This package is used to count the CFUs in petri dishes. In order to use it you must provide it with a shared dropbox link to a directory containing the images in a png format. Only images must be contained in the directory no subdirectories.
If it misses a few its ok! it gets them by cutting and rotating the original image!
This package requires python==3.6 to install type: pip install p-tree or pip install --upgrade p-tree
- Copy the example script bellow and run it.
- An image will pop-up containing a petri dish with some (or all colonies circled).
- If not all are circled close the image and enter
Accept?message of the prompt.
- Enter after that the pixel size of the colony (estimate). The x and y axis are numbered.
- Once the pixel size is added a new image will pop-up.
If satisfied close the image and type
Accept?field. If not satisfied go back to 3.
from p_tree.run import CountColonies if __name__ == '__main__': url = 'https://www.dropbox.com/sh/vq4wb9fd9k1fz49/AADLR3IIgj8lMWs8m9QLzdPoa?dl=1' cc = CountColonies(url=url) dfs = cc.main() print(dfs)
IMPORTANT (for now)
At the end of your dropbox-link it says ?dl=0. change it to ?dl=1.
- Provide rotate angle option and number of rotations.
- Accept directory as input and not only dropbox shared link.
- Save DataFrames to csv.
- Output analytics such as bell curves (mean and std) for all petries.
- modify makefile to pip install reqs.
- modify makefile to unistall reqs.
- modify makefile to build and deploy.
- convert it into a flask extension for www.engineer-it.org.
- make video on youtube on how to use it.
- show jupyter notebook example steps.
- include images