Adjust README for better clarity

[sghandian]

Tangle-tracer repo feedback

• Move to docker instead of conda @ShivamSRS
• Move the example usage lines from inside of the script to the README.md
• Add the directory structure diagram to the README.md
• Change README from annotations/cz/ in point 1
• Also, in step 2 preprocessing, add a subsection for novel datasets instead of just "optional"; make sure that it doesn't mess up directory structure (could be confusing)
• Add aicspylibczi to REQUIREMENTS.txt (and adjust the reading in czi_to_zarr.py to include a selection of dimensions to avoid OOM issues @ShivamSRS
• Change the expected table loading location for NFT_Case_List_Batch1.csv to be in the repo instead of the data_dir. Otherwise, the user has to copy the table from the repo to the data_dir.

Object Detection

• Add a note that the dataset preprocessing has to be done after the segmentation preprocessing

[mjke]

@sghandian for:

• Move to docker instead of conda @ShivamSRS
• Add aicspylibczi to REQUIREMENTS.txt ... [snip]

I propose moving to uv instead. check out quick howto in keiserlab/rad#9