04 Image Field Candidates

04 Image Field Candidates is a small discovery helper that answers one annoying question: "which dot-paths might hold image data in my ZMongo documents?" You give it a field path, and it returns a list of candidate paths the pack's image machinery would look at. It's a debugging and compatibility tool, not a storage node - it doesn't touch the database and doesn't need a session.

Here's the situation it solves. ZMongo stores images inside documents as binary envelopes - historically at a field like image_data - and documents sometimes use a legacy layout, or the field you're pointing at is nested in a slightly different way. The pack's image load/save nodes try multiple candidate paths to stay compatible. This node just tells you which candidates those are, so you can see why a workflow found (or didn't find) your image.

How it works

It takes your field_path (default image_data) and generates the candidate list the image nodes use: the cleaned public field path, the exact path you gave, and a legacy data-path variant. The output is three shapes again:

• paths - the candidate list, useful for inspection or for feeding into a selector.
• indexed - "0: path" preview text.
• json - the full response payload.

It's AlwaysDirty, pure local computation, and effectively instant. There's no error mode worth worrying about - worst case you get an empty or short list.

The inputs and outputs that matter

• field_path - the only required input. Start with the default image_data if you're not sure; that's the pack's conventional field for images.
• Outputs: paths (list), indexed (preview), json (payload). The paths output is the one to look at.

Install

Standard pack install:

cd ComfyUI/custom_nodes
git clone https://github.com/CentralFloridaAttorney/ComfyUI-ZMongo
cd ComfyUI-ZMongo
pip install -r requirements.txt

or ComfyUI Manager → "ComfyUI-ZMongo" → restart.

Troubleshooting

• Empty or one-item paths - you gave it an unusual path; the candidates are derived from your input, so a nonstandard field_path yields fewer variants. If you're troubleshooting a load failure, compare your path against what 04 Metadata Flattened Paths reports for the actual document.
• Why multiple candidates? - legacy documents and the current layout differ, and the image nodes deliberately try the family of paths so old saved records still load. This node just exposes that family.
• The node "does nothing" - correct, it's meant to. Use it when an image won't load or you inherited a workflow with an unfamiliar field name; it tells you which names to try in 04 Load Image from Field.

It's not a node you'll keep in a production graph - it's the diagnostic you wire up, learn the answer from, and delete. That's fine; it's cheap and it does its one job.