMs" --num-ideas 8
+```
+
+Pipeline: literature survey -> idea generation (8-12 ideas) -> novelty check -> critical review -> ranked report
+
+### Auto Review Loop
+
+```bash
+aurora-ai review --target ./experiment_results/ --rounds 4
+```
+
+Cross-model adversarial review: reviewer evaluates work -> executor fixes issues -> re-review -> repeat until score >= 7/10
+
+### Paper Writing
+
+```bash
+aurora-ai paper "Efficient Attention via Hierarchical Token Pruning" --assure draft
+```
+
+Pipeline: paper plan -> figures -> LaTeX writing -> compilation -> improvement loop
+
+### Full Research Pipeline
+
+```bash
+aurora-ai pipeline "continual learning without catastrophic forgetting" --write
+```
+
+End-to-end: idea discovery -> experiments -> review -> paper
+
+## Model Configuration
+
+Aurora uses **model aliases** -- user-friendly names that hide the underlying provider:
+
+| Alias | Description |
+|-------|-------------|
+| `fast` | Fast & affordable (default) |
+| `pro` | Balanced performance |
+| `deep` | Deep reasoning |
+
+```bash
+# Use a specific model
+aurora-ai run --model pro "Complex reasoning task"
+
+# Add a custom model
+aurora-ai config add turbo --provider openai --model gpt-4o --key sk-xxx
+```
+
+## Architecture
+
+```
+CLI (aurora-ai) -> Agent Loop (tool-use engine) -> LLM Backend (multi-model)
+                         |
+              Tools: read, write, edit, bash, glob, grep, web_search, web_fetch
+                         |
+              Skills: idea-discovery, auto-review, paper-writing, pipeline
+```
+
+## Permission Modes
+
+| Mode | Behavior |
+|------|----------|
+| `default` | Ask before destructive operations |
+| `acceptEdits` | Auto-approve file edits (default for `run`) |
+| `bypassPermissions` | Auto-approve everything (use `-y` flag) |
+| `plan` | Read-only, research only |
+
+## Development
+
+```bash
+git clone https://github.com/user/aurora-research.git
+cd aurora-research
+uv sync
+uv run pytest
+```
+
+## License
+
+MIT

aurora_research-0.1.0/pyproject.toml

@@ -0,0 +1,36 @@
+[project]
+name = "aurora-research"
+version = "0.1.0"
+description = "Aurora 🌌 — Autonomous AI research scientist. Multi-model, ARIS-powered."
+readme = "README.md"
+authors = [
+    { name = "dududu", email = "1816619485@qq.com" }
+]
+requires-python = ">=3.12"
+dependencies = [
+    "httpx>=0.28.1",
+    "openai>=2.43.0",
+    "pydantic>=2.13.4",
+    "python-dotenv>=1.2.2",
+    "rich>=15.0.0",
+    "typer>=0.26.7",
+]
+
+[project.urls]
+Homepage = "https://github.com/dududu/aurora-research"
+Repository = "https://github.com/dududu/aurora-research"
+
+[project.scripts]
+aurora-ai = "aurora_research.main:app"
+
+[build-system]
+requires = ["uv_build>=0.11.23,<0.12.0"]
+build-backend = "uv_build"
+
+[dependency-groups]
+dev = [
+    "litellm>=1.89.3",
+    "pytest>=9.1.1",
+    "pytest-asyncio>=1.4.0",
+    "ruff>=0.15.18",
+]

aurora_research-0.1.0/src/aurora_research/__init__.py

@@ -0,0 +1,6 @@
+"""Aurora 🌌 — Autonomous AI research scientist.
+
+Multi-model, ARIS-powered scientific research agent.
+"""
+
+__version__ = "0.1.0"

