What happened?
Plot only uses vlm provider. However, when triggering the command, it still validates if a GEMINI API Key exists for image gen.
Steps to reproduce
- Do not set GOOGLE_API_KEY env.
- Run the below command：
uv run paperbanana plot --data .\data.csv --intent "Model accuracy comparison" --aspect-ratio 16:9 --vlm-provider openai
Input used
uv run paperbanana plot --data .\data.csv --intent "Model accuracy comparison" --aspect-ratio 16:9 --vlm-provider openai
Expected behavior
Should not validate if this api key exists and process the plot task
Actual output or error
uv run paperbanana plot --data .\data.csv --intent "Model accuracy comparison" --aspect-ratio 16:9 --vlm-provider openai
╭───────────────────────────────────────────╮
│ PaperBanana - Generating Statistical Plot │
│ │
│ Data: data.csv │
│ Intent: Model accuracy comparison │
│ VLM: openai / gpt-5.5 │
│ Iterations: 3 │
╰───────────────────────────────────────────╯
╭────────────────────────────────────────── Traceback (most recent call last) ──────────────────────────────────────────╮
│ D:\test\cli.py:2039 in plot │
│ │
│ 2036 │ │ pipeline = PaperBananaPipeline(settings=settings) │
│ 2037 │ │ return await pipeline.generate(gen_input) │
│ 2038 │ │
│ ❱ 2039 │ result = asyncio.run(_run()) │
│ 2040 │ console.print(f"\n[green]Done![/green] Plot saved to: [bold]{result.image_path}[/bold]") │
│ 2041 │ vector_paths = result.metadata.get("vector_output_paths", {}) │
│ 2042 │ for fmt, path in vector_paths.items(): │
│ │
│ C:\Users\test\AppData\Roaming\uv\python\cpython-3.12.12-windows-x86_64-none\Lib\asyncio\runners.py:195 in run │
│ │
│ 192 │ │ │ "asyncio.run() cannot be called from a running event loop") │
│ 193 │ │
│ 194 │ with Runner(debug=debug, loop_factory=loop_factory) as runner: │
│ ❱ 195 │ │ return runner.run(main) │
│ 196 │
│ 197 │
│ 198 def _cancel_all_tasks(loop): │
│ │
│ C:\Users\test\AppData\Roaming\uv\python\cpython-3.12.12-windows-x86_64-none\Lib\asyncio\runners.py:118 in run │
│ │
│ 115 │ │ │
│ 116 │ │ self._interrupt_count = 0 │
│ 117 │ │ try: │
│ ❱ 118 │ │ │ return self._loop.run_until_complete(task) │
│ 119 │ │ except exceptions.CancelledError: │
│ 120 │ │ │ if self._interrupt_count > 0: │
│ 121 │ │ │ │ uncancel = getattr(task, "uncancel", None) │
│ │
│ C:\Users\test\AppData\Roaming\uv\python\cpython-3.12.12-windows-x86_64-none\Lib\asyncio\base_events.py:691 in │
│ run_until_complete │
│ │
│ 688 │ │ if not future.done(): │
│ 689 │ │ │ raise RuntimeError('Event loop stopped before Future completed.') │
│ 690 │ │ │
│ ❱ 691 │ │ return future.result() │
│ 692 │ │
│ 693 │ def stop(self): │
│ 694 │ │ """Stop running the event loop. │
│ │
│ D:\test\cli.py:2036 in _run │
│ │
│ 2033 │ from paperbanana.core.pipeline import PaperBananaPipeline │
│ 2034 │ │
│ 2035 │ async def _run(): │
│ ❱ 2036 │ │ pipeline = PaperBananaPipeline(settings=settings) │
│ 2037 │ │ return await pipeline.generate(gen_input) │
│ 2038 │ │
│ 2039 │ result = asyncio.run(_run()) │
│ │
│ D:\test\core\pipeline.py:190 in __init__ │
│ │
│ 187 │ │ │ self._demo_mode = True │
│ 188 │ │ else: │
│ 189 │ │ │ self._vlm = ProviderRegistry.create_vlm(self.settings) │
│ ❱ 190 │ │ │ self._image_gen = ProviderRegistry.create_image_gen(self.settings) │
│ 191 │ │ │ self._demo_mode = False │
│ 192 │ │ │
│ 193 │ │ # Cost tracking (optional — active when budget is set or always for reporting) │
│ │
│ D:\test\providers\registry.py:173 in create_image_gen │
│ │
│ 170 │ │ logger.info("Creating image gen provider", provider=provider, model=settings.image_model) │
│ 171 │ │ │
│ 172 │ │ if provider == "google_imagen": │
│ ❱ 173 │ │ │ _validate_api_key(settings.google_api_key, "GOOGLE_API_KEY") │
│ 174 │ │ │ from paperbanana.providers.image_gen.google_imagen import GoogleImagenGen │
│ 175 │ │ │ │
│ 176 │ │ │ return GoogleImagenGen( │
│ │
│ D:\test\providers\registry.py:60 in _validate_api_key │
│ │
│ 57 │ """Raise a helpful error if the required API key is missing.""" │
│ 58 │ if key_value is None or not key_value.strip(): │
│ 59 │ │ hint = _API_KEY_HINTS.get(env_var_name, f"{env_var_name} is not set.") │
│ ❱ 60 │ │ raise ValueError(hint) │
│ 61 │
│ 62 │
│ 63 def _validate_bedrock_auth(region: str, profile: str | None) -> None: │
╰───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
ValueError: GOOGLE_API_KEY not found.
To fix this:
1. Get a free API key at: https://makersuite.google.com/app/apikey
2. Run: paperbanana setup
Or set it manually:
export GOOGLE_API_KEY=your-key-here
Operating system
Windows 11
Python version
3.12
PaperBanana version or commit hash
No response
What happened?
Plot only uses vlm provider. However, when triggering the command, it still validates if a GEMINI API Key exists for image gen.
Steps to reproduce
uv run paperbanana plot --data .\data.csv --intent "Model accuracy comparison" --aspect-ratio 16:9 --vlm-provider openaiInput used
Expected behavior
Should not validate if this api key exists and process the plot task
Actual output or error
Operating system
Windows 11
Python version
3.12
PaperBanana version or commit hash
No response