Return ent.text and ent.label_ in dictionary format

[musicnandpeace]

I created a custom NER spacy model and intend to create a single page app that takes in input and gives out json result of extracted entities.

nlp = spacy.load("my-model")
doc = nlp(''' Patient is taking dolo 100mg for fever. Complains of sleeplessness and pain in left leg. History of CKD. Underwent gall bladder laparoscopy in feb 2022.''')

Now to get entities and create a dictionary out of it I used the following code:

entities = list(nlp.get_pipe('ner').labels)
extracted_entities = dict.fromkeys(entities, [])

extracted_entities looks like this:

{'DIAGNOSIS': [],
 'DRUG': [],
 'LAB TEST/ RECORDING': [],
 'PROCEDURE/TREATEMENT': [],
 'SYMPTOMS_COMPLAINTS': []}

Now all i want to do is get the entities into their respective list in the extracted_entities dictionary by using this:

 for ent in doc.ents:
    label = ent.label_
    extracted_entities[ent.label_].append(ent)

But this code adds all entities in all the lists in extracted_entities dictionary. How can i get the resultant labels with their ent.text in desired dictionary / json format?

[ljvmiranda921]

Hi @musicnandpeace , this may be more of how Python's dict.fromkeys works than in spaCy. One solution is to use a dictionary comprehension:

extracted_entities = {ent: [] for ent in entities}

Also, next time please format the codeblocks using backticks (```), it helps us better debug and copy-paste your sample code

[musicnandpeace]

Replacing dict.fromkeys with dictionary comprehension is giving the proper result . Thank you so much @ljvmiranda921!

Sorry, I am a noob and would need some more guidance. The folder of best model I custom trained is named 'my-model'. It contains ner - toke2vec - vocab - config - meta -tokenizer files. When i try to load it using nlp = spacy.load("my-model") in my py file, it gives the following error:
OSError: [E050] Can't find model 'my-model'. It doesn't seem to be a Python package or a valid path to a data directory.
Although it gives no such error when i run it on my jupyter notebook.

This is what my folder structure looks like:
├── FastAPI
│ ├── pychache
│ ├── endo_ner_env
│ ├── ner.py
│ └── model
│ │ ├── init.py
│ │ ├── ner_main.py
│ │ └── my-model (contains ner - toke2vec - vocab - config - meta -tokenizer files)
│ │ └── pychache

How can i resolve this problem? Thanks in advance.