aurora_research-0.1.0/src/aurora_research/branding.py

@@ -0,0 +1,135 @@
+"""Aurora brand identity -- system prompts, agent personality, term filtering.
+
+All user-facing content is branded as "Aurora". Underlying model/provider
+names are stripped from responses before display.
+"""
+
+from __future__ import annotations
+
+# ── Core system prompt for the Aurora agent ──────────────────────────────
+AURORA_SYSTEM_PROMPT = """You are Aurora 🌌, an autonomous AI research scientist.
+
+You help researchers with the full scientific workflow:
+- **Literature search & survey** — multi-source search, summarization, gap analysis
+- **Idea generation & novelty checking** — brainstorming, prior art search, feasibility analysis
+- **Experiment design & execution** — planning, code implementation, GPU deployment, monitoring
+- **Code review & debugging** — quality checks, bug fixes, performance optimization
+- **Paper writing & figure generation** — LaTeX drafting, data visualization, illustration
+- **Review & rebuttal** — adversarial review, scoring, rebuttal drafting
+
+## Operating Principles
+- **Be thorough** — verify claims with evidence, not speculation
+- **Be honest** — acknowledge uncertainty and limitations clearly
+- **Be efficient** — plan before executing, avoid unnecessary work
+- **Be scientific** — follow rigorous methodology, document your process
+- **Be safe** — never run destructive commands without clear user approval
+
+## Tool Use
+You have access to tools for reading and writing files, running shell commands,
+searching the web, fetching web content, and more. Use them to accomplish
+research tasks. When you need information, search for it rather than guessing.
+
+## Output Style
+- Use clear, academic English appropriate for scientific communication
+- Cite sources with URLs when referencing papers or web content
+- Present structured analysis before drawing conclusions
+- When unsure, state your uncertainty explicitly
+
+## Research Workflow
+When conducting research, follow this general pattern:
+1. Understand the research question deeply
+2. Search and survey existing work comprehensively
+3. Analyze gaps and generate hypotheses
+4. Design experiments to test hypotheses
+5. Execute carefully and document results
+6. Draw evidence-based conclusions
+"""
+
+# ── Agent name ─────────────────────────────────────────────────────────
+AGENT_NAME = "Aurora"
+AGENT_DESCRIPTION = "Autonomous AI research scientist"
+AGENT_VERSION = "0.1.0"
+
+# ── Terms to filter from LLM output ────────────────────────────────────
+# The agent's system prompt should never mention these.
+# If the LLM produces them anyway, we can optionally filter in display.
+BLOCKED_TERMS: list[str] = [
+    # Anthropic / Claude
+    "Claude",
+    "Anthropic",
+    "claude-code",
+    "claude_code",
+    # DeepSeek
+    "DeepSeek",
+    "deepseek",
+    # OpenAI
+    "OpenAI",
+    "GPT-4",
+    "GPT-5",
+    "ChatGPT",
+    # Generic
+    "LLM provider",
+    "underlying model",
+]
+
+# ── Display helpers ────────────────────────────────────────────────────
+AGENT_BANNER = r"""
+   🌌  Aurora Research Agent
+  ═══════════════════════════════
+  Autonomous AI research scientist
+  Type /help for commands, /quit to exit
+"""
+
+# ── System prompt presets for different modes ──────────────────────────
+SYSTEM_PROMPT_PLAN = (
+    AURORA_SYSTEM_PROMPT
+    + """
+## Plan Mode
+You are currently in PLAN MODE. In this mode:
+- You may ONLY use read-only tools (read files, search web, glob, grep)
+- You may NOT modify files, run commands, or make changes
+- Your goal is to understand the task and create a detailed execution plan
+- Present the plan to the user for approval before any execution
+- Use clear section headers and numbered steps in your plan
+"""
+)
+
+SYSTEM_PROMPT_REVIEWER = """You are an independent scientific reviewer working alongside Aurora.
+
+Your role is adversarial quality assurance:
+- Critically evaluate research, code, and writing produced by the executor
+- Identify weaknesses, flaws, missing controls, overclaimed results
+- Score work on a 1-10 scale with specific, actionable feedback
+- Suggest concrete improvements for each weakness found
+
+## Review Principles
+- **Independence** — read source materials yourself, don't trust the executor's framing
+- **Skepticism** — assume claims are wrong until evidence proves otherwise
+- **Specificity** — every criticism must cite a specific line, figure, or claim
+- **Constructiveness** — every weakness should come with a proposed fix
+- **Calibration** — match review intensity to the venue (workshop < conference < journal)
+
+## Output Format
+For each review, provide:
+1. Overall score (1-10)
+2. Verdict: accept / minor revision / major revision / reject
+3. Major strengths (3-5)
+4. Major weaknesses with suggested fixes (3-5)
+5. Minor issues and typos (as a list)
+6. Detailed comments (line-by-line where applicable)
+"""
+
+
+def filter_response(text: str) -> str:
+    """Optionally filter blocked terms from LLM output text.
+
+    This is a safety net — ideally the system prompt prevents the LLM
+    from mentioning these terms at all.
+    """
+    result = text
+    for term in BLOCKED_TERMS:
+        # Case-insensitive replacement
+        import re
+
+        result = re.sub(re.escape(term), "[Aurora]", result, flags=re.IGNORECASE)
+    return result