Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Allow input of image properties #1337

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Allow input of image properties #1337

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
63327c9
Allow input of image properties
zhugeyicixin Jun 12, 2024
11221e9
Merge branch 'geekan:main' into main
zhugeyicixin Aug 28, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
30 changes: 24 additions & 6 deletions metagpt/provider/base_llm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -47,25 +47,43 @@ class BaseLLM(ABC):
def __init__(self, config: LLMConfig):
pass

def _user_msg(self, msg: str, images: Optional[Union[str, list[str]]] = None) -> dict[str, Union[str, dict]]:
def _user_msg(
self,
msg: str,
images: Optional[Union[str, dict[str, str], list[str | dict[str, str]]]] = None,
) -> dict[str, Union[str, dict]]:
if images:
# as gpt-4v, chat with image
return self._user_msg_with_imgs(msg, images)
else:
return {"role": "user", "content": msg}

def _user_msg_with_imgs(self, msg: str, images: Optional[Union[str, list[str]]]):
def _user_msg_with_imgs(
self,
msg: str,
images: Union[str, dict[str, str], list[str | dict[str, str]]],
):
"""
images: can be list of http(s) url or base64
"""
if isinstance(images, str):
if isinstance(images, str) or isinstance(images, dict):
images = [images]
content = [{"type": "text", "text": msg}]
for image in images:
# image_info should be a dict like this:
# {'url': str, 'detail': 'high' | 'low' | 'auto'}
# (Ref: https://platform.openai.com/docs/guides/vision/low-or-high-fidelity-image-understanding)
if isinstance(image, str):
image_info = {"url": image}
elif isinstance(image, dict):
image_info = image.copy()
else:
raise NotImplementedError
# image url or image base64
url = image if image.startswith("http") else f"data:image/jpeg;base64,{image}"
if not image_info["url"].startswith("http"):
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

doesn't check the image_info field name if passing a dict without url. maybe you should add validation.

image_info["url"] = f"data:image/jpeg;base64,{image_info['url']}"
# it can with multiple-image inputs
content.append({"type": "image_url", "image_url": {"url": url}})
content.append({"type": "image_url", "image_url": image_info})
return {"role": "user", "content": content}

def _assistant_msg(self, msg: str) -> dict[str, str]:
Expand Down Expand Up @@ -130,7 +148,7 @@ async def aask(
msg: Union[str, list[dict[str, str]]],
system_msgs: Optional[list[str]] = None,
format_msgs: Optional[list[dict[str, str]]] = None,
images: Optional[Union[str, list[str]]] = None,
images: Optional[Union[str, dict[str, str], list[str | dict[str, str]]]] = None,
timeout=USE_CONFIG_TIMEOUT,
stream=None,
) -> str:
Expand Down
Loading